(Revised) - NCRTM
(Revised) - NCRTM
(Revised) - NCRTM
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
EVALUATION:<br />
REHABILITATION<br />
APPLICATION GUIDELINES<br />
SOME<br />
(<strong>Revised</strong>)<br />
James A. Bitter and Don L. Goodyear<br />
Rehabilitatlon Research Institute<br />
Regional<br />
of Social and Rehabil1"tation Service<br />
Department<br />
of Special Education and Rehabilitation<br />
School<br />
of Northern Colorado<br />
University<br />
Greeley, Colorado<br />
840.031<br />
manual was supported, in part, by Research Grant Number 15-P-55254/8<br />
This<br />
the Social and Rehabilitation Service, Department of Health, Education,<br />
from<br />
and Wel fare.<br />
July, 1974
@ 1974<br />
Copyright<br />
A. Bitter<br />
James<br />
-11-
PREFACE<br />
manual was developed as an outgrowth of a Region VIII study group<br />
Thls<br />
program evaluation Initiated by the Regional Rehabilitation Research In-<br />
on<br />
at the Universlty of Northern Colorado. Early efforts by the study<br />
stitute<br />
in 1972 indicated that much of the program evaluation lterature in<br />
group<br />
emphasizes the nature of program evaluation, i.e. what it<br />
rehabilitation<br />
types of evaluation, etc. It was felt by study group members that a<br />
is,<br />
existed for application guidelines to program evaluation methodology<br />
need<br />
rehabilitation agency program evaluators. This manual, then, repre-<br />
for<br />
a beginning attempt to address this need and is offered as a technl-<br />
sents<br />
assistance effort to agencies by the Region VIII Regional Rehabilita-<br />
cal<br />
Research Institute. This revised edition of the manual is only In-<br />
tion<br />
as a working copy to be improved and further developed through co-<br />
tended<br />
operative efforts between researchers and practitioners<br />
acknowledgement is due many indivduals who helped to m-<br />
Grateful<br />
the February, 1974 draft of this document. Special recognition and<br />
prove<br />
s expressed to William G. Hills and Kenneth R. Reagles for de-<br />
gratitude<br />
examples for inclusion in this revision. Alo deeply appreciated<br />
veloping<br />
the many helpful suggestions offered by Dennis A. Gay, John P.<br />
were<br />
Joseph T Kunce and John E. Muthard relative to the manual's<br />
Giovannini,<br />
appropriateness of material, technical accuracy, organl-<br />
communicability,<br />
and format. Acknowledgement is also extended to Kevin G. Marshall,<br />
zation<br />
R. Rabbe and James E Tripp for their helpful assistance in edit-<br />
Stanley<br />
the February, 1974 draft. And a very special thank you is extended to<br />
ing<br />
D. Tanner for her patience and skill in editing and typing a diffl-<br />
Linnea<br />
t manuscrl pt.<br />
cul<br />
A. Btter<br />
James<br />
rector<br />
Di<br />
Rehabi litation Research Institute<br />
Regional<br />
of Northern Colorado<br />
University<br />
-111-
TABLE OF CONTENTS<br />
PREFACE III<br />
TABLE OF CONTENTS v<br />
PURPOSE 1<br />
INTRODUCTION<br />
Process<br />
Evaluation<br />
Evaluation Prlnclples<br />
Some<br />
11 ng<br />
Samp<br />
signlficance<br />
Statistical<br />
1- Random Numbers<br />
Table<br />
2 Sample Critical Values of Ch-square, the Correlation<br />
Table<br />
t and F at the .05 Probability Level<br />
Coefficient,<br />
Evaluation Objectives<br />
Sample<br />
PAR____T_T ONE E__X POST FACTO STUDY- DESIGN<br />
I<br />
EXAMPLE<br />
Sample<br />
Criteria<br />
Methodology<br />
Evaluation 6bjectlve<br />
Procedures Means, percentages<br />
Statistical<br />
Interpretation<br />
2<br />
EXAMPLE<br />
Evaluation Objective<br />
Sample<br />
Criterla<br />
Methodology<br />
Approac Questionnaire on client satisfaction<br />
Measurement<br />
Procedures One-scruple ohi-square test<br />
Statistical<br />
Interpretation<br />
3<br />
EXAMIPLE<br />
Evaluatlon bet;'ve<br />
Sample<br />
teri a<br />
Cri<br />
Methodology<br />
Approach State Agency's Case Service Report<br />
Measurement<br />
data<br />
(R-300)<br />
Procedures" Chi-square test for two independent<br />
Statistical<br />
samples<br />
I nterp ret ati on<br />
-V-<br />
3<br />
3<br />
4<br />
4<br />
5<br />
6<br />
13<br />
15<br />
15<br />
15<br />
16<br />
20<br />
26<br />
33<br />
35<br />
35<br />
35<br />
36<br />
38<br />
46<br />
49<br />
51<br />
51<br />
51<br />
52<br />
52<br />
62
Eaiution betv 65<br />
Sample<br />
65<br />
Crlteria<br />
65<br />
Methodology<br />
Approach Fo llow-p survey of rehabilitated<br />
Measurement<br />
cases<br />
Procedures Product-moment correlation<br />
Statistical<br />
66<br />
coefficients<br />
Interpretati on<br />
Evaluation betv 79<br />
Sample<br />
79<br />
Criteria<br />
79<br />
Methodology<br />
Approach. Cas__.e__e Difficulty, Index 80<br />
Measurement<br />
Procedures One-way analysis of variance 82<br />
Statistical<br />
te rp retati on<br />
102<br />
In<br />
PART TWO" ONE GROUP PRETEST POSTTEST STUDY DESIGN 103<br />
Eaiuto )bletve 107<br />
Sample<br />
107<br />
Criterla<br />
107<br />
Methodology<br />
Approach" Goal Attainment Scalng 108<br />
Measurement<br />
Procedures t-es for correlated samples 112<br />
Statistical<br />
te rpret ati on<br />
122<br />
In<br />
Ezaiuton ble:tve 125<br />
Sample<br />
125<br />
Criteria<br />
Methodol ogy<br />
Approach. The Human Service Scale 126<br />
Measurement<br />
Procedures Linear regression 134<br />
Statistical<br />
In te rp re tati on<br />
PART THREE PRETEST POSTTEST CONTROL GROUP STUDY DESIGN 147<br />
8 149<br />
EXAMPLE Eaiuaton betlve 151<br />
Sample<br />
151<br />
Criteria<br />
thodo I ogy 151<br />
Me<br />
Approach Questionnaire for client knowledge<br />
Measurement<br />
the rehabilitation process 152<br />
of<br />
Procedures t-test for ndependent samples 154<br />
Statistical<br />
Interpretatl on<br />
-vl-<br />
66<br />
74<br />
125<br />
142<br />
160
9<br />
EXAMPLE Evaluto betiv<br />
Sample<br />
teria<br />
Crl<br />
Methodology<br />
Approach. The Work Austment Rating<br />
Measwrement<br />
Procedures Analysis of covariance<br />
Statistical<br />
Interpretation<br />
PAR____T FOUR- SOLOMON FOUR_____-,G.ROUP STUDY DESIG______EN 201<br />
teri a 205<br />
Cri<br />
205<br />
Methodology<br />
Approach Test of counselor knowledge of<br />
Measurement<br />
mthods and resources 206<br />
placement<br />
Procedures. 2 x 2 factorial analysis of<br />
Statistical<br />
208<br />
variance<br />
Interpretatl on 220<br />
REFERENCES ............................. 221<br />
GLOSSARY OF SYMBOLS .................. inside back cover<br />
-vll<br />
161<br />
163<br />
163<br />
163<br />
164<br />
168<br />
196
PURPOSE<br />
Evaluatlon Some Application Guidellnes Is intended<br />
Rehabllitatlon<br />
be a working manual for rehabllitation practitioners responsible for,<br />
to<br />
interested in, program evaluar.ion and for graduate students of research<br />
or<br />
statistics It can be used i'or agency application, in-service train-<br />
and<br />
ing, or self-instruction.<br />
document attempts to be a how-to-do-t manual and workbook, thus,<br />
The<br />
the purpose of other documents which describe the nature of pro-<br />
extending<br />
evaluatlon. The manual offers 10 application examples for sample<br />
gram<br />
objectCves. Four evaluation designs which necessitate only a<br />
evaluation<br />
calculator are presented, i.e. ex post facto, one group pretest-<br />
hand<br />
pretest posttest control group, and Solomon four-group designs.<br />
posttest,<br />
manual is organized by parts corresponding to the four evaluatlon<br />
The<br />
and by examples. Each of the four parts is preceeded by a brief<br />
deslgns<br />
of the evaluation design, including its lmitations and/or<br />
description<br />
The content for each of the 10 examples is organized in the fol-<br />
value.<br />
format.<br />
lowing<br />
1 Sample evaluation objective for a rehabllitation program component.<br />
2 Criterla for evaluating the program component.<br />
Methodology for evaluation, including the general procedure, mea-<br />
3.<br />
approach (five recently developed procedures are presented in the<br />
surement<br />
manual ), and statistical procedures.<br />
4 Interpretation of the results for service delivery practice.<br />
statistical techniques, a dfferent one for each example, are pre-<br />
Ten<br />
in the followlng order" means and percentages, one-sample chi-square,<br />
sented<br />
test for two ndependent samples, product-moment correlation,<br />
chi-square<br />
analysis of variance, t-test for correlated samples, linear regres-<br />
one-way<br />
t-test for independent samples, analyss of covariance, and a 2 x 2<br />
sion,<br />
analysis of variance.<br />
factorial<br />
10 hypothetical examples are ntended to assist practitioners n<br />
The<br />
program evaluation, whether on a State agency, distrct office,<br />
conducting<br />
facility level. The materlal is offered only as worked examples or<br />
or<br />
for application of evaluation methodology. The user can sub-<br />
guidelines<br />
his own program evaluation objective and use this manual to help<br />
statute<br />
to understand and "walk" through the methodologCcal steps of evalua-<br />
him<br />
tlon and analysis.<br />
s, of course, the danger of oversimpllfcaton In a manual<br />
There<br />
as ths. It should be emphaslzed that ths manual is not intended<br />
such<br />
a substitute for knowledge and understanding about resear--E-and sta-<br />
as<br />
methodology. There are many textbooks which offer background<br />
tistical<br />
information and each user should supplement this manual with textbook n-
or quallfled consultants. Rather, It is the Intent of this gulde<br />
formation<br />
help rehabilitation practitioners apply evaluation methodology to their<br />
to<br />
Those using It are encouraged to make suggestions for its im-<br />
programs<br />
and development.<br />
provement<br />
some background n evaluation methodology, the user wll find a<br />
For<br />
by F. N Kerlnger entitled Foundations of behavioral research (New<br />
book<br />
Holt, Rinehart, & Winston, 'i964) to be an excellent source. Addl-<br />
York<br />
recommendatlons Include: J. L. Bruning & B. L. Keintz. Comput.a ,-<br />
tional<br />
handbook of statistics. Dallas Scott, Foresman and Co., 1968,<br />
tional<br />
J. R. Amos, F. L. Brown, & O. J. Mink. Statistical concepts- A_<br />
and<br />
program. New York" Harper and Row, 1965 Other excellent resources<br />
bsic<br />
E. C. Bennett & M. Weisinger. Progr.am evaluation A resource handbook<br />
are<br />
vocational rehabilitation. New York ICD Rehabilita--tion and Research<br />
for<br />
1974, and H P. Harry, R. E. Winnie, & D. M. Fisk. Practical 9__.-<br />
Center,<br />
for state and local B.overnment officials. Washington, D.C.<br />
evaluation<br />
Institute, 1973.<br />
Urban<br />
-2-
INTRODUCTION<br />
evaluation is essentially an effort to determine what<br />
"Program<br />
occur as the result of a planned program by comparing<br />
changes<br />
changes (results) with desired changes (stated goals)<br />
actual<br />
by identifying the degree to which the activity (planned<br />
and<br />
is responsible for the changes." (Trantow, 1970)<br />
program)<br />
need for program evaluation has developed rapidly in the past<br />
The<br />
years. There are man), reasons for this trend, but specifically<br />
several<br />
70's have been termed the age of accountability a fact which has a<br />
the<br />
affect on all of the human services programs including rehabllita-<br />
direct<br />
welfare, education, public health, etc. No longer are publlc funds<br />
tion,<br />
without questions asked about program value to the public and<br />
available<br />
to target service populations.<br />
we look at the actions of legislators in the past few years, it<br />
If<br />
be evident that there is an increasing need to justify the "worth"<br />
should<br />
human service programs. Federal monies have increasingly been made<br />
of<br />
to State agencies to expand current evaluation activitles and<br />
available<br />
create new evaluation units in agencies where none existed. The Reha-<br />
to<br />
Act of 1973, PL 93-112, specifically requires rehabilitation<br />
billtation<br />
project evaluation. Title IV of the new Act (Sec. 401. (a) (1)) indi-<br />
and<br />
that:<br />
cates<br />
Evaluatl on<br />
Secretary shall measure and evaluate the impact of all<br />
"The<br />
authorized by this Act, in order to determine their<br />
programs<br />
In achleving stated goals, in general, and In<br />
effectiveness<br />
to their cost, their impact on related programs, and<br />
relation<br />
structure and mechanisms for delivery of serwces, in-<br />
their<br />
where approprlate, comparisons with appropriate con-<br />
cluding,<br />
groups composed of persons who have not participated in<br />
trol<br />
programs."<br />
such<br />
Process<br />
might be expected there are no deflnite rules for determining the<br />
As<br />
of program evaluatlon. Trantow (1970) suggests several basic<br />
structure<br />
in the evaluation process (a_) define specific goals or objectives,<br />
steps<br />
establish quantifiable criteria from which to evaluate these goals, (c_)<br />
(b_]<br />
methods of comparing program outcome to the criteria, (d) design<br />
develop<br />
evaluation, (e__) collect the information needed, (f_) analyze-the re-<br />
the<br />
sults; and (.) develop conclusions and recommendations.<br />
Suchman (1971) feels that prior to undertaking evaluation,<br />
Similarly,<br />
elements must be present- (a_] the establishment of objectives or<br />
three<br />
(b) a program of intervention capable of attaining the goal, and<br />
goals,<br />
a method of determining f the goal or objective has been obtained.<br />
(.)<br />
approach s llustrated by Collignon, Zawada, Thompson, and<br />
Another<br />
(1971). They suggest that the evaluator understand what exists,<br />
Markowitz<br />
-3-
the program Inputs, and understand the results of program actl-<br />
determlne<br />
tles and the program outcomes. In order to understand program outcomes,<br />
w<br />
recommend that the evaluator ask the following questions: has the<br />
they<br />
achieved the criteria set forth why was the program effective<br />
program<br />
or not effective? and, what actions need to be taken<br />
another perspective for evaluation is suggested by Denlston,<br />
Still<br />
and Getting (1968). They list four areas of concern which are<br />
Rosenstock<br />
expressed by program directors. (a) appropriateness has the pro-<br />
most<br />
undertaken been directed toward the important concerns, (b) adequacy<br />
gram<br />
what extent have the problem areas been eliminated, (c) ffectiveness<br />
to<br />
what degree have the program objectives been attained, and (d) effi-<br />
to<br />
ciency --what were the costs in achieving the objectives<br />
excellent book whlch describes how to specify goals and objectives<br />
An<br />
R. F. Mager. Preparing i.nstruction..al obectlves. Palo Alto, CA<br />
is-<br />
Publishers, 1962. Mager (1962) describes an objective as an ntent<br />
Fearon<br />
in performance terms and contains the following conslderations-<br />
stated<br />
identlfication of the terminal performance/outcome, (b) describes<br />
(a)<br />
appropriate, conditions under which performance/outcome is expected<br />
If<br />
occur, and (c) specifies the criteria for acceptable performance/<br />
to<br />
outcome.<br />
Evaluatlon Prncipjles.<br />
Some<br />
is important, of course, that the design of a study, the measure-<br />
It<br />
approach, and the statistical procedures are appropriately applied.<br />
ment<br />
adequacy of interpretation of study results depends upon such assur-<br />
The<br />
Ex post facto studies (Examples 1 5), for example, have many<br />
ances.<br />
limitations for interpretation than do experlmental designs (Examples<br />
more<br />
10) which offer more control for the effects of unstudied variables<br />
8-<br />
must also be taken to assure that the measurement approach s valld<br />
Care<br />
reliable and the statistical procedures are appropriate. Positive<br />
and<br />
are relatively easy to interpret while negative findings are much<br />
findings<br />
difficult. Negative results may be due to nappropriate methodology,<br />
more<br />
and/or statistical procedures. On the other hand, if these<br />
measurement<br />
are adequate, then negative results can contribute to our<br />
consderatlons<br />
edge of program (Kerl inger, 1964).<br />
knowl<br />
When we sample, we are drawing a portion from some popu-<br />
Sampllng.<br />
of possible subjects. The main concern for the evaluator is to n-<br />
lation<br />
the representativeness of the sample. Large samples are more accu-<br />
sure<br />
than small samples. The smaller the sample, the greater s the pos-<br />
rate<br />
of error. However, it s often more costly to obtain larger sam-<br />
sibility<br />
more time consuming, and not always available A good pollcy to<br />
ples,<br />
is to use as large a sample as possible under the evaluator's<br />
follow<br />
rcumstances.<br />
ci<br />
are several approaches to sampling. If an evaluator wanted to<br />
There<br />
a sample of all the closed cases within his State during the past year<br />
use<br />
compare the personal characteristics of rehabilitated cases with the<br />
and<br />
cases, the evaluator could take a smple random sample.<br />
not-rehabilitated<br />
-4-
is a method of drawing a portlon of the populatlon (all closed cases<br />
Thls<br />
the fiscal year) so that each member of the population has an equal<br />
during<br />
of being selected, for example, drawing names from a hat or enter-<br />
chance<br />
a table of random numbers. To use a table of random numbers, merely<br />
ing<br />
a number to each member of the population, enter a table of random<br />
assign<br />
at any point (a sample appears as Table 1) and, moving n any pre-<br />
numbers<br />
direction, read the numbers of the individuals to be included<br />
determlned<br />
the sample. If our sample has been selected by a random method, we can<br />
in<br />
it is representative of the populatlon from which it was drawn. The<br />
say<br />
can then make inferences about the population from his sample.<br />
evaluator<br />
type of sampllng procedure Is systematic random sampl.ng.<br />
Another<br />
procedure could be used in doing a case rewew n which the evalua-<br />
Thls<br />
could take every tenth name from an alphabetical lsting or every nth<br />
tor<br />
in the file. If he chooses the latter procedure, he could select a<br />
case<br />
from 1 to 10, for example the number 6, to identify every nth case.<br />
number<br />
would then proceed to the files and pull case numbers 6, 16, 26, 36...<br />
He<br />
he has completed the selection of cases for hls study.<br />
until<br />
the other hand, an evaluator may wsh to study the comparative<br />
On<br />
of a particular training program on men and women. He would, then,<br />
mpact<br />
a representative sample of the number of men and women who went through<br />
want<br />
program. To insure such representativeness, he might use a stratified<br />
the<br />
sample. For example, if the proportion of men to women was 2 to 1,<br />
random<br />
sample should include a 2.1 proportion of men to women in order to be<br />
the<br />
considered representative.<br />
s.inificance. A level of statistical signlfcance s<br />
Statistical<br />
chosen in conductlng research In social science research com-<br />
generally<br />
levels of sgnificance are the .05 probability level and the .01 prob-<br />
mon<br />
level. This level refers to the number of times in 100 an obtained<br />
ability<br />
could occur merely by chance. For example, the .05 level means<br />
result<br />
a finding which is significant at this level could occur by chance<br />
that<br />
5 times in 100 trials. Generally, in published articles statistical<br />
only<br />
is dentified in the following manner-<br />
sgnficance<br />
9_ refers to "probabllty", < means "less than", and 05 and .01 refer<br />
where<br />
the probability levels.<br />
to<br />
critical values at the .05 probability level for chl-square,<br />
Sample<br />
correlation coefficient, t, and F statistical tests are gven n Table 2<br />
the<br />
the statistical test results equal or exceed the appropriate critical<br />
If<br />
the findings are said to be "significant" Table 2 gves critical<br />
value,<br />
only for examples used in this manual. More complete tables may<br />
values<br />
found in statistics textbooks.<br />
be<br />
-5-
1 Table<br />
Numbers<br />
Random<br />
00277 23723 27229 60600 27072 98180 44936 06731 58702<br />
55251<br />
80332 17163 93678 95418 03934 64810 43191 00792 61497<br />
35610<br />
46870 45244 88590 09738 49869 09608 96140 27241 80951<br />
62490<br />
98988 29169 33104 81989 67473 38782 31043 52586 29608<br />
15320<br />
95874 66629 19575 41256 72689 78437 89534 02501 35254 46750<br />
97797 86893 52474 71759 73424 16981 20350 93878 46078<br />
02355<br />
87757 24159 65268 20207 77187 32943 98105 75475 55627<br />
87802<br />
09042 50053 83910 56058 27350 30693 27593 50763 67489<br />
29026<br />
38973 24865 75755 05689 91453 43532 93256 34239 00556<br />
55914<br />
91057 74053 20835 89891 87082 99493 99167 38131 08877 52010<br />
10025 34448 54945 84908 44824 02226 18043 08241 90611<br />
82178<br />
76392 76037 78673 33227 01277 75409 46970 23682 34549<br />
69057<br />
25263 44620 13646 14226 80434 61480 47434 41334 46802<br />
01113<br />
22189 76654 44636 43451 23517 16904 07576 72731 65695<br />
69790<br />
29064 84525 31338 39175 72737 96858 43466 85200 77787 65988<br />
59031 65879 23226 71133 79899 83070 73786 45688 44753<br />
28904<br />
89245 89880 98245 79844 35192 97610 90245 97330 88783<br />
48379<br />
79387 29928 18678 80856 05069 85954 74623 50411 09027<br />
29325<br />
94092 85788 17796 02489 03399 01805 21021 02941 86703<br />
59029<br />
06192 20941 70417 45023 90896 91031 62830 89165 75397 44559<br />
35543 07683 45709 40591 23906 24101 48848 49240 85308<br />
07200<br />
18578 83597 77850 44418 51795 99619 61936 80796 04160<br />
67294<br />
16035 22564 41561 96559 27966 40835 13123 15966 38718<br />
26022<br />
58729 18571 49902 82539 11874 44595 89911 41204 91050<br />
47879<br />
23501 57801 32724 06785 04660 78417 40603 22602 84439 45927<br />
98259 50424 75466 47799 70064 59143 59231 43002 27199<br />
65139<br />
02174 58179 68393 92596 39349 26498 45450 47906 76317<br />
74442<br />
72281 25322 12956 27041 14839 80546 08719 95632 69105<br />
17478<br />
60992 64944 99329 72651 78197 91740 02277 51333 11367<br />
20664<br />
31772 62729 89920 76780 54712 22238 21460 04224 96783 73379<br />
29344 94207 33028 02422 15117 65932 28689 57077 73370<br />
36291<br />
44405 97704 11636 86533 29509 73252 83632 40990 19167<br />
83519<br />
47964 25765 04775 55801 89173 02722 80844 69856 07645<br />
92923<br />
82058 46808 98769 09570 41860 69274 93487 10305 22885<br />
95433<br />
83999 50298 74743 50424 76343 78247 98027 66897 24249 30608<br />
55721 86437 33564 05909 69041 68287 77580 14197 11526<br />
78964<br />
63862 17575 55461 22896 40835 42641 94695 24955 25430<br />
49278<br />
89191 04185 29797 40889 03305 87797 77840 90361 83638<br />
76883<br />
18734 53515 55274 10056 02608 78262 58530 56068 84681<br />
95431<br />
45503 99586 65880 41935 96575 60548 12842 22233 07973 88195<br />
Rand Corporation. A mllion random dlglts with 100,000<br />
Source<br />
deviates. New York. Free Press, 1955, reprinted wth permisslon<br />
nomal<br />
of the publisher.<br />
-6-
df<br />
1<br />
2<br />
3<br />
4<br />
5<br />
6<br />
7<br />
8<br />
9<br />
10<br />
11<br />
12<br />
13<br />
14<br />
15<br />
16<br />
17<br />
18<br />
Sample<br />
Chisquare<br />
Table 2<br />
Values of Chi-square, the Correlation<br />
Critical<br />
and F at the .05 Probability Level<br />
t,<br />
3.84<br />
5.99<br />
7.82<br />
9.49<br />
11.07<br />
(X 2)<br />
ati on<br />
Correl<br />
(r)<br />
Coefficient<br />
.99<br />
.95<br />
.88<br />
-7-<br />
t<br />
12.71<br />
4.30<br />
3.18<br />
2.78<br />
2.57<br />
2.45<br />
2.37<br />
2.31<br />
2.26<br />
2.23<br />
2.20<br />
2.18<br />
2.16<br />
2.15<br />
2.13<br />
2.12<br />
2.11<br />
2.10<br />
Coefficient,<br />
F<br />
161.45<br />
18.51<br />
10.13<br />
7.71<br />
6.61<br />
5.99<br />
5.59<br />
5,32<br />
5.12<br />
4,96<br />
4.84<br />
4.75<br />
4 67<br />
4.60<br />
4.54<br />
4.49<br />
199.50<br />
19.00<br />
9.55<br />
6.94<br />
5.79<br />
5.14<br />
4.74<br />
4.46<br />
4.26<br />
4.10<br />
3.98<br />
3.88<br />
3.80<br />
3.74<br />
3.68<br />
3.63
S.ample Evaluatlon O.b.ectives<br />
following represent examples of evaluation objectives in reha-<br />
The<br />
agencies The list 1"s not exhaustive, of course; nor are some<br />
bilitatlon<br />
appropriate or of interest to all agencies. They are offered<br />
examples<br />
merely to orlent the manual user to some typlcal program evaluation<br />
here<br />
which can be measured. The first 10 sample evaluation objec-<br />
objectives<br />
correspond to the 10 worked examples of this manual.<br />
tives<br />
To determine client perceptions of rehabilitation agency servlces<br />
1.<br />
employment outcomes of clients.<br />
and<br />
To determlne client satisfaction with rehabilitation services for<br />
2.<br />
eligible cases closed during three fiscal years.<br />
all<br />
To determine differences for male and female clients in services<br />
3.<br />
or arranged for by the agency during the rehabilitation process.<br />
provided<br />
To determlne the relationships among some client and service varl-<br />
4.<br />
(age, number of counselor contacts, length of employment and weekly<br />
ables<br />
ncome) for clients closed as rehabilltated (status 26) for one fiscal year<br />
To compare effectiveness of rehabilitation counselor performance<br />
5.<br />
three rehabilitation offices based on the number and case difficulty<br />
n<br />
of successful closures.<br />
To compare the effectiveness of different service delivery tech-<br />
6.<br />
with a particular type of client.<br />
niques<br />
To determine the relatlonshlp between client and/or rehabilitation<br />
7<br />
variables, e.g. cost of training and change n client need satis-<br />
process<br />
faction.<br />
To compare the effectiveness of two dfferent methods of orienta-<br />
8.<br />
to rehabilitation services for new clients.<br />
tion<br />
To compare the effectiveness of three dfferent approaches to work<br />
9.<br />
training in a rehabilitation facility<br />
adjustment<br />
To determine the effectiveness of In-service tralning on counselors'<br />
10.<br />
of placement methods and resources.<br />
knowledge<br />
To compare service effectiveness wlth the same type of cllent of<br />
11.<br />
working with specialized caseloads and counselors with gener-<br />
counselors<br />
al I zed caseloads.<br />
To compare characteristics of clients that are closed successfully<br />
12<br />
after having received rehabilitation services with those who<br />
rehabilitated<br />
have not been successfully rehabilitated.<br />
-8-
To compare the employment stability of clients who have recelved<br />
13.<br />
job placement by the counselor with clients who found their own<br />
direct<br />
empl oymen t.<br />
To detemine the benefits of rehabilitatlon service, e g. weekly<br />
14.<br />
by comparing clents who had been prov1"ded servl'ces with a contrast<br />
income,<br />
group of eligible individuals who bad not rece'ved rehabilitation service.<br />
To compare the success rate of rehabilitated clients that were re-<br />
15.<br />
public assistance payments at application with rehabilitated cli-<br />
ceivlng<br />
ents who were not receiv1"ng public assistance.<br />
-9-
PART ONE<br />
EX POST FACTO STUDY DESIGN<br />
deslgn calls for a s±ngle group that is studed only<br />
Thls<br />
tme subsequent to services. The effects are presumed to<br />
one<br />
have been the result of service (X).<br />
The evaluator dent±f±es a group of sub3ects that<br />
Method.<br />
undergone services (X) and observes (0) ths group. He<br />
have<br />
attempts to nfer that the service had an effect on the<br />
then<br />
results.<br />
obta±ned<br />
Because the observation or measurement occurs<br />
Lmtat±ons.<br />
service has already been prov±ded, the des±gn does not<br />
after<br />
the evaluator to control or manipulate variables or<br />
allow<br />
(X). It also lacks the advantage of ob3ect±ve assign-<br />
serv±ces<br />
of cl±ents to serv±ces (randomization) to control for<br />
ment<br />
dfferences. Therefore, there ±s considerable rsk for<br />
clent<br />
nterpretatons snce some results may not be due to<br />
accurate<br />
serv±ces wh±ch were provided.<br />
the<br />
Kerlnger (1964) feels that desp±te the weaknesses<br />
Value.<br />
ex post facto studies, they are necessary n social science<br />
of<br />
quest±ons which do not lend themselves to controlled<br />
research<br />
nqury.
PART ONE<br />
EX POST FACTO STUDY DESIGN<br />
design calls for a single group that is studied only<br />
This<br />
tlme subsequent to services. The effects are presumed to<br />
one<br />
have been the result of servlce (X).<br />
The evaluator Identifies a group of sub3ects that<br />
Method.<br />
undergone services (X) and observes (0) thls group. He<br />
have<br />
attempts to infer that the service had an effect on the<br />
then<br />
results.<br />
obtained<br />
Because the observation or measurement occurs<br />
Limtatlons.<br />
service has already been provided, the design does not<br />
after<br />
the evaluator to control or manipulate variables or<br />
allow<br />
(X). It also lacks the advantage of ob3ectlve assign-<br />
services<br />
of clients to services (randomization) to control for<br />
ment<br />
differences. Therefore, there is considerable rlsk for<br />
client<br />
interpretations since some results may not be due to<br />
accurate<br />
services which were provided.<br />
the<br />
Kerlinger (1964) feels that despite the weaknesses<br />
Value.<br />
ex post facto studies, they are necessary in social science<br />
of<br />
questions which do not lend themselves to controlled<br />
research<br />
Inquiry.
i<br />
i<br />
EXAMPLE<br />
A. General Procedure: Ex post facto study deslgn.<br />
Measurement approach: The Consumer Measurement<br />
B.<br />
and Consumer Measurement Follow-up<br />
Scale<br />
Questlonnalre.<br />
C. Stat±stcal Procedures: Means, percentages.<br />
example was drafted by Wllam G. Hills, Assoclate<br />
iTh±s<br />
of Management and Publlc Admnstrat±on, Pennsylvania<br />
Professor<br />
Un±versty, M±ddletown. Dr. Hills was formerly the Drector<br />
State<br />
the Reg±onal Rehabltaton Research Inst±tute, the University<br />
of<br />
Oklahoma. The Consumer Measurement Scale and the Consumer<br />
of<br />
Follow-up Quest±onna±re are reprinted wlth permission<br />
Measurement<br />
of the authors.
EXAMPLE 11<br />
A. General Procedure: Ex post facto study design.<br />
Measurement approach: The Consumer Measurement<br />
B.<br />
and Consumer Measurement Follow-up<br />
Scale<br />
Questlonnalre.<br />
C. Statlstcal Procedures: Means, percentages.<br />
1 example was drafted by Wlllam G. Hllls, Assoclate<br />
Thls<br />
of Management and Publc Administration, Pennsylvania<br />
Professor<br />
University, Middletown. Dr. Hills was formerly the Dlrector<br />
State<br />
the Regional Rehabltaton Research Institute, the University<br />
of<br />
Oklahoma. The Consumer Measurement Scale and the Consumer<br />
of<br />
Follow-up Questionnaire are reprinted wlth perm±sson<br />
Measurement<br />
of the authors.
I Sample Evaluation Objectlve<br />
determine cllent perceptlons of rehabilitatlon agency services<br />
To<br />
employment outcomes of clients.<br />
and<br />
II Criterla<br />
of satisfactlon expressed by clients for nne agency service<br />
Levels<br />
related factors at the tlme of closure as determined by the Con<br />
delivery<br />
Measurement S..c.a]e and the employment status of clients slx months<br />
sume___._r<br />
as reported on the Consumer Measurement Follow-u_ Qu.estionnaire.<br />
later<br />
A. General Procedure s<br />
III Methodology<br />
Ex post facto study design XO<br />
Steps.<br />
X service/treatment<br />
where<br />
0 observation/measurement<br />
and<br />
Administer the Consumer Measurement Scal____e (CM__S) to all cll-<br />
i<br />
of the State agency as they are closed from servlce (status 26,<br />
ents<br />
30, 08) for an entlre fiscal year period. If maled, include a<br />
28,<br />
letter and a return addressed, stamped envelope.<br />
cover<br />
Mall the Consumer Measurement (C_M) Follow-u_ Questlo.n.nalre<br />
2<br />
the same closed clients sx months after closure wth a cover let-<br />
to<br />
ter and a return addressed, stamped envelope.<br />
Upon return, score and tabulate the results for the CMS and<br />
3<br />
responses for the C__M Follow-u_ Questlonnalre.<br />
tabulate<br />
D. T. & Stanley, J. C. Experimental and quas-expermental<br />
2Campbell,<br />
for research. Chicago Rand McNally, 1966.<br />
s<br />
-15-
Calculate means for the CMS total scores and for each of nlne<br />
4<br />
service delivery rlated factors<br />
CM_._S<br />
Develop a profile of the means for the nine factors<br />
5<br />
may also be developed relative to speclfc variables, e g.<br />
Profiles<br />
dlstrict, closure status, age, sex, disability, etc.<br />
Tabulate response category frequencies for the CM Follow-u<br />
6<br />
questions and determlne percentages of response for<br />
Q.u.estionnalre<br />
response category. If desired, a one-sample ch-square test 3<br />
each<br />
be applied to determlne if there is a sgnificant difference In<br />
may<br />
for each question (application procedures are described in<br />
responses<br />
Example 2).<br />
If deslred, compare scores for the nne CMS factors wlth the<br />
7.<br />
of the same nlne factors on the CM Follow- Questionnalre to<br />
scores<br />
agreement n satCsfaction perceptlon after six months. One<br />
assess<br />
for determining agreement is to calculate a product-<br />
possibility<br />
correlation coefficient (appICcation procedures are described<br />
moment<br />
Example 4).<br />
in<br />
If not all clients respond to the CMS or CM Follow-u_ (].-<br />
8.<br />
you may want to compare characterlstcs of the obtained<br />
tonnare,<br />
wth the total client population to determine if the sample<br />
sample<br />
representative of the population. One way this may be done is by<br />
is<br />
a ch-square test for two independent examples s (applicatlon<br />
using<br />
procedures are descrlbed in Example 3).<br />
Measurement Approach The Consumer Measurement Scale and the Consumer<br />
B.<br />
Follow-u_9 questionnaire<br />
Measurement<br />
Measurement Scale_. The purpose of the CQ.nsumer Measurement<br />
Consumer<br />
s to obtain an ndicatlon of client satisfaction/dssatsfacton<br />
Scale<br />
wth rehabilitation services (the CMS appears at the back of this example).<br />
S. N_onparametrlc statistics for the behavioral sciences.<br />
3Segel,<br />
York. McGraw-Hll Book Company, 1956, pp. 4T-47<br />
New<br />
G. A. Statlstical ana!,si s i._n and educatlon.<br />
WFerguson,<br />
York McGraw-Hill ,' 1966, pp 105-115<br />
New<br />
S. NQn.parametrc statistics for thee behavloral sciences<br />
SSiegel,<br />
York McGraw-Hill, 1956, pp. 104-111<br />
New<br />
W. G. & Viaille, H. V Consumer's measurement of vocational<br />
6Hills,<br />
Norman, OK. University of 'Oklahoma Press, 19---73.<br />
rehabilitatlon.<br />
-16-
WORK SHEET
CM__S contains 28 questions which represent nlne service dellvery<br />
The<br />
aspects (factors) of agency services. Through factor analysis,<br />
related<br />
questions were found to be relevant to each of eight factors, and<br />
three<br />
questions to the ninth.<br />
four<br />
The questions which relate to each of the nne factors are as follows<br />
Factor C.ues ti ons<br />
Speed of Service Delivery<br />
1.<br />
Medical Services<br />
2.<br />
Training Services<br />
3.<br />
Employment Satisfaction<br />
4.<br />
Participation in Planning<br />
5.<br />
Counselor Effort n Placement<br />
6.<br />
Agency Policies<br />
7.<br />
Physical Facilities<br />
8.<br />
9. Personal Treatment<br />
16, 22<br />
3,<br />
9, 28<br />
6,<br />
4, 10<br />
1,<br />
8, 26<br />
7,<br />
14, 15<br />
2,<br />
17, 23<br />
13,<br />
19, 21, 24<br />
11,<br />
12, 25<br />
5,<br />
18, 20, 27<br />
to CMS questions involves placing a check mark n the box<br />
Responding<br />
best explains how the respondent feels about each statement. For<br />
that<br />
example-<br />
1. the quality of training I received...<br />
Does<br />
Very Not<br />
Very<br />
Satlsfied Satisfied Neutral Dissatisfied Dissatisfied Apply<br />
of the questionna1"re involves the asslgnment of negative<br />
Scoring<br />
for responses indicating dCssatisfaction and positive points for<br />
polnts<br />
indicating satisfaction. Thus, a client's response to each<br />
responses<br />
is scored -2 f he is "very dissatCsfied", -1 if he is "dissat-<br />
question<br />
isfied" 0 if he is "neutral" +I f he s "satlsfied" or +2 f he s<br />
"very satisfied".<br />
describe the levels of satisfaction non-technically, these re-<br />
To<br />
are reported as "high", "medium", or "low" depending upon the de-<br />
sponses<br />
of satisfaction (or dissatisfactlon) expressed for each individual<br />
gree<br />
Low levels of satisfaction range from 0 to +.70, medium levels<br />
question.<br />
from +.71 to +1.40, and high levels range from +1.41 to +2.00. The<br />
range<br />
ranges, only negative, apply to the levels of dissatisfaction The<br />
same<br />
of consumer satsfaction/dlssatisfacton levels are shown in the<br />
continuum<br />
lowing chart.<br />
fol<br />
-18-
WORK SHEET
Continuum of Satisfaction/Dissatlsfaction Levels<br />
Di ssatisfactl on<br />
Satl sfactl on<br />
-2 -1.40 -.70 0 +.70 +1.40 +2<br />
HIGH MEDIUM LOW LOW MEDIUM HIGH<br />
questions pertaining to each factor are scored according to the<br />
The<br />
of respondents and are combined to yeld a mean (average) score<br />
replies<br />
each factor. For example, if every client rated all three questions<br />
for<br />
speed of service delivery as "satisfied", the mean score for<br />
concerning<br />
of service delivery would be speed 1 1 + 1<br />
+<br />
or 1.00. The scores for all<br />
3<br />
factors are then profiled to portray a general picture of client sat-<br />
nne<br />
within an agency for a specified time interval (usually one fis-<br />
isfaction<br />
year).<br />
cal<br />
Measurement Follow-up questionnaire. The follow-up ques-<br />
Consumer<br />
seeks information about the former client's employment during<br />
tionnaire<br />
six months since closure. It contains ( six questions regarding<br />
the<br />
(b_) two related to the rehabilltation program, and (c_) nine<br />
employment,<br />
for again rating the service delivery related factors on the Co.__n-<br />
stems<br />
sumer Measurement Scale.<br />
C. Statistical Procedures<br />
1_. Score the CMS by computing response levels for each ques-<br />
Step<br />
and each of the nne factors. For example-<br />
tion<br />
Client<br />
I<br />
2<br />
3<br />
4<br />
Factor" Speed of Service Delivery<br />
Question<br />
16 22<br />
3<br />
1 0 1<br />
2 1 1<br />
1 0 -1<br />
-2 -1 0<br />
-1 1 0<br />
0<br />
5<br />
1 1 1 3<br />
Sum<br />
Sum<br />
2<br />
4<br />
0<br />
-3<br />
Mean<br />
.67<br />
1.33<br />
O0<br />
-1. O0<br />
Mean .20 ,20 .20 .20<br />
-20-<br />
.00
ORK SHEET
Client<br />
factor mean (.20) can be obtained in any of three ways: (a_) di-<br />
The<br />
the sum of all scores by the number of scores, e.g. 3 15 .20,<br />
viding<br />
dividing the sum of client means by the number of clients, e.g<br />
(b)<br />
+ 1.33 + O0 + (-1.00) + .00<br />
67<br />
20, or (c_) divlding the sum of ques-<br />
5<br />
+ .20 + .20<br />
.20<br />
tlon means by the number of questlons, e.g. 3 .20.<br />
mean satisfaction level for this sample of 5 clients for the fac-<br />
The<br />
"Speed of Service Dell"very" is .20 which suggests that this group was<br />
tor<br />
"low satisfied" with the speed of service delivery.<br />
same procedure s followed for each of the other eight factors<br />
The<br />
the CMS.<br />
of<br />
2_. Respondents' overall satisfaction with agency services is<br />
Step<br />
determined. For example"<br />
similarly<br />
FACTOR<br />
2 3 4 5 6 7 8 9<br />
1<br />
1.20 .40 .75 .90 1.00 .68 .70 .90<br />
.67<br />
1 33 1.47 .90 .89 .90 .49 1.00 .67 .45<br />
O0 70 1.30 -.60 .75 .89 91 -.91 .00<br />
-1.00 -.90 .87 45 1.25 1 54 .65 .78 .34<br />
1.00 -.98 .57 .48 1.15 -1.00 .19 .47<br />
O0<br />
1.00 2.07 2.49 2.06 4.28 5.07 2.24 1.43 2.16<br />
Sum<br />
Sum<br />
7.20<br />
8.10<br />
1 64<br />
3.98<br />
88 I.<br />
22.80<br />
Mean .20 41 .50 .41 86 1.01 .45 .29 .43 51<br />
in ths example the general level of satisfaction wlth rehabll-<br />
Thus,<br />
service delivery is posltve, but low (.51).<br />
ltaton<br />
A profile of factor means can then be made to provide a de-<br />
3_.<br />
visual picture of client satisfaction with services (See Figure<br />
scr!ptive<br />
1.1).<br />
4_ Similarly, a profile can be developed for any specific vari-<br />
Step<br />
e g. district, closure status, age, sex, disability, etc.<br />
able,<br />
-22-<br />
Mean<br />
.8O<br />
.90<br />
.18<br />
44<br />
21
WORK SHEET
-24--
WORK SHEET
5. Tabulate responses by category for each question on the CM<br />
Step<br />
questionnaire and determine percentages of response for each---<br />
Follow-up<br />
questlon. For exmple"<br />
1. Are you employed at this time?<br />
390 (65%)<br />
Yes<br />
210 (35%)<br />
No<br />
one-sample chl-square test described in Example 2 may then be<br />
The<br />
if an agency w'shes to know 1"f significantly more former clients<br />
applled<br />
employed than unemployed. The same procedure is used for each of<br />
are<br />
other questions on the CM Follow-up ..u..e.stionna're.<br />
the<br />
6. An agency may also wish to know if former clients perceived<br />
Step<br />
services similarly at closure and six months after closure.<br />
rehabilitation<br />
such a case, the product moment correlation procedures described in<br />
In<br />
4 could appropriately be used to determine the degree of agreement<br />
Example<br />
each of the nine factors between scores on the CMS and the CM Follow-<br />
for<br />
questionnaire.<br />
up<br />
7_. Rarely is there 100% response to questionnaires. This is<br />
Step<br />
true for mailed questionnalres. If the agency has some doubts<br />
especially<br />
whether the respondents are representative of the population, t is<br />
about<br />
to compare the characteristics of the sample and the population which<br />
wise<br />
consldered important by the agency for interpretation and declsion<br />
are<br />
making.<br />
example, if the agency wanted to know f the proportion of males<br />
For<br />
females was the same between the respondents (sample) and all closed<br />
and<br />
(population), a chi-square test for two independent samples could<br />
clients<br />
appropriately conducted to determine if there was a significant dif-<br />
be<br />
between the two groups. Procedures for calculating a chi-square<br />
ference<br />
for two independent samples are described in Example 3.<br />
test<br />
IV Interpretation<br />
general mean for scores on the Consumer Measurement Scale was .51.<br />
The<br />
the descrlptlve continuum of satisfaction/dissatisfactl'on levels as<br />
Using<br />
guide, this value is in the low positive range of satisfaction. Compar-<br />
a<br />
the profile of factor means with the continuum satisfaction/dissatis-<br />
ing<br />
faction levels ndicates that seven of the service dellvery related factors<br />
-26-
ORK SHEET
in the low satisfaction range and two factors were in the medlum sat-<br />
were<br />
range. Clients were most sat1"sf'ed wth "Counselor Effort in<br />
isfaction<br />
and least satisfied with "Speed of Service Delivery" and "Phys-<br />
Placement"<br />
Facilities". Addit1"onally, 65% of all closed clients were employed<br />
ical<br />
months after closure as reported on the Consumer Measurement Follow-<br />
six<br />
Qqestl onnaire.<br />
u_<br />
-28-
Consumer Measurement Scale<br />
]ER'S MEASUREIVIEHT OF VOCATIONAL REHABILITATION<br />
CONSU<br />
1973, Wllllam G. Hllls, Ph.D., Harold D. Vlallle, Ph.D.)<br />
(Copyrlght<br />
you for your time and assistance Your response to tls questionnaire m|l bo kept conhdent, ai and will be<br />
Thank<br />
only to ,denhfy the strengths and weaknesses of rehabilitation services.<br />
interpreted<br />
complete the following 28 questions while asking yourself"<br />
Please<br />
satlshed am with this aspect of rehabilitation service';<br />
How<br />
sat means am very Very satisfied<br />
means am satisfied<br />
Sat<br />
means can't decide whether am N. satisfied not or<br />
means am dissatisfied<br />
Dissot<br />
dlssat means am very Very dssatlsfled<br />
N A. means this item does not apply to me<br />
D<br />
Please place a check mark in the box that best explains how you feel about eeck statement.<br />
2. My counselor's w,||n,aess to listen to my ideas<br />
s. , 0,,,. ..,,o.<br />
s. . emp..,, .0. ...<br />
m. x,.d ...,.,.o, ,.--,-"<br />
. vo.,,... R.bo,,,,..', ab,,,y,.... ..,,,0.<br />
. Tbo..,. ,,..<br />
'. ,*-,. pan.,.o , - ,ab<br />
S. tbo ',--,' '00 ,* o.' ,. ,.--,--. -*-<br />
20 The pers.nol treatment re{e,ved from ,bo med,{.l<br />
21. Tbo nomr of fos and omoo.t of pape¢wrk<br />
26. How my employment hts my mental and phys,cai abd,t,es<br />
PLEASE TURN THE PAGE AND COMPLETE THE FOLLOWING QUESTIONS<br />
Very<br />
Very<br />
t. N D,$st. Dist D 14 A.<br />
t.<br />
Very<br />
Very<br />
Sat. N DiJsat. Dinsat. D N A<br />
Sat.
D,sabhty<br />
l.ast grade completed n<br />
school<br />
training you have received<br />
Other<br />
Wh,te [] Negro [<br />
Race<br />
Sex Male<br />
Present Job<br />
Years completed ,n<br />
ind,en O sp ,,sh 0 Other 0<br />
Referral referred myself to VR [ Someone else referred me<br />
Please answer each of the following<br />
college<br />
(1) Would you return to Vocational Rehabll,tatlon (VR) should you again have need of VR services Yes [-1 No<br />
Do you think the serwces prowded by VR helped you get a better lob than you would have found without VR services 9<br />
(2)<br />
NoJ--]<br />
J"<br />
Yes<br />
(3) In your contact wsth VR, did you ever in any way experience d, scrlmmatlen+ Yes J"]<br />
How do you think rehabd,tutlon services could be improved<br />
ADDITIONAl COMMENTS<br />
are interested n knowing what happens to people after rehabilitation services end. Would you be willing to<br />
We<br />
another questionnaire for us about six months from now<br />
complete<br />
Yes/"] No<br />
yes, please print your name and address so we can send you the follow-up questionnaire<br />
Name<br />
Address.,<br />
Street Cty State Ztp Code<br />
Please return this questionnaire in the enclosed envelope to<br />
Rehabdtatlon Research Institute<br />
Reglono|<br />
University of Oklahoma<br />
The<br />
District Number
Measurement Fol 1 ow-up questl onnal re<br />
Consumer<br />
|E SURE , E IT OF VOCATIO AL REHABILITATIO<br />
I<br />
(Copyrzght, 1973, Wzllzam G. HzlIs, Ph.D., Harold D. Vzazlle, Ph.D.)<br />
Today's Date<br />
1. Are you employed at th,s t,me Yes No.<br />
Note: If no, please skip questions 2, 3, and 4.<br />
2. Are you worksng for the same employer you were six months ago<br />
Yes No<br />
you do,ng the same k,nd of work you were doing sx months ago<br />
Are<br />
No<br />
Yes<br />
4. Now sat, sfted are you wtth your present<br />
Sat, stied Sat, slued<br />
Very<br />
Dssatsfled<br />
Very<br />
5. How many months during the last sx have you been employed ?<br />
0 2 3 4 5 6<br />
6. How many lobs have )ou had ,n the last sx months 9<br />
0 2.........._3 4 5<br />
7. Have you again appl,ed for rehabilitation services 9<br />
Yes. No<br />
Neutral Dssatshed<br />
6 or more<br />
What else could the Vocational Rehabilitation program have done that would have been of help to you n hnd-<br />
8.<br />
or keeping su,table employment 9<br />
Ing
Ask Yourself How satisfied am with this aspect of rehabilitation services';<br />
sat. means am very Very satisfied<br />
means am satisfied<br />
Sat.<br />
means can't decde whether am N satisfied not or<br />
means am dissatisfied<br />
Dssat.<br />
dlssat, means am very dssatsfled<br />
Very<br />
A. means this tem does not apply to me<br />
D.N<br />
Please place a check mark In the box that best explains how you feel about each statement.<br />
Choose an answer for all statements.<br />
IN MY EXPERIENCE WITH VOCATIONAL REHABIL.ITATION, THIS IS THE WAY FEEL. ABOUT<br />
9 The tme it took to get the serwces started<br />
10 Results of medical services<br />
11 The quality of traanlng receIved<br />
12<br />
The employment I now have<br />
13 My counselor's w, llmgness to listen to my adeas and suggestions..<br />
14 The part my counselor played In actually helpIng me get my lob.<br />
15. Vocational Rehab, lltatlon's ab, ity to make decisions<br />
16 l:ase with which could enter the office<br />
17 Personal treatment received from Vocational Rehabilitation<br />
Very<br />
Very<br />
Sat. N Dissat. Dissat.<br />
Sat.<br />
E]<br />
DE]DE]<br />
DE] 0 CI<br />
CI<br />
n D ODD<br />
OOO O<br />
...0<br />
O O O CI<br />
O<br />
O []] O O<br />
l,l<br />
mOO[]]<br />
O<br />
..mOOD<br />
put this cluestonnalre n the attached enveiope and ma, to the Regional Rehabilitation Research<br />
Please<br />
Thank you for your cooperation<br />
Institute<br />
REHABILITATION RESEARCH INSTITUTE<br />
REGIONAL<br />
Un:vers,ty of Oklahoma<br />
The<br />
Questionnaire Number<br />
DI stnct Number<br />
D.N.
EXAMPLE 2<br />
A. General Procedure: Ex post facto study des±gn.<br />
B. Measurement Approach: Questlonnalre on cllent satsfact±on.<br />
C. Statistical Procedures: One-sample ch±-square test.
EXAMPLE 2<br />
A. General Procedure: Ex post facto study deslgn.<br />
B. Measurement Approach: Questlonnare on clent satisfaction.<br />
C. Statlstlcal Procedures: One-sample ch-square test.
Sample Evaluation Objectlve<br />
determine cllent satisfaction with rehab111tation services for<br />
To<br />
eligible cases closed durlng three fiscal years.<br />
all<br />
II Criteria<br />
of clients expressing satsfactlon with rehabllitation serv-<br />
Number<br />
after having received servlce.<br />
ices<br />
A. General Procedure7<br />
III Methodology<br />
Ex post facto study design XO<br />
Steps.<br />
X servlce/treatment<br />
where<br />
0 = observation/measurement<br />
and<br />
Identlfy clients who have been closed as rehabilitated or<br />
1.<br />
for the past 3 fiscal years.<br />
not-rehabilitated<br />
2 Develop a questonnalre to determine cllent satsfactlon<br />
Mal survey to former cllents explaining the survey and the<br />
3.<br />
for them to respond.<br />
need<br />
Send a follow-up letter again seeklng cooperation to all cli-<br />
4<br />
who did not respond.<br />
ents<br />
D. T. & Stanley, J. C. Experimental and quasi-experimental<br />
7Campbell,<br />
for research Chicago" Rand McNally, 1966.-'--<br />
designs<br />
-35-
Tabulate results and do the necessary statistics to make the<br />
5.<br />
meaningful. One possibility Cs a one-sample chi-square test e.<br />
results<br />
B. Measurement Approach- Questionnalre on client satisfaction<br />
way to learn about client satisfaction with services provlded by<br />
One<br />
rehabilitatlon agency is to use a questionnaire with former clients.<br />
the<br />
typical related questions taken from a questionnaire used by one<br />
Some<br />
VIII State rehabilitation agency are<br />
Region<br />
How much do you feel your counselor and services provided by the<br />
1.<br />
agency helped you<br />
rehabilitation<br />
Great deal<br />
Only a little<br />
Not at all<br />
Were you satisfied wlth the help the counselor gave you In flnd-<br />
2.<br />
a job<br />
ing<br />
Yes<br />
No<br />
3. On the average how often dd you see your counselor<br />
2. etc.<br />
Once<br />
Every couple of weeks<br />
About once a month<br />
Every couple of months<br />
Less than every couple of months<br />
Never<br />
S Nonparametric statistics for the behavioral sciences.<br />
eSiegel,<br />
York- McGraw-Hill Book Company, 1956, pp. 42-47.<br />
New<br />
-36-
ORK $ttEET
C. Statlstical Procedures<br />
1. Tabulate frequencies for the various response categorles for<br />
Step<br />
question as a first step in the chi-square analysis. For example<br />
each<br />
How much do you feel your counselor and servlces provlded<br />
1<br />
the rehabllitation agency helped you<br />
by<br />
78 Great deal<br />
Only a little<br />
16<br />
Not at all<br />
2<br />
the 101 questlonnaire respondents, 5 dld not respond to<br />
Of<br />
question.<br />
tbls<br />
Were you satisfied wlth the help tne counselor gave you in<br />
2.<br />
a 3ob'<br />
finding<br />
29 Yes<br />
19 No<br />
the 101 questionnaire respondents, 52 did not respond to<br />
Of<br />
questlon. Ths may indicate that the question did not<br />
this<br />
former clients adequate categories to respond, e.g.<br />
allow<br />
satisfied nor dissatisfied.<br />
neither<br />
3. On the average bow often dd you see your counselor?<br />
7 Once<br />
26 Every couple of weeks<br />
30 About once a month<br />
19 Every couple of months<br />
12 Less than every couple of months<br />
2 Never<br />
-38-
WORK SHEET
the 101 questionnaire respondents, 5 dld not respond to<br />
Of<br />
question.<br />
this<br />
2. Ut111ze the tabulated frequencies in the chi-square formula,<br />
Step<br />
is as follows.<br />
which<br />
: z (o ,E),<br />
x<br />
E<br />
=1<br />
degrees of freedom<br />
df=k-1<br />
/¢ = number<br />
where<br />
classification<br />
of<br />
categories.<br />
r. directs us to sum the resultant calculation for<br />
where:<br />
each category over all categories (/c).<br />
=1<br />
0 = the observed frequency in a gven category.<br />
= the frequency we would expect in a given category.<br />
E<br />
a one-sample chi-square problem such as ths,<br />
In<br />
is expected that there will be equal dstribu-<br />
t<br />
of frequencles for all categories.<br />
tion<br />
for the responses to Question Number I the observed and expected<br />
Thus,<br />
would look llke this<br />
frequencles<br />
0 E<br />
78 Great deal<br />
Only a little<br />
Not at all<br />
Step 3_. Apply the chi-square formula to the data.<br />
: . (o E}<br />
x<br />
E<br />
=1<br />
-40-
WORK SHEET
Number 1 How much do you feel your counselor and servlces<br />
Questlon<br />
by the reha'bTlitation agency helped you<br />
provided<br />
78 Great deal<br />
16 Only a little<br />
2 Not at all<br />
the calculations necessary for the chi-square formula,<br />
Performing<br />
results can be shown in the following way:<br />
the<br />
dr= /¢- 1<br />
=3-I<br />
=2<br />
E O-E (O-E) 2 (O-E) 2<br />
0<br />
E<br />
78 32 46 2116 66.13 ÷ Major contributor<br />
16 32 -16 256 8.00<br />
2 32 -30 900 28.13<br />
X = 102.26<br />
a table of crit1"cal values for X (found in statistics text-<br />
Using<br />
for 2 degrees of freedom, the necessary chi-square value for sig-<br />
books)<br />
is 5.99 at the .05 probabil1"ty level. Th's means that only 5<br />
nificance<br />
out of 100 would we expect a X 2 of 5.99 or greater to happen by<br />
times<br />
There is, of course, the possibility of error, but when we select<br />
chance.<br />
probability level, we do so to make a decision, one way or the other,<br />
a<br />
significance. The X value of 102.26 for question 1 exceeds a X 2<br />
about<br />
5.99, therefore the 9_for ths calculate'on is less than .05 (in fact,<br />
of<br />
it is less than .001).<br />
The calculations for Question Number 2 are done n the same manner.<br />
-42-
WORK SHEET
Number 2: Were you satisfied with the help the counselor<br />
question.<br />
you in f1"nding a job<br />
gave<br />
29 Yes<br />
19 No<br />
df=/-I<br />
=2-1<br />
=1<br />
0 E O-E (O-E) 2 (O-E) 2<br />
29 24 5 25 1.04<br />
19 24 -5 25 1.04<br />
using a table of critical values, the X 2 value for 1 degree of<br />
Again<br />
is 3.84 at the .05 probability level. Thls means that only 5 times<br />
freedom<br />
of 100 would we expect a X 2 of 3.84 or greater to happen by chance.<br />
out<br />
the X 2 value of 2.08 s not significant.<br />
Therefore,<br />
calculations for Question Number 3 are done in the same manner<br />
The<br />
questions I and 2.<br />
as<br />
Number 3. On the average how often did you see your counselor?<br />
q.u.estion<br />
Once<br />
7<br />
26 Every couple of weeks<br />
30 About once a month<br />
19 Every couple of months<br />
12 Less than every couple of weeks<br />
2 Never<br />
-44-<br />
X 2<br />
E<br />
--<br />
2.08
WORK SHEET
0<br />
7<br />
26<br />
3O<br />
19<br />
12<br />
2<br />
df=k-1<br />
=6-1<br />
:5<br />
O-E (O-E) 2 (O-E) 2<br />
E<br />
E<br />
16 -9 81 5.06<br />
16 10 100 6.25<br />
16 14 196 12.25 ÷ Maor contrlbutor<br />
16 3 9 .56<br />
16 -4 16 1.00<br />
16 -14 196 12.25 ÷ Maor contributor<br />
X<br />
=<br />
37.37<br />
using a table of crit1"cal values, the X 2 value for 5 degrees<br />
Again<br />
freedom is 11.07 at the .05 probabClty level The X 2 value of 37.37<br />
of<br />
indicates a probabil1"ty of less than .05.<br />
IV Interpretation<br />
can be seen by the obtained X for Question 1, signCficant differ-<br />
As<br />
in response frequency from the expected were obtained at the .05<br />
ences<br />
level because the obtained X 2 value exceeded 5.99. (In fact,<br />
probability<br />
obtained X 2 value exceeded the critical value of 13.82 at the .001 prob-<br />
the<br />
level.) As can be observed from the calculation for this question<br />
ability<br />
much do you feel your counselor and services provided by the rehabil-<br />
"How<br />
agency helped you" the greatest contributor to the chi-square<br />
itation<br />
(102.26) was the response to the category "great deal" (66.13). More<br />
value<br />
(78) indicated "great deal" than expected (32) and fewer than<br />
respondents<br />
indicated "only a 11"ttle" (16) or "not at all" (2). Ths indi-<br />
expected<br />
that former clients were very satisfied (significantly) with their<br />
cates<br />
counselor and the services provided.<br />
outcome was different in Questlon 2 where the responses were not<br />
The<br />
different than expected at the .05 probabllity level. For<br />
signlficantly<br />
-46-
WORK SHEET
question many clients did not respond to the item (52 of them). For<br />
this<br />
that did respond, there was not sufficient difference between the<br />
those<br />
observed and the expected frequencies to suggest sgnCficance.<br />
question 3, statistical slgnfCcance was attained at the .05 prob-<br />
For<br />
level, indCcating that there was a dl'fference n the number of<br />
ability<br />
clients were seen by their counselors. The most frequent category<br />
times<br />
about once a month w'th very few clients never seen by their counselor.<br />
was<br />
-48-
EXAMPLE 3<br />
A. General Procedure: Ex post facto study des±gn.<br />
Measurement Approach: State Agency's Case Servlce Report<br />
B.<br />
data.<br />
(R-300)<br />
Statlstcal Procedures: Ch-square test for two ndepen-<br />
C.<br />
samples.<br />
dent
EXAMPLE 3<br />
A. General Procedure: Ex post facto study design.<br />
Bo<br />
Co<br />
Approach: State Agency's Case Servlce Report<br />
Measurement<br />
data.<br />
(R-300)<br />
Procedures: Ch-square test for two ndepen-<br />
Statistical<br />
samples.<br />
dent
(r- 1)(c- 1)<br />
df<br />
(2 1)(2 1)<br />
=<br />
=1<br />
a table of critical values for X 2 for 1 degree of freedom, the<br />
Using<br />
chi-square value for significance is 3.84 at the .05 level. Thus,<br />
necessary<br />
the X 2 value of 68.66 is significant at the .05 probability level.<br />
IV Interpretation<br />
can be seen by the obtained X for example 1 relative to diagnos-<br />
As<br />
and evaluation services for males and females, sgnlficance was not<br />
tic<br />
at the .05 probability level because the obtained X value (.41)<br />
reached<br />
not exceed 3.84. Th1"s Ind'cates that there is not a signflcant dif-<br />
did<br />
between males and females who were prov'ded diagnostic and evalu-<br />
ference<br />
ation services.<br />
contrast, the outcome for example 2 relative to business school<br />
In<br />
college training indicates that significance was reached at the 05<br />
or<br />
A X value of 68.66 is greater than the needed critical value of<br />
level.<br />
for statistical significance. Thus, differences do ex1"st between<br />
3.84<br />
number of men and women who are sent to business school or college.<br />
the<br />
of the major contributors to the X value indicates that cells<br />
Inspection<br />
and D offer the most to the X 2 total. Fewer men and more women than<br />
B<br />
be expected by chance alone (except 5 times out of 100) are sent<br />
could<br />
to business school or college for training.<br />
-62-
WORK SHEET
= (r- 1)(c- 1)<br />
df<br />
(2- )(2 )<br />
=<br />
=1<br />
()()<br />
a table of critical values for X 2 for 1 degree of freedom, the<br />
Using<br />
chi-square value for sgnficance is 3.84 at the .05 probability<br />
necessary<br />
level. Thus, the X 2 of .41 n our example is not significant.<br />
calculations for the second serw'ce category "Business School or<br />
The<br />
are done in the same manner:<br />
College"<br />
SEX<br />
Male<br />
Female<br />
A)<br />
'C)'<br />
Business School or College<br />
NO<br />
130<br />
Yes<br />
180 70<br />
Ce I l 0 E O-E (O-E)<br />
A 130 100.80 +29.2 852.64<br />
B 10 39.20 -29.2 852.64<br />
C 50 79.20 -29.3 852.64<br />
D 60 30.80 +29.2 852.64<br />
-60-<br />
X 2<br />
(O-E)<br />
E<br />
=<br />
140<br />
110<br />
250<br />
8.46<br />
21.75 ÷ Major contributor<br />
10.77<br />
27.68 ÷ Major contributor<br />
68.66
WORK SHEET
3_. Calculate the expected frequencies for each cell. This is<br />
Step<br />
using the following formula-<br />
accomplished<br />
Total x Column Total<br />
Row<br />
Grand Total<br />
E<br />
Example. Calculate the expected frequency for Cell A.<br />
x 85<br />
140<br />
250<br />
E<br />
the expected frequency for each cell, the Diagnostic &<br />
Calculating<br />
table would look like this-<br />
Evaluation<br />
A)<br />
c)<br />
Diagnostic & Evaluation<br />
No<br />
50<br />
47.60<br />
the calculations necessary for the chl-square formula, the<br />
Performing<br />
can be shown in the following way-<br />
results<br />
I 1 0 E O-E (O-E) 2 (O-E) 2<br />
Ce<br />
E<br />
A 50 47.60 +2.40 5.76 .12<br />
B 90 92.40 -2.40 5.76 .06<br />
C 35 37.40 -2.40 5.76 .15<br />
D 75 72.60 +2.40 5.76 .08<br />
-58-<br />
Yes<br />
90<br />
X 2 .41
WORK SHEET
dlrects us to sum the resultant calculation for<br />
where.<br />
category over all categories (/
WORK SHEET
SRS-RSA.300<br />
Form<br />
7-74<br />
Revtsed<br />
Agency Code<br />
Number<br />
Case<br />
Last Name<br />
A<br />
37<br />
th, Sml<br />
Street and Number<br />
Address<br />
Wr ,qht<br />
I992<br />
ty County<br />
C,<br />
Mes a<br />
SUc Sac No<br />
A<br />
SSDI Status. 0<br />
B.<br />
OF HEALTH, EDUCATION, AND WELFARE<br />
DEPARTMENT<br />
and Rehabl,tation 5erv,ce<br />
ocmi<br />
Center for Sac,el Stat, shcs<br />
Natmnel<br />
D C 20201<br />
Washington,<br />
SERVICE REPORT STATE.FEDERAL PROGRAM OF VOCATIONAL REHABILITATION<br />
CASE<br />
(TO 8E RECORDED AT TIME OF FIRST REFERRAL)<br />
PART<br />
371-00-3406<br />
SSl Status<br />
C.<br />
eo,.. 'i<br />
D.<br />
Date Referral<br />
E.<br />
Completed<br />
Process<br />
10-1-71<br />
Months ,n Statuses 00-02 2<br />
F<br />
Spomsh Surname<br />
G<br />
Outcome of Referral Process<br />
H<br />
ACCEPTED Reason_.._.<br />
NOT<br />
F-] from (IX)), 2 !"1 from (02)<br />
FOR<br />
ACCEPTED<br />
[] Extended Evaluat,on (06)<br />
04<br />
05VR Surv,ces (10)<br />
Data of Closure<br />
A<br />
Referred by Sac Sec Adm,n<br />
B<br />
Suctal Security C] at m Type<br />
C<br />
Federal Special Program Identtf, cat,on*<br />
D<br />
TF VET MAW PO WIN [SEC4<br />
None<br />
X<br />
001 002 '004 010 020 040 '100<br />
000<br />
Cost for Case Services<br />
E<br />
All Serv, cos- Total<br />
Rehab, l,tat,on Fac, ltJes Total<br />
2.<br />
Socal Security Trust Funds Total<br />
3<br />
First Name Imtial C. Referral Date 8--20-7. D<br />
Mary R.<br />
JZ,p Code<br />
(code*)<br />
182222<br />
71<br />
E Ag. 3_.!_7 dab 8-29-33,, F<br />
G D,sab, hty as Reported (descrmbe)<br />
a Mes<br />
2 (TO BE RECORDED AT COMPLETION OF REFERRAL PROCESS)<br />
PART<br />
0<br />
9-29-73<br />
2 F"]Yes<br />
k-No,<br />
4. Supplemental security Income Funds Total<br />
Date Ext Eval Completed (,f appl,coble)<br />
F<br />
SSDI Status<br />
G<br />
IH<br />
SISI Status<br />
CLOSED FROM STATUS 00, COMPLETE<br />
IF<br />
3 PART<br />
CLOSED FROM STATUS 02, COMPLETE<br />
IF<br />
THRU 2T AND PART 3ON<br />
ITEMS21<br />
2 PAGE<br />
ACCEPTED FOR EXTENDED EVALUATION<br />
IF<br />
VR SERVICES COMPLETE ITEMS 2.1<br />
OR<br />
2.T DO NOT COMPLETE PART 3<br />
THRU<br />
THIS TIME,<br />
AT<br />
sabhng Condit,ons (descr,be)<br />
D,<br />
Orthopedic<br />
Molar<br />
mpal<br />
upper llmb<br />
Secondary<br />
2<br />
appl cable<br />
Not<br />
rment,<br />
Code<br />
C,d,<br />
340<br />
999<br />
PART 3 (TO BE RECORDED AT TIME OF CLOSURE)<br />
SF<br />
SD<br />
2e0 4o0<br />
(Dollars)<br />
1474<br />
status 1<br />
o,k<br />
eekly Earn,ngs $ 40<br />
J<br />
ubi,c Ass,stance I Type 0<br />
K<br />
SSI) |Monthly Amount $<br />
(tnclud,ng<br />
qccupot, on (.tie) ervce food shop L.<br />
Code<br />
Number of Months on Agency Rolls<br />
M.<br />
Extended Evaluahon (Status 06)<br />
In<br />
From A©cptance to Closure (Statuses 10-24) 2].<br />
2<br />
In Tm,nmg (Status 18)<br />
3.<br />
4 Ready For or In Employment (Statuses 20-22)<br />
*These ,terns ore to be coded<br />
STATE AGENCY COPY<br />
3118<br />
42<br />
ct Code<br />
Dstr,<br />
lJm<br />
Referral Source<br />
Approved<br />
Form<br />
t¢0. 83-R0040<br />
OMB<br />
Sex [] Male, 2 [] Femai-<br />
Cod,<br />
699<br />
Prev,ous Closure wmthn 36 Months<br />
J<br />
C-Jl, Yes-Outcome Rahab. I- 2<br />
No<br />
Rehab [--3<br />
Not<br />
Yes-Months Stnce Last Closure.___._._<br />
If<br />
Marital Status<br />
K<br />
Number of Dependents<br />
L<br />
Total Number ,n Family<br />
M.<br />
Highest Grade Completed<br />
N<br />
Work Status<br />
0<br />
Weekly Eam:ngs<br />
P<br />
Total Monthly Family Income<br />
Q<br />
earmngs)<br />
(,ncludng<br />
Pubhc ( Type<br />
R / Me Amt $<br />
Assistance<br />
SSI) Time on P A<br />
(,ncluding<br />
Primary Source at Support<br />
S<br />
T Type of Institutmn<br />
Outcome of Extended Evaluation or VR 5erv,ces<br />
N<br />
Closed from Extended Eveluahon (Status 08) Reason<br />
[<br />
Closed Rehabl,tated (status 26)<br />
2<br />
[] Closed Not Rehab,htoted (status 28) Reason<br />
3<br />
[] Closed Not Rahab, htated (Status 30) Reason<br />
4<br />
Serwces Provtded<br />
O<br />
of Service Provtded<br />
Type<br />
Arranged for by Agency<br />
or<br />
'D;ag'nst;c a.d 'Evaiuot,0n<br />
10<br />
Restorahon (Phys,cai or Mental)<br />
11<br />
12 T<br />
1611<br />
18<br />
G<br />
or Umvers,ty<br />
College<br />
cedem',c' (EIe. or MS)<br />
Other<br />
School or College<br />
Busme.ss<br />
School<br />
Vocatmnal<br />
On-the-Job<br />
& Voc Adlustment<br />
Personal<br />
'Mscellaneous<br />
Mm ntanance<br />
19<br />
Other Services<br />
20<br />
21 Services to Other Family Members<br />
P 5ate Agency Specml Program Identsf, cahon*<br />
None<br />
X<br />
Cost<br />
Status<br />
000 001 002 1,004 010 020 040 100 '200<br />
9-30-73<br />
of Report Counselor Sgrrature and<br />
Date<br />
i<br />
5<br />
12<br />
8<br />
000<br />
000<br />
serv, ca<br />
No<br />
th agen.<br />
W,<br />
only<br />
cost<br />
W,thout<br />
cost<br />
agency<br />
only<br />
and<br />
W,th<br />
thout<br />
w<br />
agency c<br />
SUM<br />
91<br />
Code<br />
L
WORK SHEET
B. Measurement Approach. State Agency's Case Service Report (R-300) data<br />
this example, the evaluator utilized case information from the<br />
For<br />
category entitled "Type of Service Provided or Arranged for by the<br />
R-300<br />
The lst of var'ous categories can be found on the R-300, Part 3,<br />
Agency."<br />
O, 10-21. (See sample SRS-RSA-300 form on next page.)<br />
Section<br />
C. Statlstical Procedures<br />
1. Tabulate frequencles for the various categories according<br />
Step<br />
whether services were provided or were not provided for men and for<br />
to<br />
women. A convenl'ent way to record results is as follows-<br />
of service provided/arranged Male (N=140) Female (N=110)<br />
T_<br />
b_z th__e agency No Yes No Yes<br />
fo___r<br />
1. Diagnostic & Evaluation 50 90 35 75<br />
2. Buslness School or College 125 15 70 40<br />
Two service categorles are included for demonstration purposes.<br />
2_. Utilize the tabulated frequencles n the chi-square formula,<br />
Step<br />
is as follows1:<br />
whlch<br />
i=1<br />
E<br />
degrees of freedom<br />
df = (r- 1)(c- 1)<br />
r number of rows<br />
where<br />
c number of columns<br />
and<br />
In the special case of a 2 x 2 table, it is possible to cal-<br />
IZNote<br />
the X 2 from the raw data wthout calculating the expected values.<br />
culate<br />
formula for thls purpose offered by Ferguson, G. A. Statstlcal analy-<br />
A<br />
n psychology and education. New York McGraw-Hill, 1966, p 204 is-<br />
sis<br />
BC) 2<br />
Xe N(AD<br />
(A + B)(C + D)(A + C)(B + D)<br />
=<br />
in a 2 x 2 table if any of the expected frequencies are less<br />
Also,<br />
5, a Yate's correction for continuity may be applied. Ths results<br />
than<br />
a slightly more conservative X 2 value. In the case of very small ex-<br />
in<br />
frequencies, a FCsher exact probability test may be used (Ferguson,<br />
pected<br />
1966, pp. 206-210).<br />
-52-
I Sample Evaluation Objectlve<br />
determine differences for male and female clients in servlces pro-<br />
To<br />
or arranged for by the agency during the rehabilitation process.<br />
vided<br />
II Criteria<br />
of clients (by sex) who are provided services in any given<br />
Number<br />
category<br />
service<br />
A. General Procedure<br />
Ill Methodology<br />
Ex post facto study design XO<br />
Steps.<br />
X : service/treatment<br />
where<br />
0 : observation/measurement<br />
and<br />
Identify clients who bare been provided services and have<br />
1.<br />
closed during two fiscal years. Divide them nto two groups,<br />
been<br />
male and female.<br />
From the information relatlng to servlces contained on the<br />
2.<br />
agency's case service report (R-300), determine if services<br />
State<br />
were provided or not.<br />
3. Tabulate results for each group,<br />
Conduct necessary statistics to make results meaningful<br />
possi lity l's a chi-square test for two independent samples I°.<br />
b4<br />
D. T. & Stanley, J. C. Experimental and quas-experimental<br />
9Campbell,<br />
for research. Chicago- Rand McNally, 1966.<br />
designs<br />
S. Nonparametric statistics for the behavioral sciences.<br />
10Siegel,<br />
York- McGraw-Hill, 1956, pp 104-111.<br />
New<br />
-51-<br />
One
EXAMPLE 4<br />
A. General Procedure: Ex post facto study deslgn.<br />
Measurement Approach: Follow-up survey of rehab-<br />
B.<br />
cases.<br />
±ltated<br />
Statistical Procedures: Product-moment correlation<br />
C.<br />
coefflc±ents.
EXAMPLE 4<br />
A. General Procedure: Ex post facto study deslgn.<br />
Measurement Approach: Follow-up survey of rehab-<br />
B.<br />
cases.<br />
llltated<br />
Statlstcal Procedures: Product-moment correlation<br />
C.<br />
coefficients.
I Sample Evaluation Objective<br />
determine the relatlonships among some client and servlce varia-<br />
To<br />
(age, number of counselor contacts, months of employment, and weekly<br />
bles<br />
for clients closed as rehabilitated (Status 26) for one fiscal<br />
Income)<br />
year.<br />
I I Criterl a<br />
of relatlonship between varlables for clients successfully<br />
Degree<br />
I itated.<br />
rehabi<br />
A. General Procedure 12<br />
I I I Methodology<br />
Ex post facto study deslgn XO<br />
Steps.<br />
X servlce/treatment<br />
where<br />
0 observation/measurement<br />
and<br />
Identify clients who have been closed rehablltated (Status<br />
1.<br />
during one fiscal year.<br />
26)<br />
Develop a survey instrument to obtain the needed data con-<br />
2.<br />
variables you wish to investigate.<br />
cernlng<br />
3. Conduct survey by mat1, phone, or personal contact.<br />
D. T. & Stanley, J. C. E..xperimental and quasl-experlmental<br />
12Campbell,<br />
for research Chlcago: Rand McNally, 1966.<br />
designs<br />
-65-
Tabulate results and do necessary statstlcs to make the<br />
4.<br />
meanlngful. One ossib111ty Is to conduct product-moment<br />
results<br />
correlatlon coefficients I<br />
B. Measurement Approach Fol low-up survey of rehab 11 tared cases<br />
way to obtain the needed nformaton is to adminlster a question-<br />
One<br />
to former clients relatlve to their present employment and income.<br />
nacre<br />
Some questlons that might be asked are:<br />
How many months have you been employed snce rehablitatlon<br />
1.<br />
were terminated<br />
services<br />
2. What is yur present weekly ncome in dollars?<br />
addltional method of obtaining information is from the former cli-<br />
An<br />
case record. Some questions for which data would be available are:<br />
ent's<br />
Age of client at closure .<br />
I.<br />
Number of contacts made by the counselor durlng the rehabl-<br />
2.<br />
tatlon process )<br />
C. Statistical Procedures<br />
1_. Tabulate the obtained information for each followed-up cli<br />
Step<br />
For example, if data was collected for 5 clients, it would be pre-<br />
ent.<br />
sented n the following format:<br />
of #<br />
Months of Weekly<br />
Counselor<br />
Cllent Age Contacts EmRloymept Income<br />
i 23 3 I 35.00<br />
2 29 8 6 60.00<br />
3 33 10 15 75.00<br />
4 42 7 25 80.00<br />
5 53 9 18 55.00<br />
G. A. Statlstcal analysls In psychology and education.<br />
13Ferguson,<br />
York McGraw-H II, 1966, pp. 105-115.<br />
New<br />
-66-
WORK SHEET
2. Calculate the correlation coefficient 14 for each combination<br />
Step<br />
variables (age, number of counselor contacts, months of employment, and<br />
of<br />
income) to determine degree of relationships. For example, to de-<br />
weekly<br />
the relationsh1"ps between age and @ncome, the follow'ng procedure<br />
termine<br />
could be util ized.<br />
Age<br />
Client<br />
X<br />
Example i<br />
of the Correlation Coefficient<br />
Calculatlon<br />
Age and Weekly Income<br />
for<br />
Income<br />
Weekl},<br />
X 2 Y2 XY<br />
Y<br />
23 35. O0 529 1225. O0 805<br />
29 60.00 841 3600.00 1740<br />
33 75. O0 1089 5625.00 2475<br />
42 80. O0 1764 6400. O0 3360<br />
53 55. O0 2809 3025. O0 2915<br />
zX 180 7.X 305.00 7.X 7032 .Y 19875.00 XY 11295<br />
this instructs us to do is to identify age as the X variable<br />
What<br />
income as the Y variable. You then square each X value and enter the<br />
and<br />
under X . The same procedure is followed for the Y variable. To<br />
result<br />
XY we simply multlply each X by its corresponding Y value to obtain<br />
obtain<br />
XY. After you have done this, then sum each column.<br />
Utilize the previously calculated values to calculate the<br />
3_.<br />
correlation coefficients (r). The formula is:<br />
product-moment<br />
zXzY<br />
Ny.XY-<br />
=V f lNzX (zx)2l INzY (zY) I<br />
r<br />
In this example X represents the square of the value of X;<br />
14Note:<br />
chi-square.<br />
not<br />
-68-
WORK SHEET
where N total number of cases<br />
.XY : sum of the products of X tlmes Y<br />
sX sum of X values<br />
sY sum of Y values<br />
sX 2 sum of each X value squared<br />
sY2 sum of each Y value squared<br />
you are ready to substitute the prevlous obtained data into the<br />
Now<br />
a-<br />
formul<br />
180(305)<br />
..5(11295)<br />
=/(5 r 7032 1802 (5 x 19875 3052<br />
x<br />
54900<br />
_56475<br />
32400) (99375 93025)<br />
=/(35160-<br />
54900<br />
:(2760)<br />
56475<br />
(6350)<br />
1575<br />
17526000<br />
:/<br />
1575<br />
4186.41<br />
determine the significance of r, one can use a table of cr%tical<br />
To<br />
found n statistics textbooks (a sample 1"s glven In Table 2 of the<br />
values<br />
To use the table, determine the degrees of freedom (df)<br />
Introduction).<br />
find the r value wh#ch Cs needed for sign1"ficance at the .05 probabil-<br />
and<br />
level. In th's case df N 2 or 5 2 3. For 3 df a critical<br />
ity<br />
of .88 or higher is necessary for a sgn1"ficant correlation. Thus,<br />
value<br />
.38 is not a statist'cally sign'f1"cant r at the .05 probability level.<br />
-70-
WORK SHEET
Cllent<br />
1<br />
2<br />
3<br />
4<br />
5<br />
of<br />
Months<br />
I oymen t<br />
Emp<br />
X<br />
1<br />
6<br />
15<br />
25<br />
18<br />
zX 65<br />
2<br />
Example<br />
of Correlation Coefficient for<br />
Calculation<br />
Months of Employment and Weekly Income<br />
Weekly<br />
Income<br />
y X 2 y2<br />
35.00 1 1225.00<br />
60.00 36 3600.00<br />
75.00 225 5625.00<br />
80.00 625 6400.00<br />
55.00 324 3025.00<br />
.Y = 305.00 zX 2<br />
sXzY<br />
NsXY<br />
:/INzX 2 (.x)l IN.Y 2 (sY)21<br />
r<br />
) 65(305)<br />
514510<br />
x 1211 652 (5 x 19875 3052<br />
"h/(5<br />
19825<br />
22550-<br />
(6055 4225) (99375 93025)<br />
"<br />
2725<br />
(1830) (6350)<br />
=,-<br />
2725<br />
11620500<br />
2725<br />
3408.89<br />
-72-<br />
XY<br />
35<br />
360<br />
1125<br />
2000<br />
990<br />
= 1211 sY 2 = 19875.00 .XY = 4510
3 df the r must equal or exceed 88 to be signlficant at the .05<br />
For<br />
level. Thus, there is not a statistically significant corre-<br />
probability<br />
lation between months of employment and weekly income.<br />
procedure Is followed for all comparisons. The addtlonal cor-<br />
This<br />
coefficients were calculated and results are presented in the<br />
relation<br />
correlation matrix table. This table allows us to conveniently<br />
following<br />
the inter-correlation of each of the four variables.<br />
examine<br />
Age<br />
of Counselor<br />
#<br />
Contacts<br />
of Em-<br />
Months<br />
I oyment<br />
p<br />
Weekly Income<br />
degrees of freedom<br />
df:N-2<br />
=5-2<br />
=3<br />
Correlation Matrix Table<br />
of #<br />
Months of Weekly<br />
Counselor<br />
Contacts Employment Income<br />
.55 .79 .38<br />
.55 69<br />
a correlation coefficlent table (found in statistics textbooks)<br />
Using<br />
3 degrees of freedom, the necessary critical value of .88 was not<br />
for<br />
for a correlation coefficient significantly different from a zero<br />
reached<br />
at the .05 probability level for any of the relationships.<br />
correlate'on<br />
IV Interpretation<br />
can be observed in the correlation matrix table, the highest cor-<br />
As<br />
found was between months of employment and weekly 1"ncome (.80).<br />
relation<br />
-74-<br />
.80
WORK SHEET
indicates a relatively high relationship between these two variables<br />
This<br />
O0 would be a perfect correlation) However, because of the small num-<br />
(1<br />
of subjects (5) for thls analysis, it as not high enough to be sig-<br />
ber<br />
different than a zero correlation. (.00 or no relationship at<br />
nificantly<br />
at the .05 probabllity level. Thus, we could not say that this rela-<br />
all)<br />
is significant and would occur by chance only 5 times out of 100<br />
tionship<br />
none of the other correlation coefficients were statistically<br />
Likewise,<br />
significant<br />
a correlation is not significantly different from zero, then sta-<br />
If<br />
t must be assumed that either no relatonshlp exists between<br />
tlstically<br />
or that the sample was not large enough to detect the difference.<br />
variables<br />
our example with 5 cases, a correlation coefficient of .88 is needed to<br />
In<br />
statistically significant. On the other hand, if we had 30 cases, a<br />
be<br />
of .36 would be significant at the .05 level, and with 100<br />
coefficient<br />
a coefficient of 20 would be significant. Whether a correlation<br />
cases<br />
statistically significant or not, the size of the correlation suggests<br />
is<br />
degree of relationship. Thus, correlations may also be viewed in terms<br />
the<br />
their practical meaning Borg & Merideth (1968) offer some general<br />
of<br />
for interpreting simple correlation coefficients assuming a<br />
guidelines<br />
of 100 or more subjects. According to Borg & Meredith Is, correla-<br />
sample<br />
ranging from .20 to 35 show a very slight relationship between<br />
tions<br />
and may have very limited meaning. Correlations from .35 to 65<br />
varlables<br />
crude predictions for groups but are of limlted value for ndl-<br />
provide<br />
dual predictions Correlations ranging from .65 to .85 are good for<br />
w<br />
group preductions, e g predicting the proportion of successful<br />
making<br />
Correlations over 85 indicate a very good relatonshlp between<br />
clients<br />
two variables and are useful for either groups or ndlwdual predictions<br />
the example the correlation coefficients between months of employ-<br />
In<br />
and age (.79), months of employment and weekly ncome (.80), and num-<br />
ment<br />
of counselor contacts and weekly ncome (.69) are good for making group<br />
ber<br />
The correlations between the number of counselor contacts<br />
predictions.<br />
age (55), number of counselor contacts and months of employment (.55),<br />
and<br />
weekly income and age (38) provide only crude predictions for groups.<br />
and<br />
of the correlation coefficients In our example were posltve.<br />
All<br />
suggests, for example, that the higher the months of employment of<br />
This<br />
clients, the higher the weekly income, the hgher the number of<br />
former<br />
contacts, the higher the weekly income of former clients, etc<br />
counselor<br />
the correlatlons were negative, then they would suggest that the hgher<br />
If<br />
months of employment, the lower the weekly Income.<br />
the<br />
a correlatlon coefficient does not reflect a cause-effect rela-<br />
Though<br />
it does suggest the possibillty of one. It can be used as an<br />
tionship,<br />
ndicator for improving service dellvery.<br />
W. R. & Merideth, D. G. Educational research- An introduc-<br />
ISBorg,<br />
(2nd Ed.) New York David McKay Company, Inc., 1963, pp. 357-360.<br />
tion.<br />
-76-
EXAMPLE 5<br />
A. General Procedure: Ex post facto study deslgn.<br />
B. Measurement Approach: Case Dlff!culty Index.<br />
C. Statistical Procedures: One-way analyss of variance.
EXAMPLE 5<br />
A. General Procedure: Ex post facto study deslgn.<br />
B. Measurement Approach: Case Dlffculty Index.<br />
C. Statistical Procedures: One-way analyss of varlance.
I Sample Evaluation Objectlve<br />
compare effectiveness of rehabilitatlon counselor performance in<br />
To<br />
rehabil itation offices based on the number and case dCfficulty of<br />
three<br />
successful closures.<br />
II Criteria<br />
effectiveness of rehabilitation counselors-in three<br />
Servcedelivery<br />
as determined by Case Di.ffculty Index scores for rehabilitated<br />
offices<br />
cases (Status 26).<br />
A. General Procedurel 6<br />
III Methodology<br />
Ex post facto study design XO<br />
Steps.<br />
X servlce/treatment<br />
where<br />
0 = observation/measurement<br />
and<br />
Identify all clients closed as rehab11tated (Status 26) for<br />
1.<br />
in a given office.<br />
counselors<br />
Determine the service difficulty of each case by uslng the<br />
2.<br />
Difficulty Index. (CDI).<br />
Case<br />
Determine the Total Weighted Closure Index (TWCI) by adding<br />
3.<br />
CDI values for a gven counselor.<br />
all<br />
Divlde the TWCI by the number of successful closures for a<br />
4<br />
counselor to obtain the Average Weighted Closure Index (AWCl).<br />
given<br />
D. T. & Stanley, J. C. Experimental and quas'-experlmental<br />
16Cmnpbell,<br />
for resear.c.h. Chlcago Rand McNally, 1966.<br />
desl.qns<br />
-79-
Conduct necessary statlstlcs to make the results meanlngful<br />
5<br />
possibility Is a one-way analyss of variance test zT.<br />
One<br />
B. Measurement Approach Case Dfficult.y Inde___x 18<br />
Case Dffculty Index (CDI) was deslgned to expand and complement<br />
The<br />
number of successful closures as a measure of counselor performance<br />
the<br />
1972). In doing so, the CDI contains two additional dlmensons<br />
(Sermon,<br />
counselor performance. The first of these is the Total Weighted Clo-<br />
of<br />
Index (TWCl). Sermon (1972) defines the TWCl as a "quantity-tempered-<br />
sure<br />
measure of a counselor's performance and represents total<br />
wth-dfflculty"<br />
counselor effort. The second index is the Average Weighted Closure Index<br />
This reflects the average difficulty of all cases successfully<br />
(AWCl).<br />
by a counselor regardless of the actual number of cases closed<br />
closed<br />
index Is Intended to provide a quality dimension for counselor per-<br />
Thls<br />
The Cas___e Diffcu!t.y Index, then, allows us to look at counse-<br />
formance.<br />
performance n terms of cases successfully closed (quantity), the<br />
lor<br />
counselor effort represente by these closures (work effort), and<br />
total<br />
average difficulty of successfully closing these kinds of cases<br />
the<br />
(quality).<br />
Case Difficulty Index formula<br />
CDI (1 000 P) x 10<br />
CDI is the Cas___e Dfficulty Inde_.._x for<br />
where<br />
given impairment (represented<br />
a<br />
by 4).<br />
is the rehablltatlon success<br />
P<br />
experienced nationally<br />
rate<br />
way to express the same thing is CDI O x 10, where O is<br />
Another<br />
rehabilitation failure rate experienced nationally with an impairment<br />
the<br />
G A. Statlstlcal analysls in psychology and education<br />
17Ferguson,<br />
York McGraw-Hill, 1966, pp. New 21-297<br />
D. T. The dfficult>, index an expanded measure of coun-<br />
18Sermon,<br />
performance. St.---Paul" Minnesota Division of Vocational Rehablli-<br />
selor<br />
Research Monograph 1, 1972.<br />
tatlon,<br />
-80-
WORK SHEET
CDI; is then determined by taking the percentage of fallures<br />
category.<br />
28/30) for each impairment and multlplyng it by 10.<br />
(Status<br />
CDI indices (Table 3) on the next 4 pages are taken from natlonal<br />
The<br />
for FY 68, 69, and 70 as reported by the Statistical Analysis<br />
statistics<br />
Systems Branch of the Rehabilitation Services Administration in the<br />
and<br />
S. Department of Health, Education, and Welfare. For local comparisons<br />
U.<br />
agency could develop a similar scale based on local norms within a<br />
an<br />
or region.<br />
State<br />
C. Statistical Procedures<br />
1_. Convert the number of rehabilitated case closures to a Case<br />
Step<br />
Inde____x Score (See CD__.I table). For example<br />
Difficulty<br />
Counselor #1<br />
Disability<br />
hearing impair-<br />
Other<br />
ment<br />
220-229)<br />
(Disability Code<br />
and unspeci-<br />
Benlgn<br />
neoplasms (Dis-<br />
fied<br />
11 ty Code 609)<br />
ab<br />
retardat on-<br />
Mental<br />
(Disability<br />
m11d<br />
Code 530)<br />
of suc-<br />
No.<br />
closures CDI<br />
cessful<br />
.94<br />
.90<br />
1.96<br />
TWCI<br />
x .94 = 4.70<br />
5<br />
4 x .90 = 3.60<br />
1 x 1.96 = 1.96<br />
TWCI<br />
:<br />
AWCI<br />
10 FTTI<br />
10.26<br />
ths counselor has 10 successfully rehabilitated cases, a Total<br />
Thus,<br />
Closure Index of 10.26 and an Average Weighted Closure Index of<br />
Weighted<br />
1.03.<br />
-82-
WORK SHEET
IMPAIRMENTS<br />
VISUAL<br />
Both Eyes (No Light<br />
Blindness<br />
rcepti on<br />
Pe<br />
Both Eyes (Light<br />
Blindness<br />
Perception)<br />
i ndness One Eye Other Eye<br />
Bl<br />
Defective<br />
One Eye Other Eye<br />
Bllndness<br />
Good<br />
Visual Impal rments<br />
Other<br />
IMPAIRMENTS<br />
HEARING<br />
Unable to talk<br />
Deafness-<br />
-Able to talk<br />
Deafness<br />
Hearing Impairments<br />
Other<br />
IMPAIRMENTS<br />
ORTHOPEDIC<br />
3 or more limbs<br />
Involves<br />
re Body<br />
Ent<br />
1 upper and 1 lower<br />
Involves<br />
imb<br />
1<br />
1 or both upper llmbs<br />
Involves<br />
1 or both lower limbs<br />
Involves<br />
Ill Defined Impairments<br />
Other<br />
OR AMPUTATION OF MAJOR AND<br />
ABSENCE<br />
MEMBERS<br />
MINOR<br />
of at least 1 upper and 1<br />
Loss<br />
major extremity<br />
lower<br />
of both major upper<br />
Loss<br />
ti es<br />
extreml<br />
of 1 major upper extremity<br />
Loss<br />
of 1 or both major lower<br />
Loss<br />
t es<br />
extreml<br />
of other unspecified parts<br />
Loss<br />
TABLE 3<br />
CASE DIFFICULTY IBDEX<br />
FY FY FY<br />
Disability<br />
1968 1969 1970<br />
Code<br />
100-109 1.96 1.91 1.97<br />
110-119 1.70 1.84 1.75<br />
120-129 1.54 1.76 1.69<br />
1.55 1.59 1.37<br />
130-139<br />
1.44 1.41 1.21<br />
140-149<br />
I. 80 I. 86 1.82<br />
200-209<br />
1.10 1.18 1.18<br />
210-219<br />
220-229 .98 .99 .85<br />
300-319 2.79 2.99 2.75<br />
2.76 2.96 2.75<br />
320-339<br />
2.17 2.26 2.17<br />
340-359<br />
2.11 2.26 2.06<br />
360-379<br />
2.71 2.74 2.60<br />
380-399<br />
400-409 1.61 1.77 1.84<br />
1.50 1.68 2.15<br />
410-419<br />
1.36 1.67 1.51<br />
420-429<br />
1.55 1.68 1.51<br />
430-439<br />
2.12 2.49 1.94<br />
440-449<br />
-84-<br />
Dl ffl cul ty<br />
Case<br />
(Composite<br />
Index<br />
for the 3 years)<br />
1.94<br />
1.77<br />
1.67<br />
1.50<br />
1.35<br />
1.83<br />
1.16<br />
.94<br />
2.85<br />
82 2<br />
2.20<br />
2.15<br />
2.68<br />
1.73<br />
1.79<br />
1.51<br />
1.58<br />
2.20
PSYCHONEUROTIC, AND PER-<br />
MENTAL,<br />
DISORDERS<br />
SONALITY<br />
c<br />
Psychotl<br />
c<br />
Psychoneurotl<br />
Alcoholism<br />
Addiction<br />
Drug<br />
Character, Personality, and<br />
Other<br />
oral Disorders<br />
Behavl<br />
Retardatl on Mlld<br />
Mental<br />
Retardation Moderate<br />
Mental<br />
Mental Retardation Severe<br />
DISABLING CONDITIONS<br />
OTHER<br />
NOT KNOWN OR NOT APPROPRIATE<br />
ETIOLOGY<br />
FROM NEOPLASMS<br />
RESULTING<br />
from Malignant<br />
Colostomes<br />
I asms<br />
Neop<br />
from Mal gnant<br />
Laryngectomles<br />
I asms<br />
Neop<br />
and Aleukema<br />
Leukemia<br />
Mali gnant Neoplasms<br />
Other<br />
and Unspecified Neoplasms<br />
Benign<br />
ENDOCRINE SYSTEM,<br />
ALLERGIC,<br />
AND NUTRITIONAL DISEASES<br />
METABOLIC<br />
Fever and Asthma<br />
Hay<br />
A1 lergies<br />
Other<br />
abetes Mel i tus<br />
D<br />
Endocrlne System Dsorders<br />
Other<br />
tam nos s and Other Metabol c<br />
Aw<br />
seases<br />
D1<br />
OF THE BLOOD AND BLOOD<br />
DISEASES<br />
ORGANS<br />
FORMING<br />
I i a<br />
Hemoph<br />
and Other Dseases<br />
Anemia<br />
SPECIFIED DISORDERS OF THE<br />
OTHER<br />
SYSTEM<br />
NERVOUS<br />
lepsy<br />
Ep<br />
DIsorders<br />
Other<br />
TABLE 3 (Continued)<br />
sabi ty FY FY FY<br />
Di<br />
1968 1969 1970<br />
Code<br />
3.37 3.58 3.23<br />
500<br />
2.41 2.55 2.23<br />
510<br />
3.55 4.04 3.81<br />
520<br />
5.69 4.94 4.89<br />
521<br />
3.33 3.47 3.20<br />
522<br />
2.07 1.96 1.90<br />
530<br />
2 31 2.22 2.24<br />
532<br />
2.88 2.71 2.61<br />
534<br />
600 2.77 2.14 2.74<br />
4.31 3.47 3.25<br />
601<br />
5.56<br />
602<br />
2.30<br />
605<br />
.94 94 .82<br />
609<br />
2.03 2.13 1.98<br />
610<br />
1.84 1.91 1.65<br />
611<br />
2.21 2.20 1.94<br />
614<br />
1 46 1.47 1.18<br />
615<br />
619 3.19 2.33 2.18<br />
2.43 2.27 2.36<br />
620<br />
2.33 2.95 2.52<br />
629<br />
2.76 2.76<br />
630<br />
639<br />
-85-<br />
2.62<br />
D ff cul ty<br />
Cas____e<br />
(Compos te<br />
Index<br />
the 3 years)<br />
for<br />
3.40<br />
2.39<br />
3.84<br />
5.09<br />
3.33<br />
1.96<br />
2.26<br />
2.73<br />
2.48<br />
3 67<br />
.90<br />
2.05<br />
1.79<br />
11 2<br />
1.38<br />
2.54<br />
2.36<br />
2.62<br />
2.72
AND CIRCULATORY CONDITIONS<br />
CARDIAC<br />
Heart Disease<br />
Congenltal<br />
Fever and Chronic<br />
Rheumatic<br />
Heart Disease<br />
Rheumatic<br />
and Degenerative<br />
Arteriosclerotic<br />
DIsease<br />
Heart<br />
Diseases and Conditlons of<br />
Other<br />
Heart<br />
the<br />
Heart Disease<br />
Hypertensive<br />
Hypertensive Disease<br />
Other<br />
Veins and Hemorrhoids<br />
Varicose<br />
Conditions of the<br />
Other<br />
Circulatory System<br />
DISEASES<br />
RESPIRATORY<br />
ratory Tubercul os s<br />
Resp<br />
Emphysema<br />
and Aspestosis<br />
Pneumoconoiosis<br />
ectasl s<br />
Bronchi<br />
Bronchltis and Snusltus<br />
Chronic<br />
Dseases of Respiratory<br />
Other<br />
System<br />
OF THE DIGESTIVE SYSTEM<br />
DISORDERS<br />
of the Teeth and<br />
Conditions<br />
Structures<br />
Supporting<br />
of Stomach and Duodenum<br />
Ulcer<br />
Enteritis and Ulceratlve<br />
Chronic<br />
Colitis<br />
a Hernl<br />
other than from<br />
Colostomes<br />
gnant Neoplasms<br />
Mal<br />
Conditions of Digestlve<br />
Other<br />
Sys tern<br />
GENITO-URINARY SYSTEM CONDITIONS<br />
TABLE 3 (Continued)<br />
D ffl cul ty<br />
Case<br />
FY FY FY Index (compos te<br />
Disability<br />
Code 1968 1969 1970 for the 3 years)<br />
640 2.50 2.56 2.15 2 41<br />
641 2.68 2.75 2.40 2.62<br />
642 3.31 3.67 3.26 3.42<br />
2.97 2.93 3.01 2.97<br />
643<br />
3.17 3.41 3.31 3.30<br />
644<br />
2.56 2.40 2.53 2 51<br />
645<br />
1.38 1.25 1.16 1.27<br />
646<br />
649 2.82 2.74 2.48 2.68<br />
2.97 3.31 2.94 3.08<br />
650<br />
3.58 3.88 3.66 3.71<br />
651<br />
4.82 2.80 2.73 3.88<br />
652<br />
2.68 3.00 2.54 2.76<br />
653<br />
654 2.74 2.66 2.28 2.56<br />
659 2.60 2.41 2.64 2.55<br />
.73 .77 65 .71<br />
660<br />
1.81 1.71 1.58 1.70<br />
661<br />
1.26 2.40 2.20 2.28<br />
662<br />
1.05 1.14 .99 I. 06<br />
663<br />
664 1.87 1.96 1.49 1.76<br />
669 1.31 1.19 1.12 1.20<br />
670 .99 1.01 .92 .97<br />
-86-
IMPAIRMENTS<br />
SPEECH<br />
Palate and Harelip<br />
Cleft<br />
TABLE 3 (Contlnued)<br />
Dl ffl cul ty<br />
Case<br />
lty FY FY FY Index (Composite<br />
Dsabl<br />
1968 1969 1970 for the 3 years)<br />
Code<br />
Imperfections 680 1.32 1.72 1.39 1.47<br />
Speech<br />
and Stuttering 682 1.92 1.75 1.53 1.74<br />
Stammering<br />
from other than<br />
Laryngectomies<br />
i gnant Neoplasms 684 1.90 3.73 3.09 2 97<br />
Mal<br />
from Intercranial<br />
Aphasla<br />
Embolism or<br />
Hemorrhage,<br />
(Stroke) 685 3.50 3.2 2.78 3.16<br />
Thrombosis<br />
Speech Impairments 689 1.92 2.20 1.63 1.91<br />
Other<br />
DISEASES AND CONDITIONS<br />
DISABLING<br />
ELSEWHERE CLASSIFIED (N.E.C.)<br />
NOT<br />
Skin and Cellular Tssue 690 2.05 2.04 1.73 1.93<br />
Of<br />
(N.E.C.) 699 2.17 2.19 2.24 2.21<br />
Other<br />
from Duane T. Sermon, Minnesota D1wson of Vocatlonal Reha-<br />
Recelved<br />
St. Paul, Minnesota.<br />
b11taton,<br />
-87-
Counselor<br />
three counselors in each of three offices then, the CDI mlght<br />
For<br />
like this<br />
look<br />
Office #I<br />
Rehab<br />
TWCI AWCI<br />
Closures<br />
10 10.26 1.03<br />
9 11 36 1.26<br />
10 12.10 1.21<br />
29 33.72 1.17<br />
Office #2<br />
Rehab<br />
TWCI AWCI<br />
Closures<br />
5 23.76 4.75<br />
4 18.61 4.65<br />
5 19.00 3.80<br />
14 61.37 4.40<br />
Rehab.<br />
l os u res<br />
C<br />
9<br />
I0<br />
9<br />
28<br />
Office #3<br />
values 1.17, 4.40, and 1.31 represent AWCI means for the three<br />
The<br />
in each of the three offices.<br />
counselors<br />
2. Calculate F-test for a one-way ana]ysls of varlance. The<br />
Step<br />
tests the difference between the means of more than two groups. (In<br />
F<br />
TWCI AWCI<br />
12.00 1.33<br />
Ii.01 i. I0<br />
13.50 1.50<br />
36.51 1.31<br />
case the means of the AWCI for the three rehabilitation offices under<br />
this<br />
The F is represented by the ratlo of variance between groups<br />
evaluation)<br />
our example there are three groups) and the variance within groups.<br />
(in<br />
the variability between group means s large enough in comparison with<br />
If<br />
variability withln groups, a statistically significant F will result<br />
the<br />
we can infer that the groups are different The F-ratio is expressed<br />
and<br />
as<br />
follows-<br />
F =<br />
square between Is<br />
Mean<br />
square within<br />
Mean<br />
arrive at an F value, the following procedures should be followed<br />
To<br />
would look like ths for the AWCI comparisons.<br />
and<br />
#1 Office #2 Office #3<br />
Office<br />
X2 X3<br />
X1<br />
1.03 4.75 1.33<br />
1.26 4 65 1.10<br />
1.21 3.80 1.50<br />
G. A. Statis.tica.l analysis n psycholog}, and education<br />
19Ferguson,<br />
York- McGraw-Hill, i966, pp. 281-297.<br />
New<br />
-88-
WORK SHEET
Squaring each X value and surnaming the columns results In the following-<br />
#1 Office #2 Office #3<br />
Office<br />
XI 2 X2 X22 X3 X3<br />
Xz<br />
1.03 1.06 4.75 22.56 1.33 1.77<br />
1.26 1.59 4.65 21.62 1.10 1.21<br />
1.21 1.46 3.80 14.44 1.50 2.25<br />
3.50 4.11 13.20 58.62 3.93 5.23<br />
3 3 3<br />
n#<br />
3.50 13.20 3.93<br />
T#<br />
1.17 4.40 1.31<br />
[<br />
nj<br />
. Xi# 2 4.11 58.62 5 23<br />
T z<br />
4.08 58.08 5.15<br />
N=9<br />
T = 20.63<br />
T/N<br />
425.60<br />
9<br />
47.29<br />
nj<br />
. Xi# 67 96<br />
z<br />
=i i=I<br />
k<br />
j=l n<br />
where the number of counselors in each column (office)<br />
total number of counselors In the evaluation<br />
67.31<br />
total of the AWCI in each column (e.g. in office #1"<br />
T#<br />
1.03 + 1.26 + 1.21 or 3.50)<br />
T#<br />
total of T#'s for all three offlces (e.g. T 3.50 +<br />
T<br />
+ 3.93)<br />
13.20<br />
# T_#_ or the average of the AWCI's for each office (e.g.<br />
: :<br />
-90-
WORK SHEET
the previous calculated T value squared dlvlded by the<br />
T2/N<br />
number of counselors.<br />
total<br />
T2/N 20.63 x 20.63 425.60<br />
9 = 9 47 29)<br />
(e.g.<br />
the sum of all three sXij2 (e.g. 4.11 + 58 62 + 5 23)<br />
s.Xij2<br />
T<br />
2<br />
take the value of T# for each office and multlply it by<br />
-#--=<br />
its own value (square it). Then divide by n# for each<br />
n#<br />
zT_=<br />
col umn.<br />
(e.g. for office #1<br />
T 3.50 x 3.50 12.25 4.08)<br />
j 3 3<br />
the three values calculated for and add them<br />
take<br />
together<br />
3. Using the values calculated in step 2, you can now deter-<br />
St<br />
the sum of the squares for between, within, and the total. The for-<br />
mlne<br />
and calculation would look like this.<br />
mulas<br />
Sum of Squares Between<br />
Degrees of freedom k 1<br />
:2<br />
3-1<br />
-92-<br />
k<br />
s<br />
T#<br />
#:1 T 2<br />
67.31 47.29<br />
12o,021<br />
where- k number of groups
WORK SHEET
of Squares Within s s X{# 2<br />
Sum<br />
j=l<br />
Degrees of freedom N k<br />
Sum of Squares Total<br />
:9-3<br />
=6<br />
Degrees of freedom N I<br />
=9-1<br />
--8<br />
-94-<br />
= 67 96 67 31<br />
k<br />
.<br />
i=I<br />
#=1<br />
67.96 47.29<br />
k<br />
s<br />
#=1<br />
T# 2<br />
N : total number of<br />
wheresubjects<br />
k number of groups<br />
If the sum of squares between<br />
120.,671<br />
within are added together,<br />
and<br />
value will equal the sum<br />
thelr<br />
squares total (Sum of<br />
of<br />
Total Sum of Squares<br />
Squares<br />
Between + Sum of Squares W1thln)
WORK SHEET
4_. Fill in the Summary Table and do the necessary calculatlons<br />
Step<br />
obtain an F. The Summary Table would appears as follows before cal-<br />
to<br />
culations are conducted.<br />
Source of varlation<br />
Between<br />
Within<br />
Total<br />
Summary Table<br />
Degrees of freedom<br />
N-1<br />
Sum of squares Mean squares<br />
the prewously calculated nformation in the summary<br />
Substituting<br />
results in the followlng-<br />
table<br />
Source of varation<br />
Between<br />
Within<br />
Total<br />
*p < .05.<br />
Summary Table<br />
Degrees of freedom Sum of ,squares<br />
20.02<br />
.65<br />
20.67<br />
B<br />
SS<br />
k-1<br />
SS<br />
W<br />
N-/<<br />
Mean Squares,<br />
10.01<br />
obtained F value for our example Is 91.00. Using a table of<br />
The<br />
values for F tests (found in statistics books) for 2 and 6 degrees<br />
critical<br />
-96-<br />
.11<br />
F<br />
B MS<br />
MS<br />
W<br />
91.00"
WORK SHEET
,"reedom, the necessary F value is 5.14 at the .05 probab111ty level.<br />
of<br />
means that only 5 times out of 100 would we expect a F of 5.14 or<br />
This<br />
greater to happen by chance alone.<br />
now know from our significant F value that there are significant<br />
We<br />
between our three groups. However, we do not know where the<br />
dlfferences<br />
exists. To determine this we must conduct tests between each<br />
difference<br />
of groups. This, of course, would not be necessary if a significant<br />
pair<br />
F was not obtained.<br />
5_. Run a Scheff test Step 2° determine which mean(s) of the three<br />
to<br />
belng studied is significantly different from the others. (Again,<br />
offices<br />
is only conducted if the obtained F is significant. The significant<br />
this<br />
tells us there is a difference between the three off'ces but does not<br />
F<br />
us where iIs difference Is.) Comparisons for this example would be<br />
tell<br />
follows:<br />
as<br />
mean I compared to mean II<br />
mean I compared to mean III<br />
mean II compared to mean<br />
To make these comparlsons It is necessary to use the following formula-<br />
Sw2("i + 2)<br />
SW2 mean square wthin (taken from<br />
where:<br />
sumary table)<br />
the<br />
of subjects n each group<br />
number<br />
examinati on.<br />
under<br />
k total number of groups.<br />
G. A. Statlstcal analysls in psychology and education.<br />
2°Ferguson,<br />
York. McGraw-Hll, 1966, pp. 295-297.<br />
New<br />
-98-
WORK SHEET
are as follows for the comparison between the means of<br />
Calculations<br />
#1 and office #2"<br />
office<br />
F<br />
4.40) 2<br />
(1.17<br />
(3 + 3)/(3)(3)<br />
(.11)<br />
(-3.23)<br />
.66/9<br />
10.43<br />
.07<br />
the other two comparisons are conducted with the fol]owlng<br />
Similarly,<br />
results<br />
Comparl s on<br />
F_<br />
II 149.04<br />
I,<br />
I, III .28<br />
II, II! 136.40<br />
the determnaton of F values for the several comparisons,<br />
Following<br />
is necessary to consult a F table for the necessary crltical value re-<br />
it<br />
for significance at the .05 probabllty level for df /< 1 and<br />
qulred<br />
N k (2 and 6 in this example). For ou' example, a F of 5.14 is<br />
df<br />
necessary.<br />
calculate a new quantity F' which is /
WORK SHEET
compare the values of F and F' For any difference to be slg-<br />
Then,<br />
at the required .05 probability ievel, F must be equal to or<br />
nificant<br />
greater than F'.<br />
F_ F'<br />
Comparison<br />
II 149.04 10.28<br />
I<br />
I, III .28 10.28<br />
II, Ill 136.40 10.28<br />
F values for comparisons between off1"ce #I and office #2, and<br />
Te<br />
office #2 and office #3 are greater than the required F' value of<br />
for<br />
10.28<br />
IV Interpretation<br />
can be observed by the obtained F value of 91.00 for counselors<br />
As<br />
three rehabilitation offices, significance was reached at the .05 prob-<br />
in<br />
level because the obtained F value exceeded 5.14. This indicates<br />
abllity<br />
differences exist between the Average Weighted Closure Index means<br />
that<br />
the three offices. However, it does not tell us where these differ-<br />
of<br />
are. To obtain this information, the evaluator must conduct a<br />
ences<br />
test to compare means two at a time. The Scheffe tests resulted<br />
Scheff<br />
slgnificant differences between offices #1and #2, and between offices<br />
in<br />
and #3, but no significant differences between offl'ces #I and #3. The<br />
#2<br />
again, were as follows.<br />
means,<br />
#1 1.17<br />
Office<br />
#2 4.40<br />
Office<br />
Office #3 1.31<br />
office #2 counselors, as a group, were significantly more ef-<br />
Thus,<br />
than both offices #1 and #3. Office #2 was more effective based<br />
fective<br />
a combination of effectiveness measures including quantity, work<br />
upon<br />
and quality of rehabilitations, even though they had fewer reha-<br />
effort,<br />
(14) than either office #1 (29 rehabilitations) or office #3<br />
b11itations<br />
rehabilitations).<br />
(28<br />
-102-
PART TWO<br />
One Group Pretest Posttest Study Des±gn<br />
01 X 0 2<br />
where: X service/treatment<br />
observat±on/measurement<br />
01<br />
service<br />
before<br />
observatlon/measurement<br />
02<br />
service<br />
after<br />
design calls for a smngle group that ms studied prior<br />
This<br />
services and again after services. The "gain" that is ob-<br />
to<br />
from 0 to 0 is presumed to have been a result of<br />
served<br />
serv±ce (). 2<br />
the<br />
The evaluator identlfmes a group of sub3ects<br />
Method.<br />
they are to undergo a particular servlce/treatment.<br />
before<br />
measures this group (0 I) prior to service. The evaluator<br />
He<br />
measures this group after service (02). From the<br />
again<br />
of the change from 01 to 02 he then attempts to<br />
results<br />
the effect of the service.<br />
infer<br />
This deslgn ms an improvement over the ex<br />
Limitations.<br />
facto study design because ±t enables the evaluator<br />
post<br />
manipulate services, and it permlts comparing a group<br />
to<br />
itself. Its weakness is that it does not allow the<br />
with<br />
the opportunity to know or test the effect of<br />
evaluator<br />
mnfluences in addlton to the provided service<br />
possible<br />
e.g. the effect of pretesting, maturation, other<br />
(X),<br />
etc.<br />
events,
PART TWO<br />
One Group Pretest Posttest Study Deslgn<br />
X 0 where:<br />
0<br />
2<br />
1<br />
X<br />
servlce/treatment<br />
observatzon/measurement<br />
01<br />
service<br />
before<br />
0 2<br />
observatlon/measurement<br />
servlce<br />
after<br />
deslgn calls for a sngle group that s studied prior<br />
Thls<br />
services and again after servlces. The "galn" that s ob-<br />
to<br />
from 0 to 0 s presumed to have been a result of<br />
served<br />
service (). 2<br />
the<br />
The evaluator dentfes a group of sub3ects<br />
Method.<br />
they are to undergo a partlcular service/treatment.<br />
before<br />
measures ths group (0 I) prior to servlce. The evaluator<br />
He<br />
measures thls group after service (0) From the<br />
agaln<br />
of change from 01 to 0_, he thn attempts to<br />
results the<br />
the effect of the 2<br />
service.<br />
nfer •<br />
Ths deslgn s an mprovement over the ex<br />
Lmtatons.<br />
facto study deslgn because t enables the evaluator<br />
post<br />
manipulate services, and t permits comparing a group<br />
to<br />
tself. Its weakness s that t does not allow the<br />
wlth<br />
the opportunity to know or test the effect of<br />
evaluator<br />
influences n addition to the provided service<br />
possible<br />
e.g. the effect of pretestlng, maturation, other<br />
(X),<br />
etc.<br />
events,
EXAMPLE 6<br />
General Procedure: One group pretest posttest<br />
A.<br />
deslgn.<br />
study<br />
B. Measurement Approach: Goal Attainment Scaling.<br />
Statlst±cal Procedures: t-test for correlated<br />
C.<br />
samples.
EXAMPLE 6<br />
General Procedure: One group pretest posttest<br />
A.<br />
deslgn.<br />
study<br />
B. Measurement Approach:<br />
Statlstlcal Procedures:<br />
C.<br />
samples.<br />
Goal Attainment Scalng.<br />
t-test for correlated
Sample Evaluatlon Objective<br />
compare the effectiveness of different service delivery technlques<br />
To<br />
a particular type of client.<br />
wltb<br />
I I Criteri a<br />
in ratings before and after service using Goal Attainment Scallng<br />
Change<br />
co res.<br />
s<br />
A. General Procedure zz<br />
III Methodology<br />
One group pretest posttest study design<br />
Steps.<br />
O X Oz<br />
O1 observation/measurement<br />
where<br />
service<br />
before<br />
X : serwce/treatment<br />
on/measurement<br />
observati<br />
servlce<br />
after<br />
Identify clients that are to receive a particular servlce or<br />
I.<br />
i.e.; behavior modification, reality therapy, sheltered<br />
technique,<br />
workshop, etc.<br />
Measure the level of client functioning before he begins re-<br />
2.<br />
the service.<br />
ceiving<br />
D. T. & Stanley, J. C. Experlmental and quasi-experimental<br />
ZICampbell,<br />
for research. Chicago- Rand McNally, 1966.<br />
designs<br />
-107-
3. Conduct the servlce.<br />
Measure the level of client functloning az termlnation of<br />
4.<br />
service.<br />
the<br />
Tabulate the results for the first and second measurements<br />
5.<br />
conduct the necessary statlstics to make the results meaningful.<br />
and<br />
One possibility is a t-test for correlated samples 22.<br />
Measurement Approach. Goal Attainment Scaling (G.A.S.) 23<br />
B.<br />
Program Evaluation<br />
for<br />
A Technlque<br />
Attalnment Scalng is a procedure for identlfying behaviorally<br />
Goal<br />
goals and outcomes for clients on a simple scale ranging from<br />
defined<br />
unfavorable outcome to best antlcpated outcome. The procedure has<br />
most<br />
following cbaracterlstlcs (a_) client objectives are devised for or<br />
the<br />
the individual, (b) there is a system for assigning weights among these<br />
by<br />
(c_) expected outcomes are identified for each objective, (d)<br />
objectives,<br />
is a quantifiable follow-up system for these outcomes, and (e_) a<br />
there<br />
can be obtained which summarizes outcome across all objectives. A<br />
score<br />
feature of this procedure is that goals are established for the<br />
positive<br />
client and he is compared with himself in terms of degree of<br />
individual<br />
success.<br />
problem or goal area s identlfied between counselor and clent<br />
Each<br />
weighted as to its importance n the overall service process. The<br />
and<br />
system allows the counselor and/or client to arb'trarily as-<br />
weighting<br />
a level of importance to each area One or two diglt numbers may<br />
sign<br />
used, the higher the number the more important is the goal. If no<br />
be<br />
are assigned, it is assumed that all problem areas identified are<br />
weights<br />
equal importance. After the identification and weighting of each of<br />
of<br />
problem areas, objective behavioral descriptions of outcome on fve<br />
these<br />
are scaled from the most unfavorable outcome likely (assigned a<br />
levels<br />
two) to the most favorable outcome likely (assigned a plus two).<br />
mnus<br />
expected outcome of treatment for each problem is also stated (asslgned<br />
An<br />
score of zero) and is at the middle of the scale. A sample Goal Attain-<br />
a<br />
Follow-up Guide appears on the next page (Figure 6 1)<br />
ment<br />
procedure for converting what are qualitative dmenslons to quan-<br />
The<br />
scores is computationally simple. It enables many comparisons<br />
titative<br />
transforming raw scores to standard scores ncluding comparisons be-<br />
by<br />
groups of disabled, types of servlce, and staff members. An advan-<br />
tween<br />
of the Goal Attainment Scaling procedure is its flexibility. It<br />
tage<br />
no restrlctions on posslble goals and glves freedom to assign rela-<br />
places<br />
tive weights which are appropriate for each client<br />
G.A. Statistlcal analysis in psychology and education.<br />
22Ferguson,<br />
York McGraw-Hill, 1966, pp. 169-171.<br />
New<br />
T J Goal attalnment scaling at a county mental health<br />
23Kiresuk,<br />
Evaluation, 1973, 1, 12-18.<br />
service<br />
-108-
WORK SHEET
z<br />
z<br />
o<br />
-ii0-<br />
> o<br />
X4--<br />
o<br />
0<br />
o
WORK SHEET
The steps involved In developing goal attainment scales are as follows<br />
Collect information about the clleniC at the tlme of ntake or re-<br />
1<br />
and designate "problem areas" where change s desired to Insure a<br />
ferral<br />
rehabilitation. This can be done by (a_) counselor and client,<br />
successful<br />
cllent only, (c_) counselor only, and/or (d) others (e.g. family mem-<br />
(b)<br />
ber, etc.<br />
Identify these "problem areas" by giving each one a title. The<br />
2.<br />
s then placed across the top of the follow-up guide starting wlth<br />
tltle<br />
I for problem 1, Scale 2 for problem 2, etc., until each problem<br />
Scale<br />
Identlfied has been lsted. There is no minimum nor maximum number<br />
area<br />
areas necessary. These problem areas can then be weighted as to their<br />
of<br />
if so desired.<br />
mportance<br />
The counselor then states n behavioral terms under each "problem<br />
3.<br />
what he feels will be the outcomes, ranging on a 5-point scale, from<br />
area"<br />
most unfavorable service outcome thought likely to the best anticipated<br />
the<br />
with service. He also indicates the level of client functioning<br />
success<br />
in each problem area at the time of G.A.S. construction.<br />
The counselor then designates a follow-up date for determinlng<br />
4<br />
change.<br />
client<br />
At the deslgnated follow-up tme, a follow-up rater assesses goal<br />
5<br />
levels for each problem area for the client.<br />
attainment<br />
typical question that could be asked by a rehab11taton agency<br />
A<br />
How effective s a client-centered counseling approach wth clients<br />
s<br />
havlng drug abuse problems?<br />
ths example, the sample G.A.S. Follow-up Gulde was constructed<br />
For<br />
calculations carried out. Four additional cases were gven hypotheti-<br />
and<br />
cal pre and post scores for demonstration purposes.<br />
C Statistical Procedures<br />
Score each G A.S Follow-up Guide for "level at ntake"<br />
I<br />
--This score ndicates where the client was functioning prior<br />
(pretest)<br />
to counseling services<br />
procedure involves converting the G A.S. values to a standard T<br />
The<br />
(a T score has an average of 50 wlth a standard devlaton of 10)<br />
score<br />
The formula is 24<br />
T. J & Sherman, R. E Goal attainment scaling- A general<br />
24K1resuk,<br />
for evaluating comprehensive community mental health programs.<br />
method<br />
Community Mental Health Journal, 1968, 4_(6), 443-453.<br />
-112-
WORK SHEET
T=50+<br />
ZWX<br />
I0<br />
+<br />
%/(.7)ZWi2<br />
3(sWi)2<br />
W the weight asslgned to<br />
whereproblem<br />
area<br />
each<br />
level at which the cllent<br />
X<br />
functioning (ranges<br />
is<br />
from -2 to +2).<br />
For one client the information would appear lke thls<br />
Level at Intake<br />
Wi W 2 WiX<br />
X<br />
(Functionlng (Welght (Weight (Welght<br />
Scale<br />
Assigned) Squared) times level)<br />
Level)<br />
1 -2 3 9 -6<br />
2 -1 3 9 -3<br />
3 -I 2 4 -2<br />
4 -2 3 9 -6<br />
sW 11 zW 2 31 sWiX -17<br />
these values have been determlned, they can be placed n the T<br />
Once<br />
as follows"<br />
formula<br />
T= 50+<br />
zWiXc<br />
10<br />
2 + .3(zW) 2<br />
(.7).Wc<br />
(-17)<br />
!0<br />
+ (.7)(31)+ (3)(11) 2<br />
50<br />
-114-
WORK SHEET
-170<br />
+21.7 50 (.3)(121)<br />
+<br />
-170<br />
+%)2"i'.7 50<br />
36.3<br />
+<br />
-170<br />
-170<br />
=50+<br />
7.62<br />
50 + (-22.31)<br />
T I7 7 (level at intake)<br />
2. Calculate each G.A.S. follow-up score at the concluslon of<br />
Step<br />
The procedure Is the same as for "level of ntake" and would ap-<br />
servlce<br />
like this for the first client (see sample G.A.S. Follow-up Gulde for<br />
pear<br />
).<br />
scores<br />
Level at Fol l ow-up<br />
Scale X i Wi W 2 WX<br />
1 0 3 9 0<br />
2 +2 3 9 +6<br />
3 +1 2 4 +2<br />
4 +1 3 9 +3<br />
.Wc 11 zW 2 31 sWXi 11<br />
-116-
WORK SHEET
Convert follow-up ratlngs to T score as follows<br />
(11)_<br />
10<br />
50 + ^,:<br />
7)(31) (.3)(11)<br />
T<br />
2<br />
110<br />
+ 21.7 + 36.3<br />
5O<br />
110<br />
=50+<br />
7.62<br />
50 + 14.44<br />
T -64.¢I (level at follow-up)<br />
+<br />
for this cllent and four addltional clients are given as<br />
Scores<br />
l ows<br />
fol<br />
Cl lent Pre-test Post-test<br />
1 27.7 64.4<br />
2 38.9 56.2<br />
3 40.0 59.1<br />
4 42.0 60.4<br />
5 35.3 63.6<br />
3 Conduct a statistlcal test to compare the pretest scores<br />
Step<br />
the posttest scores. A t-test for correlated samples is approprlate<br />
wlth<br />
because there are two observations for the same subjects) 25<br />
("correlated"<br />
formula is"<br />
The<br />
G.A. Statistlcal .a.nalsis n psychology an__d educatlon<br />
25Ferguson,<br />
York- McGraw-Hill, 1966, pp. 169-171.<br />
New<br />
-118-
WORK SHEET
Calculations are as follows-<br />
D<br />
--IN.D (zD)21/( t 1)<br />
N<br />
degrees of freedom N 1<br />
where D difference between 01 and 02<br />
Posttest D D z<br />
Pretest<br />
01 02<br />
Client<br />
1 27.7 64.4 36.7 1346.9<br />
2 38.9 56.2 17.3 299.3<br />
3 40.0 59.1 19.1 364.8<br />
4 42.0 60.4 18.4 338.6<br />
5 35.3 63.6 28.3 800.9<br />
83.9 303.7<br />
36.8 60.7<br />
119.8 3150.5<br />
these flgures in the t-test formula results In the<br />
Substltutlng<br />
I owing<br />
fo<br />
t INsD2_ (.D)I/(N 1)<br />
119.8<br />
(119.8)21/(5 1)<br />
15(3150.5)<br />
-120-
WORK SHEET
8 119<br />
14352 01/4<br />
115752.5<br />
119.8<br />
1400.5/4<br />
119.8<br />
350.1<br />
119.8<br />
18.7<br />
degrees of freedom = N 1<br />
=5-1<br />
=4<br />
a table of critical values for t-tests (found in statlstlcs<br />
Using<br />
for 4 degrees of freedom, the necessary t value for signifl-<br />
textbooks)<br />
cance at the 05 probability level is 2.78.<br />
IV Interpretati on<br />
can be observed by the obtalned t value, significance was reached<br />
As<br />
our example at the .05 probability level because the obtained t value<br />
in<br />
the critical value of 2.78. The mean of the pretest scores for<br />
exceeded<br />
five clients receiving the client-centered counseling service was<br />
the<br />
the mean of the posttest scores was 60.7. Thus, in this fictltous<br />
36.8,<br />
the client-centered counseling approach appears to be a signifl-<br />
example,<br />
effective technique for working with rehabilitation clients with<br />
cantly<br />
abuse problems.<br />
drug<br />
-122-
26<br />
7<br />
EXAMPLE<br />
General Proceudre: One-group pretest posttest<br />
A.<br />
deslgn.<br />
study<br />
B. Measurement Approcah: The Human Serv±ce Scale.<br />
C. Statst±cal Procedures: Lnear regress±on.<br />
example was developed by Kenneth W. Reagles,<br />
26Thls<br />
Dmrector, Reglonal Rehabllmtatmon Research Inst±tute<br />
Research<br />
the Unversmty of W±scons±n-Mad±son. The Human Servmce Scale<br />
at<br />
reprinted wmth permmsson of Human Service Systems, Inc.<br />
ms
26<br />
7<br />
EXAMPLE<br />
General Proceudre: One-group pretest posttest<br />
A.<br />
deslgn.<br />
study<br />
B. Measurement Approcah: The Human Servlce Scale.<br />
C. Statlstcal Procedures: Lnear regression.<br />
example was developed by Kenneth W. Reagles,<br />
26Thls<br />
Dlrector, Regional Rehabltaton Research Instltute<br />
Research<br />
the University of W±sconsln-Madson. The Human Service Scale<br />
at<br />
reprinted wth permlsson of Human Servlce Systems, Inc.<br />
is
I Sample Evaluation Objective<br />
determine the relationship between client and/or rehabilitation<br />
To<br />
variables, e.g. cost of trainingand change in client need sat-<br />
process<br />
I sfactl on.<br />
II Criteria<br />
of relationship between client/process variables and client<br />
Degree<br />
satisfaction as determined by the Human Service Scale.<br />
need<br />
A. General Procedure27<br />
Ill Methodology<br />
One group pretest posttest study design<br />
Steps.<br />
01 X 02<br />
01 = observatlon/measurement<br />
where<br />
service<br />
before<br />
X = service/treatment<br />
on/measurement<br />
observati<br />
servi ce<br />
after<br />
Identify clients that are to receive a partlcular service or<br />
1.<br />
e.g. alcoholism treatment, personal-social counseling,<br />
training,<br />
training, physical restoration, marital counseling,<br />
pre-vocatonal<br />
etc.<br />
Admlnlster the Human Service Scale to all clients to ascer-<br />
2.<br />
the relatlve level of need satisfaction in seven need areas,<br />
taln<br />
as well as overall need satisfaction, prior to the receipt of services.<br />
D. T. & Stanley J. C. .xperimental an__d q.uasi-experlmental<br />
27Campbell,<br />
for research. Chicago- Rand McNally, 1966.<br />
designs<br />
-125-
3 Implement and monitor the service plan.<br />
Re-administer the Human Service Scale at the completlon of<br />
4.<br />
service<br />
the<br />
Score the Human Service Scale and compile these scores for<br />
5<br />
flrst (pre) and second (post) aaministratlons of the scale. Con-<br />
the<br />
the necessary statlstical procedures to determine the relatlon-<br />
duct<br />
between cost of training and changes in need satisfaction. One<br />
ship<br />
possibility for determining relationship is linear regres-<br />
procedural<br />
2B. This is a process of determining the line which describes<br />
sion<br />
relationship between two variables. Its usefulness is n deter-<br />
the<br />
or predicting the value of one variable from knowledge of the<br />
mining<br />
other va ri ab I e.<br />
B. Measurement Approach The Human Service Scale29<br />
Human Service Scale (HSS____) was developed, in part, by the Univer-<br />
The<br />
of Wisconsin Regional Rehabilitation Research Institute (UW-RRRI).<br />
slty<br />
HSS s intended to measure the extent to which client needs are met<br />
The<br />
several life areas.<br />
in<br />
herarchy of basic human needs served as the underlying theo-<br />
Maslow's<br />
ratlonale for the development of the scale 0. From appropriate tem<br />
retical<br />
factor analyses at the UW-RRRI, seven factors, or subscales, were den-<br />
and<br />
These were labeled Physiological, Economic Securlty, Emotlonal,<br />
tfied.<br />
Social, Economic Self-Esteem, and Vocational Self-Actualizai:on.<br />
Family,<br />
a technique called "Smallest Space Analysis TM it was determlned<br />
Using<br />
the configuration of the need categorles was not a simple linear h-<br />
that<br />
but was a spherical configuration. This configuration was used<br />
erarchy<br />
construct a profile for reporting the scoring results Thus, a descrip-<br />
to<br />
of relative client need satisfaction can be presented graphcally for<br />
tion<br />
seven need dimensions (a HSS Scoring Profile form appears on the next<br />
the<br />
as Figure 7.1)o<br />
page<br />
HSS consists of 80 items (a copy of the scale appears at the back<br />
The<br />
thl s exam--'-ple).<br />
of<br />
R. B. Fundamental statistics four psyc.h..ology. New York.<br />
2BMcCall,<br />
Bruce and World, Inc., 1970, pp. 86-112.<br />
Harcourt,<br />
K. W., Wright, G. N., & Butler, A. J. Human Service Scale<br />
29Reagles,<br />
WI" Human Service Systems, Inc,, (P. O. Box 5551, Zip 53705),<br />
Madison,<br />
1973.<br />
A. H. Motivation and personality,<br />
3°Maslow,<br />
Harper and Row, 1970, pp. 97-105.<br />
IL-<br />
(2nd Ed.) Evanston,<br />
I. M. & Guttman, L. Smallest Space Analysts of Intel-<br />
31Schlesinger,<br />
and achievement test. Ps),choloi.cal Bulletin, 1969, 71, 95-100.<br />
lgence<br />
-126-
WORK SHEET
Z<br />
NEE<br />
"/IfNOS3
WORK SHEET
questions whlch relate to each of the seven subscales are as<br />
The<br />
follows<br />
Subscale<br />
1 Physlologlcal<br />
2. Emotlonal Security<br />
3 Economlc Security<br />
4 Famlly<br />
5 Socl al<br />
6 Economic Sel f-Esteem<br />
7. Vocational Self-Actuallzation<br />
ques tl ons<br />
10, 11, 13, 19, 26, 28, 31, 36, 37<br />
5,<br />
40, 41, 47, 52, 57.<br />
39,<br />
8, 9, 11, 12, 14, 16, 19, 20, 21,<br />
6,<br />
24, 25, 29, 33, 34, 35, 38, 41, 44,<br />
22,<br />
50, 59, 62.<br />
16, 18, 21, 48, 60, 61.<br />
7, 9, 17, 27, 30, 32, 42, 45, 49, 62<br />
15, 23, 43, 46, 51, 53, 54, 55, 58, 59.<br />
1, 2, 3, 4, 56, 78, 79, 80.<br />
64, 65, 66, 67, 68, 69, 70, 71, 72,<br />
63,<br />
74, 75, 76, 77, 78, 79, 80<br />
73,<br />
are a varlety of uses of the scale and a variety of statistical<br />
There<br />
in which the scores can be used 32 Scoring involves assigned<br />
techniques<br />
for each item response, weights range from one to seven A scoring<br />
weights<br />
is available, but hand-scoring is a laborious process at this time.<br />
key<br />
scale is machine-scorable and is returned to Human Service Systems,<br />
The<br />
for scoring. The scoring output provides basic identifying Informa-<br />
Inc.<br />
about the client, raw scores for the seven need areas and an overall<br />
tion<br />
and normative Information which compares the client's performance<br />
score,<br />
others with similar characteristics (e g. male alcoholics age 30-35<br />
to<br />
years<br />
scores for subscales are relatively independent of each other<br />
Raw<br />
lntercorrelatlon) and represent the extent of need satsfactlon.<br />
(low<br />
hgher the score, the greater the level of need satisfaction for that<br />
The<br />
The overall score does not represent a simple summation of the seven<br />
area<br />
scores, but s an independent welghtlng system. Raw scores may<br />
subscale<br />
used in pre and post comparisons as in this example However, the scoring<br />
be<br />
also translates raw scores into percentiles comparing the client<br />
output<br />
all subjects (approximately 4000) who have ever completed the HSS<br />
with<br />
Norm Group) In addition, the scale user may specify other com-<br />
(Total<br />
groups for obtaining more specific percentile informatlon. A sam-<br />
parison<br />
scoring output is presented as Figure 7.2.<br />
ple<br />
K. W., Kravetz, S Wright, G. N & Butler, A. J A<br />
32Reagles,<br />
yardstick The Human Servlce Scale. Madison The University of<br />
better<br />
-Regional Rehabilitation Research Institute, 1974<br />
Wisconsin<br />
-130-
WORK SHEET
2;: Z<br />
z<br />
-132-<br />
E
WORK SHEET
C. Statlstical Procedures<br />
I_ Have each client's Human Service Scale scored by Human<br />
Step<br />
Systems, Inc. A sl'ngle pr'inted page of scoring information is<br />
Service<br />
produced by computer for each suboect.<br />
Tabulate the scores for each sbscale and the HSS overall<br />
Step2<br />
for each client for both pre and post administrations. For example,<br />
score<br />
and post scores for five clients for lust one subscale "Vocatlonal<br />
pre<br />
would appear as follows<br />
Self-Actualization"<br />
Vocational Self-Actual zat on<br />
Posttest Change<br />
Pretest<br />
Ient Oi 02 (dl ffe rence )<br />
Cl<br />
1 63 67 4<br />
2 54 60 6<br />
3 57 65 8<br />
4 51 63 12<br />
5 39 53 14<br />
3_. Detemlne the linear regression for the cost of tralning on<br />
Step<br />
in vocational self-actualization. The followlng procedure can be<br />
change<br />
uti I ized-<br />
onal<br />
Vocat<br />
of Sel f-Actual izat on<br />
Cost<br />
ni ng Change<br />
Tra<br />
X Y<br />
Cllent<br />
1 $1oo 4<br />
2 150 6<br />
3 200 8<br />
4 300 12<br />
5 350 14<br />
N 5 ZX 1,100 .Y 44<br />
X 220 Y 8.80<br />
-134-<br />
X 2 y2<br />
10,000 16<br />
22,500 36<br />
40,000 64<br />
90,000 144<br />
122,500 196<br />
SX 2 285,000 456<br />
XY<br />
400<br />
900<br />
1,600<br />
3,600<br />
4,900<br />
sXY 11,400
WORK SHEET
instructs one to Identify cost of training as the X varlable<br />
This<br />
Vocational Self-Actualizatlon as the Y variable. One then squares<br />
and<br />
X value 33 and enters the result under X 2. The same procedure is fol-<br />
each<br />
for the Y variable. To obtain XY simply multlply each X by its cor-<br />
lowed<br />
responding Y value to obtain XY. After doing this, sum each column.<br />
4_. Utilize the previously computed values to calculate the<br />
Step<br />
relatlonships between cost of training and vocational self-<br />
lnear<br />
A linear relationship is an association between two varl-<br />
actualization.<br />
which can be represented on a graph by a straight line. The equa-<br />
ables<br />
tion Is"<br />
Y=X+a<br />
where Y variable Y<br />
of the lne<br />
lope<br />
constant value)<br />
(a<br />
X variable X<br />
at which the<br />
point<br />
Intercepts the<br />
line<br />
axis (a cons tant<br />
Y<br />
value)<br />
Thus, a linear relationship can be graphically represented as follows<br />
1 2 3 4 5 6 7<br />
X axis<br />
In this example X 2 represents the square of the value of X,<br />
33Note:<br />
chi-square.<br />
not<br />
-136-
WORK SHEET
Is now ready to substitute the calculated Infomation nto the<br />
One<br />
fomulas to detemine the values for /p and a.<br />
following<br />
(X)(ZY)<br />
N(XY)-<br />
2 (ZX) 2<br />
NZX<br />
c: Y -IpX<br />
For this example, the values n the formula for Zp are-<br />
=<br />
(1100)(44)<br />
5(11,400)<br />
(1100)(1100)<br />
5(285,0001-<br />
48,400<br />
57,000<br />
1,210,000<br />
1,425,000<br />
8600<br />
215,000<br />
]-,04<br />
Values In the formula for a are<br />
8.80 (.041(220)<br />
8.80 8 80<br />
-138-
WORK SHEET
the slope of the line (/p) Is .04 and the intercept (a) begins<br />
Thus,<br />
O. Presented graphically, it appears as follows:<br />
at<br />
sion<br />
N<br />
16<br />
14<br />
10<br />
4<br />
2<br />
#,<br />
s<br />
s<br />
#,<br />
s<br />
s<br />
S<br />
s<br />
s<br />
s<br />
s<br />
s<br />
S<br />
b .04<br />
i,,, I,<br />
$50 I00 150 200 250 300 350 400<br />
Cost of Training (X)<br />
7.3. Regresslon llne for cost of tralnlng on<br />
Figure<br />
in vocational self-actualization<br />
change<br />
Next, we substitute the resultant constants b and a nto the regres-<br />
equatlon<br />
Y=bX-cz<br />
Y .04X 0<br />
-140-<br />
s<br />
s<br />
•
WORK SHEET
this informatlon one can predict Y (vocational self-actualizatlon)<br />
With<br />
knowledge of X (cost of training). For example, if X $50, then<br />
from<br />
Y 2 as follows"<br />
.04(50) 0<br />
= 2.00 0<br />
2.00<br />
where = a predicted value for Y.<br />
f cost of training equals $700, one can predict the galn<br />
Similarly,<br />
vocational self-actualization satisfaction will equal 28, i e.<br />
in<br />
.04(700) 0<br />
28.00 0<br />
28.00<br />
IV Interpretation<br />
shown in Figure 3, the regression line for the linear relation-<br />
As<br />
between cost of training and vocational self-actualization suggests<br />
shlp<br />
positive or direct relatlonship. This means that as the cost of traln-<br />
a<br />
ncreases so does the ga'n in client need satisfaction for vocational<br />
ng<br />
sel f-actua I i zati on.<br />
such relationshlps between two variables and the resul-<br />
Similarly,<br />
predictability can be conducted among other variables. Lnear re-<br />
tant<br />
is the basis for conducting multiple linear regresslon analysis<br />
gression<br />
relationships among more than two variables.<br />
nvolwng<br />
-142-
WORK SHEET
THIS PART TO BE COMPLETED BY THE PROFESSIONAL<br />
(First) (M<br />
(Last)<br />
^OIbRESS<br />
EPHONE<br />
(Street or R R<br />
JCIAL SECURITY NO L..i_J<br />
GE BIRTH DATE<br />
SEX, ® ®<br />
V<br />
(State) (Z,p)<br />
MO DAY YR<br />
-.LI ENT NO () ®I<br />
OUNSELOR NO<br />
r-] IIARITAL STATUS<br />
Married<br />
)<br />
W,dowed<br />
.<br />
£)vorced<br />
()<br />
Separated<br />
)<br />
Never Married<br />
_,<br />
Marriage Annulled<br />
C)<br />
" Unknown<br />
EDUCATION<br />
None<br />
grade<br />
7<br />
grade<br />
8th<br />
grade<br />
9<br />
School Diploma<br />
Hgh<br />
Techmcal wthout<br />
Vocational<br />
Lcensure/Certflcaton<br />
Techmcal wth<br />
Vocational<br />
censu re/Certfl cat on<br />
L<br />
College one or more<br />
Attended<br />
semesters<br />
Year College Degree<br />
Four<br />
Degree<br />
Graduate<br />
Unknown<br />
OF DEPENDENTS other<br />
NUMBER<br />
yourself<br />
than<br />
r- HERITAGE<br />
Wh=te<br />
_)<br />
Black<br />
C)<br />
Amencan Indian<br />
_( panlsh<br />
)<br />
Surname<br />
()ther<br />
LJnknown<br />
O<br />
FIRET EDITION COPYRIGHT ( 1973<br />
®®!<br />
®®i<br />
@®I<br />
@®I<br />
-®®®®@®<br />
®®®®®®<br />
@®®®®®<br />
®®®®®®<br />
®®®®®®<br />
@®®®@®<br />
®®®®®®<br />
®®@®®®<br />
@®@®®®<br />
ADMIN<br />
LIVING ARRANGEMENT<br />
Liv,ng alone<br />
0<br />
L,v,ng w, th spouse<br />
0<br />
L=vlng w,th one or both parents<br />
0<br />
step parents)<br />
(=ncludlng<br />
Liwng w=th non relat=ves<br />
O<br />
Other O<br />
r] WORK STATUS<br />
DISCOOE<br />
®@®<br />
®®®<br />
@®®<br />
@®®<br />
®@®<br />
@®@<br />
®®®<br />
®®®<br />
®@®<br />
Wage or salaried worker (competlt=ve<br />
0<br />
market)<br />
labor<br />
Wage or salaried worker (sheltered<br />
O<br />
workshop)<br />
Self employed (except BEP)<br />
O<br />
State agency managed bus=ness<br />
O<br />
(BEP)<br />
enterprise<br />
Homemaker<br />
O<br />
Unpa=d famdy worker<br />
O<br />
Not work=ng student<br />
O<br />
O Unemployed<br />
r--] PRIMARY SOURCE OF SUPPORT<br />
0 Current earnings, nterest,<br />
Famdy and friends<br />
O<br />
Private rel=ef agency<br />
O<br />
Pubhc assistance, at least partly<br />
O<br />
Federal funds<br />
w=th<br />
d=v=dends, rent<br />
Pubhc assistance, without Federal funds<br />
O<br />
Pubhc nsttutlon-tax supported<br />
O<br />
Workmen's compensat=on<br />
O<br />
Soctai Security Disab=hty Insurance benefits<br />
O<br />
Other dsabl=ty, s=ckness, surwvors', or age-<br />
O<br />
benef=ts (except from private<br />
ret=rement<br />
=nsurance) unemployment nsurance<br />
Annu=ty or other non d=sab=hty<br />
O<br />
benefits (private nsurance)<br />
Cnsurance<br />
SPECIAL<br />
CODES<br />
@®@®®<br />
®®®®®<br />
®®®®®<br />
®®®®®<br />
®®®®®<br />
benefits<br />
D=sabl=ty or s=ckness benehts<br />
O<br />
Cnsurance) sav=ngs, other sources<br />
(private<br />
O Not reported<br />
r--] VETERAN<br />
®@®®®<br />
@®®®®<br />
Yes 0 No 0 Selective Serwce relectee<br />
0<br />
SERVICE SYSTEMS INC MADISON WI 53705 NCS Tran$ Optic F1706 5432<br />
HUMAN<br />
o-i111111111111111111111111111111111111111111111111<br />
ADD<br />
,0<br />
DROP<br />
I0<br />
to CHG<br />
Io<br />
OTHER
DI RECTIONS<br />
What IS your mal.__.n source of support<br />
(<br />
own earmngs (wages, workshop<br />
(your<br />
income from own bus<br />
payments<br />
hess)<br />
property or other invest<br />
()sawngs,<br />
ments<br />
of someone else m faro=Iv<br />
(earnmgs<br />
Security, pension payments,<br />
()Soc=al<br />
bnemployment Compensation<br />
or<br />
payments<br />
Public Assistance or Welfare pay<br />
()<br />
ments<br />
How much Public Assistance or Wel-<br />
2<br />
payments (but no.._t earnings, Social<br />
fare<br />
Pension Payments, or unem-<br />
Security,<br />
compensation payments) are<br />
ployment<br />
receiving per month><br />
you<br />
Pubhc Welfare Ass=stance at this<br />
(No<br />
t=me<br />
to 75 dollars per month<br />
()1<br />
to 150 dollars per month<br />
()76<br />
to 225 dollars per month<br />
()151<br />
than 225 dollars per month<br />
(more<br />
How much do you earn 3 (wges, work-<br />
payments, income from own busi-<br />
shop<br />
savings, property or other nvest-<br />
ness,<br />
per week (nearest dollar);)<br />
ments)<br />
()none<br />
to 35 dollars per week<br />
()1<br />
to 70 dollars per week<br />
(36<br />
DARKEN THE CIRCLE IN FRONT OF DESIRED RESPONSE FOR EACH QUESTION<br />
How often do you have trouble show=ng your feehngs to your famlly<br />
9<br />
very often () sometimes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
()<br />
to 105 dollars per week<br />
(as often as not<br />
()71<br />
dollars or more per week 16 How often do you<br />
()106<br />
How many jobs (either paid or unpaid<br />
4<br />
have you had ,n the last six months ;)<br />
work)<br />
(dd not work)<br />
()none<br />
job (2 jobs ()3 jobs (4 or more<br />
(1<br />
How often are you bothered by rap=d<br />
5<br />
beat<br />
heart<br />
very often () somet=mes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often are you uncertain about<br />
6<br />
you make;)<br />
decisions<br />
often ( somet=mes<br />
()very<br />
() hardly ever<br />
()often<br />
often as not<br />
(as<br />
7 How often, when you need help, can<br />
find someone to help you<br />
you<br />
very often () somet= rues<br />
(<br />
() hardly ever<br />
(often<br />
as often as not<br />
(<br />
How often do you worry about grow-<br />
8<br />
old;)<br />
=ng<br />
often ( sometimes<br />
()very<br />
( hardly ever<br />
()often<br />
as often as not<br />
How often are you bothered by shortness of breath when not exercising;)<br />
10<br />
very often () somet mes<br />
()<br />
I) hardly ever<br />
()often<br />
often as not<br />
(as<br />
How often do you feel depressed, down, or very unhappy;)<br />
11<br />
very often () somet= mes<br />
()<br />
( hardly ever<br />
()often<br />
often as not J<br />
(as<br />
How often do you feel down or discouraged because your major problems cause yu<br />
12<br />
waste time;)<br />
to<br />
very often ( sometimes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
()<br />
How often do you become so sick you have to cut down on your usual actzvltes;)<br />
13<br />
very often ( somet=mes<br />
(<br />
( hardly ever<br />
()often<br />
as often as not<br />
()<br />
How often do you feel restless;)<br />
14<br />
very often ( sometimes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often do you get together w=th<br />
15<br />
(go,ng out together or v=sit=ng ,n<br />
friends<br />
others' homeP<br />
each<br />
very often ( sometl mes<br />
()<br />
() hardly ever<br />
()often<br />
worry about the<br />
future;)<br />
very often ( somet=mes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often has your famzly faded to<br />
17<br />
you when you needed help?<br />
help<br />
very often () somet= mes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often do you worry about your<br />
18<br />
havng enough money><br />
famdy<br />
very often ( somet= rues<br />
()<br />
often ( hardly ever<br />
()<br />
as often as not<br />
()<br />
How often do you tend to go to<br />
19<br />
under pressure7<br />
pieces<br />
very often ( sometl mes<br />
()<br />
() hardly ever<br />
()often<br />
often as not<br />
(as<br />
How often are you able to solve your<br />
20<br />
problems<br />
()<br />
own<br />
often () sometl rues<br />
very<br />
often () hardly ever<br />
(<br />
often as not<br />
(as<br />
EDITION COPYRIGHT ( 1973<br />
i11111111111111111111111111111111<br />
FIRST<br />
How often do you worry about ge<br />
21<br />
ahead m the worid7<br />
tlng<br />
very often ( sometimes<br />
()<br />
often () hardly ever<br />
(<br />
as often as not<br />
()<br />
How often do you worry about ge<br />
22<br />
along with your family;)<br />
tlng<br />
very often () somet=mes<br />
()<br />
often ( hardly ever<br />
()<br />
as often as not<br />
()<br />
How often do you become interested<br />
23<br />
something new7<br />
=n<br />
often () sometimes<br />
()very<br />
often () hardly ever<br />
(<br />
as often as not<br />
(<br />
How often do you treat other people<br />
24<br />
badly;)<br />
very often ( sometimes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
()<br />
How often have you felt that you are<br />
25<br />
the kind of famdy member that you<br />
not<br />
like to be7<br />
would<br />
very often ( somet=mes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
()<br />
How often are you bothered by<br />
26<br />
twitches, trembhng, or shakes><br />
muscle<br />
very often () somet mes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often does your famdy accept<br />
27<br />
as you are<br />
you<br />
very often () someti mes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often do you have headaches;<br />
28<br />
very often () sometimes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
()<br />
lllllllllllUl
iil<br />
How many hours each week do you<br />
55<br />
on activities with other people in<br />
spend<br />
community7<br />
your<br />
hour or less () 14 to 19 hours<br />
()<br />
2 to 7 hours ( 20 hours or more<br />
()<br />
8 to 13 hours<br />
()<br />
How many weeks during the last six<br />
56<br />
were you unemployed7<br />
months<br />
none ( 17 24 weeks<br />
()<br />
8 weeks () 24 or more weeks<br />
()<br />
9 16 weeks<br />
(<br />
During the last six months, about how<br />
57<br />
days have your major problems kept<br />
many<br />
in bed all or most of the day><br />
you<br />
none () 15-21<br />
()<br />
7 (E) 22 or more<br />
(E)<br />
8 14<br />
()<br />
How many people do you know<br />
5<br />
you feel free to talk to about per-<br />
whom<br />
things and ploblems7<br />
sonal<br />
very many () a few<br />
()<br />
many (E) none<br />
()<br />
some<br />
(<br />
How satisfied are you with your<br />
59<br />
hfe<br />
social<br />
very sattsfled<br />
()<br />
sat=stied<br />
()<br />
too satsf=ed but not too<br />
not<br />
dlssattsfled<br />
d=ssat=sfled<br />
(<br />
very dissatisfied<br />
(E)<br />
Whmh of the following statements<br />
60<br />
describes your present financial<br />
best<br />
situation7<br />
very good ( poor<br />
(E)<br />
good () very poor<br />
(<br />
average<br />
(<br />
Apart from mortgages on your house,<br />
61<br />
how many debts could you pay off n<br />
two months7<br />
next<br />
none of them<br />
()<br />
a few of them<br />
()<br />
some of them<br />
()<br />
of them<br />
(all<br />
have no debts<br />
(<br />
the<br />
Taking all things together, how would<br />
62<br />
descr=be your family hfe7<br />
you<br />
very happy<br />
()<br />
happy<br />
(E)<br />
not too happy but not too unhappy<br />
(E)<br />
unhappy<br />
(<br />
very unhappy<br />
(<br />
Which of the following best descr=bes<br />
63<br />
you are presently doing:<br />
what<br />
work for wages or salary or n own<br />
()<br />
including nvestments<br />
busness<br />
homemaker or housew=fe (housework<br />
(E)<br />
own family)<br />
for<br />
work in workshop or homebound<br />
()<br />
employment<br />
student or job tramng<br />
(<br />
unemployed<br />
IIII!1111111111<br />
YOUR ANSWER TO QUESTION 71 How often do you find t hard to<br />
IF<br />
63 WAS "E' (UNEMPLOYED), make friends w=th your present co- ork-<br />
NO<br />
HERE IF NOT, PLEASE CON- ere or people who are doing what y }u do ;<br />
STOP<br />
STUDENTS, PERSONS IN () very often () sometimes<br />
TINUE<br />
AND HOUSEWIVES () often ( hardly ever<br />
TRAINING,<br />
ANSWER THE FOLLOWING ( as often as not<br />
SHOULD<br />
ABOUT JOBS WITH 72 How often are you treated fa=rl y n<br />
QUESTIONS<br />
PRESENT ACTIVITY (SCHOOL, your present work<br />
THEIR<br />
OR HOUSEWORK) IN () very often () sometimes<br />
TRAINING,<br />
AS THEIR "WORK" AT THIS () often ( hardly ever<br />
MIND<br />
) as often as not<br />
TIME<br />
How often does your tresent vt fork<br />
73<br />
How often does your present work let you do something new each day<br />
64<br />
you make dems=ons on your own () very often () somet=mes<br />
let<br />
very often ( somet= mes (E) often () hardly ever<br />
(E)<br />
often ( hardly ever ( as often as not<br />
(E)<br />
as often as not 74 How often does your present ork<br />
(<br />
How often does your present work let you try out your own ideas7<br />
65<br />
you enough to do ( very often ( sometimes<br />
give<br />
very often ( somet, mes () often () hardly ever<br />
()<br />
often () hardly ever ( as often as not<br />
()<br />
as often as not 75 How often do you find that yo,u<br />
()<br />
During the last two weeks, how many really enjoy your [}resent work<br />
66<br />
of work did you m=ss due to a mnor () very often () somet=mes<br />
days<br />
such as a cold or sore throat () often () hardly ever<br />
smkness<br />
none ( 5 6 ( as often as not<br />
()<br />
2 7 or more 76 How often are you told m you=r<br />
()<br />
3 4 present work that you have done a<br />
()<br />
How many hours do you no_._w work good job<br />
67<br />
week7 () very often ( somet=mes<br />
each<br />
10 or less ( 31 to 40 hours () often () hardly ever<br />
()<br />
11 to 20 hours () over 40 hours () as often as not<br />
()<br />
21 to 30 hours 77 How often does your resent vork<br />
(<br />
How often do you learn new th=ngs give you a chance to make use of our<br />
68<br />
your present work abll=t=es<br />
from<br />
very often ( somet=mes (E) very often (E) somet=mesl<br />
(E)<br />
often (E) hardly ever often () hardly evel<br />
(E)<br />
as often as not as often as not<br />
()<br />
this list of activities that you 78 How steady is your present jot or<br />
Read<br />
take part n where you work the work you do<br />
may<br />
1 belong to some type of club or<br />
() very steady<br />
composed of people () steady<br />
organization<br />
whom work or who have reasonably steady<br />
with<br />
work unsteady<br />
s=m=lar<br />
belong to a un=on, attend umon () 2 unsteady<br />
very<br />
meet=ngs<br />
somahze after work hours with 79 What do other people th=nk of<br />
3<br />
workers your lob7<br />
fellow<br />
other actvlt=es related to your () they think it s a very good job<br />
4<br />
( they thnk =t s a good job<br />
work<br />
many of the above do you do : ( they think it s an average job<br />
How<br />
of them ( they think it is a poor job<br />
none<br />
one of them ( they thnk It is a very poor job<br />
()<br />
two of them 80 How does your present job (work)<br />
(<br />
three of them compare wth jobs you've had n the past<br />
(<br />
four or more of them () my present job =s much better<br />
()<br />
What =s the tota._._l number of hours () my present job is better<br />
70<br />
spend each week on the above actlv- () my present job s just as good<br />
you<br />
Choose one of the following ( my present job s worse<br />
itles<br />
2 hours or less (9 to 11 hours ( my present job s much worse<br />
()<br />
3 to 5 hours ( 12 hours or more STOP<br />
()<br />
( 6 to 8 hours<br />
THANK YOU FOR YOUR COOP.RATIO<br />
IIIIIIIIIIIIIIlilIIIlIiliI/ilL<br />
,=
9<br />
,/or<br />
30<br />
*all<br />
he<br />
3<br />
tor,<br />
often do things happen to make<br />
How<br />
angry<br />
often () sometimes<br />
ery<br />
() hardly ever<br />
often<br />
often as not<br />
as<br />
often do other members of the<br />
How<br />
talk to you about what went on<br />
ily<br />
the day'<br />
mg<br />
often ( sometimes<br />
ry<br />
() hardly ever<br />
),ten<br />
often as not<br />
as<br />
often do you feel dizzy'<br />
How<br />
often sometimes<br />
very<br />
() hardly ever<br />
often<br />
often as not<br />
as<br />
speaking, how often do you<br />
Generally<br />
your famdy about what went on<br />
to<br />
the day'<br />
ing<br />
often ( sometimes<br />
very<br />
() hardly ever<br />
often<br />
as often as not<br />
general, how often do you feel<br />
In<br />
Ipless'<br />
often () sometimes<br />
very<br />
( hardly ever<br />
often<br />
often as not<br />
as<br />
often have you consulted a doc-<br />
How<br />
psychiatrist, psychologzst, or anyone<br />
about a nervous problem'<br />
else<br />
often ( sometimes<br />
(very<br />
often hard ly ever<br />
(<br />
often as not<br />
()as<br />
How often do your major problems<br />
35,<br />
you feel inferior'<br />
make<br />
very often ( sometimes<br />
(<br />
often ( hardly ever<br />
(<br />
as often as not<br />
()<br />
How often in the past year have you<br />
36,<br />
sePn a<br />
or been hosp,tahzed for<br />
doctor<br />
ur physmal problems'<br />
yo,<br />
very often ( sometimes<br />
)<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often do you have general aches<br />
37<br />
pains'<br />
and<br />
very often () somet mes<br />
()<br />
often () hardly ever<br />
(<br />
as often as not<br />
(<br />
How often do your major problems<br />
38<br />
it difficult for you to make friends'<br />
make<br />
very often () sometl rues<br />
()<br />
often () hardly ever<br />
()<br />
How often have you felt that you are<br />
41<br />
to have a nervous breakdown><br />
going<br />
very often ( somet mes<br />
@<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
About how much time a week do you<br />
42<br />
doing things together with your<br />
spend<br />
family'<br />
hours or less<br />
5<br />
6 to 11 hours<br />
()<br />
12 to 17 hours<br />
(<br />
18 to 23 hours<br />
()<br />
( 24 hours or more<br />
Read the list of clubs and orgamza-<br />
43<br />
to which people may belong<br />
t=ons<br />
parent-teachers group<br />
any<br />
church-connected groups (usher's<br />
2<br />
Ladies Aid, etc<br />
club,<br />
fraternal lodge or auxiliary<br />
3<br />
neighborhood clubs, commumty<br />
4<br />
(including YWCA, YMCA)<br />
center<br />
card clubs or social clubs<br />
5<br />
veteran's association<br />
6<br />
service club (Rotary, Lions, etc<br />
7<br />
cwc organizations (partm=patfon m<br />
8<br />
drives, Red Cross, etc<br />
charity<br />
sports team<br />
9<br />
partmfpaton m political activities<br />
10<br />
a<br />
club or party<br />
political<br />
many of the above organizations do<br />
How<br />
take an actwe part in'<br />
you<br />
none of them () 5 or 6 of them<br />
()<br />
or 2 of them () 7 or more of them<br />
()<br />
3 or 4 of them<br />
(<br />
How often do you feel bored'<br />
44<br />
very often ( sometl rues<br />
()<br />
often ( hardly ever<br />
(<br />
as often as not<br />
()<br />
the following list of things<br />
Read<br />
may do together<br />
Illes<br />
visit frmnds<br />
1<br />
go to a movie, bowling, sporting event,<br />
2<br />
some other entertainment<br />
or<br />
spend an evening just talking with<br />
3<br />
other<br />
each<br />
working on some household prolect<br />
4<br />
entertaining friends in your home<br />
5<br />
go shopping<br />
6<br />
have a good laugh together or share<br />
7<br />
joke<br />
a<br />
eat out in a restaurant<br />
8<br />
are affectionate toward each other<br />
9<br />
take a drive or go for a walk<br />
10<br />
help a family member solve some<br />
11<br />
problem<br />
often as not 12 take part in some religious act,vlty<br />
)as<br />
How often do you have a common How<br />
39<br />
or the flu<br />
cld<br />
very often ( sometl mes<br />
)<br />
often ( hardly ever<br />
)<br />
often as not<br />
@as<br />
lww'eer;o you have sk|n rashes'<br />
'<br />
very often () sometimes<br />
)<br />
often ( hardly ever<br />
many of these things does your<br />
do together'<br />
family<br />
of these things ()5 or 6 of them<br />
()none<br />
or 2 of them () 7 or more of<br />
()<br />
3 or 4 of them these things<br />
(<br />
In the last year, how many new<br />
46<br />
have you made'<br />
friends<br />
very many ( a few<br />
(<br />
many ( none<br />
(<br />
some<br />
()<br />
How often are you bothered by an<br />
47<br />
stomach'<br />
upset<br />
very often () sometimes<br />
()<br />
often ( hardly ever<br />
()<br />
as often as not<br />
(<br />
How often do you worry about not<br />
48<br />
enough money'<br />
hav,ng<br />
very often ( sometimes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
(<br />
How often do you like spending<br />
49<br />
with your famdyY<br />
time<br />
very often<br />
()<br />
often<br />
()<br />
as often as not<br />
(<br />
sometimes<br />
(<br />
ever<br />
hardly<br />
How often do your major problems<br />
50<br />
you from making use of your<br />
keep<br />
abilities'<br />
very often () sometimes<br />
()<br />
() hardly ever<br />
()often<br />
as often as not<br />
()<br />
51 About how many people did you<br />
dunng the last year, other than<br />
meet<br />
you meet where you work, that<br />
those<br />
never met before'<br />
you<br />
very many ( a few<br />
()<br />
many none<br />
()<br />
some<br />
(<br />
How often do you worry about<br />
52<br />
healthy<br />
your<br />
very often ( sometimes<br />
()<br />
often () hardly ever<br />
()<br />
as often as not<br />
()<br />
About how many friends do you<br />
53<br />
keep in touch withy<br />
usually<br />
very many ( a few<br />
()<br />
many ( none<br />
()<br />
some<br />
(<br />
Read this list of which<br />
activities you<br />
5j,<br />
take part In with other people in<br />
might<br />
your community<br />
sports football, basketball, tenms,<br />
1<br />
etc<br />
golf,<br />
outdoor activities hunting fishing,<br />
2<br />
etc<br />
hiking,<br />
indoor activities bowling, table<br />
3<br />
dancing, cards, etc<br />
tenms,<br />
4 other somai activities<br />
many of the above activities do you<br />
How<br />
part m with other people in your<br />
take<br />
communltyY<br />
none of them ( 3 of them<br />
()<br />
of them () 4 or more<br />
()<br />
2 of them of them<br />
(<br />
PLEASE CONTINUE ON NEXT PAGE<br />
often as not HUMAN SERVICE SYSTEMS INC MADISON WI 53705<br />
=(111111111111111111111111111111111111111111111111<br />
as
PART THREE<br />
Pretest Posttest Control Group Study Design<br />
R 01 X 0 2<br />
R random assignment<br />
where:<br />
sub3ects<br />
of<br />
X<br />
0 3<br />
servlce/treatment<br />
0 4<br />
observatlon/measure-<br />
01<br />
before service<br />
ment<br />
experimental<br />
to<br />
group<br />
0 2<br />
0 3<br />
0 4<br />
observatlon/measurement<br />
service to experi-<br />
after<br />
mental group<br />
observation/measurement<br />
service for con-<br />
before<br />
trast group (no service)<br />
observatlon/measurement<br />
service for con-<br />
after<br />
trast group (no servlce)<br />
design calls for a sample of sub3ects that are randomly<br />
This<br />
to one of two groups. These groups are then examined<br />
assigned<br />
but only one receives service (X). The differences in<br />
twice,<br />
change from pre to post score measures of the two groups<br />
the<br />
is attributed to the effect of the service.<br />
The evaluator identifies a sample of sub3ects. From<br />
Method.<br />
sample he randomly assigns each of the members to one of<br />
this<br />
two groups, the sub3ects in both groups are then observed<br />
the<br />
measured. One group is given a service (exper±mental group)<br />
or<br />
the other is not (control/contrast group). After the<br />
while<br />
the groups are again measured. The difference between<br />
service<br />
measures for the two groups are then evaluated ststlstl-<br />
these<br />
cally.<br />
This design involves random assignment of sub3ects to<br />
Value.<br />
thus, permitting the evaluator to assume that the two<br />
groups;<br />
have been "equalized". This means the groups are<br />
groups<br />
to be equal, within chance limits, in all possible<br />
assumed<br />
Having a contrast group allows the evaluator<br />
characterlstcs.<br />
compare the change from pretest to posttest for the experi-<br />
to<br />
group with the before and after measures of the group<br />
mental<br />
the specfled service; thus, further ad3ust±ng for<br />
without<br />
to the experimental group during the evaluation<br />
cahnges<br />
The design's limitation is the lack of control by<br />
period.<br />
the evaluator for the affect of pretesting.
PART THREE<br />
Pretest Posttest Control Group Study Design<br />
0 X 0<br />
1<br />
R<br />
2<br />
0 0<br />
3 4<br />
R<br />
R random asslgnment<br />
where:<br />
sub3ects<br />
of<br />
X<br />
servlce/treatment<br />
observatlon/measure-<br />
01<br />
before service<br />
ment<br />
exper:mental<br />
to<br />
group<br />
0 2<br />
0 3<br />
0 4<br />
bservaton/measurement<br />
service to experl-<br />
after<br />
mental group<br />
observation/measurement<br />
service for con-<br />
before<br />
trast group (no service)<br />
observatlon/measurement<br />
service for con-<br />
after<br />
trast group (no service)<br />
deslgn calls for a sample of sub3ects that are randomly<br />
Th:s<br />
to one of two groups. These groups are then examined<br />
assigned<br />
but only one receives service (X). The differences in<br />
twice,<br />
change from pre to post score measures of the two groups<br />
the<br />
is attributed to the effect of the service.<br />
The evaluator Identlfes a sample of sub3ects. From<br />
Method.<br />
sample he randomly assigns each of the members to one of<br />
thls<br />
two groups, the sub3ects in both groups are then observed<br />
the<br />
measured. One group s given a servlce (experimental group)<br />
or<br />
the other is not (control/contrast grouD). After the<br />
while<br />
the groups are again measured. The difference between<br />
service<br />
measures for the two groups are then evaluated ststst±-<br />
these<br />
cally.<br />
Ths design Involves random assignment of sub3ects to<br />
Value.<br />
thus, permttlng the evaluator to assume that the two<br />
groups;<br />
have been "equalized". Thls means the grouDs are<br />
groups<br />
to be equal, wlthn chance lmlts, in all possible<br />
assumed<br />
Having a contrast group allows the evaluator<br />
characterlstcs.<br />
compare the change from pretest to posttest for the experl-<br />
to<br />
group wlth the before and after measures of the group<br />
mental<br />
the specified service; thus, further ad3ustng for<br />
without<br />
to the experimental group during the evaluation<br />
cahnges<br />
The design's lmltatlon is the lack of control by<br />
period.<br />
the evaluator for the affect of pretesting.
ao<br />
So<br />
Co<br />
EXAMPLE 8<br />
Procedure: Pretest posttest control<br />
General<br />
study deslgn.<br />
group<br />
Approach: Questlonnare for cllent<br />
Measurement<br />
of the rehabltaton process.<br />
knowledge<br />
Procedures: t-test for ndependent<br />
Statlstcal<br />
samples.
EXAMPLE 8<br />
General Procedure: Pretest posttest control<br />
A.<br />
study deslgn.<br />
group<br />
Measurement Approach: Questlonnalre for clent<br />
B.<br />
of the rehabltatlon process.<br />
knowledge<br />
Statistical Procedures: t-test for ndependent<br />
C.<br />
samples.
I Sample Evaluation Objectlve<br />
compare the effectlveness of two dlfferent methods of orienta-<br />
To<br />
to rehabilitation services for new clients.<br />
tion<br />
II Criteria<br />
in scores of clients measured before and after a special<br />
Change<br />
program using a 10 item questionnaire for each group.<br />
orientation<br />
A. General Procedure34<br />
III Methodology<br />
Pretest posttest control group study design<br />
Steps.<br />
R = random assignment of<br />
where<br />
to experlmental<br />
clients<br />
and contrast groups<br />
X = servlce/treatment<br />
0 : observation/measurement<br />
Identify clients who have been accepted for rehabltatlon<br />
1.<br />
but who have not recelved any orientation to the rehabil-<br />
services<br />
itation program.<br />
D. T. & Stanley, J. C. Experimental and quasi-experimental<br />
34Campbell,<br />
for research. Chicago: Rand McNally, 1966.<br />
designs<br />
-151-
Randomly asslgn these clients to one of the two groups uslng<br />
2.<br />
of the random sampling methods described in the Introduction of<br />
one<br />
this manual.<br />
Administer a measure to determine the level of knowledge of<br />
3.<br />
rehabilitation process to both groups before the orientation pro-<br />
the<br />
gram is initiated.<br />
Carry out the special orientation program wth the exper-<br />
4.<br />
group and give the usual orientation or none to the contrast<br />
mental<br />
group.<br />
Measure the level of client knowledge of the rehablltaton<br />
5.<br />
after the special orientation program with the same instru-<br />
process<br />
ment used prior to the program.<br />
Tabulate results and conduct the necessary statistics to make<br />
6.<br />
results meaningful. One possibility is a t-test for independent<br />
the<br />
samples 3s.<br />
Measurement Approach Questionnaire for client knowledge of the reha-<br />
B.<br />
process.<br />
bilitation<br />
way to measure cllent knowledge of the rehabilltation process is<br />
One<br />
administer a questionnalre to new clients that have been accepted for<br />
to<br />
prior to a speclal orientation program and again after the spe-<br />
services<br />
orientation program<br />
cial<br />
Some typical questions that might be included are<br />
Rehabilitation services are provided only when a client has a<br />
1.<br />
that creates a handicap to his belng employable.<br />
disability<br />
T<br />
F<br />
Cllents who qualify for rehabllitation servlces automatically<br />
2.<br />
eligible for welfare.<br />
become<br />
T<br />
F<br />
G. A. Statistical analysis in psychology and education.<br />
3SFerguson,<br />
York- McGraw-Hill, 1966, pp. 167-169.<br />
New<br />
-152-
WORK SHEET
One service every person recelves from the rehablltaton program<br />
3.<br />
vocational training.<br />
is<br />
1. etc.<br />
C. Statlstical Procedures<br />
T<br />
F<br />
score each test given to clients making sure the pre-measures<br />
st__e_El,<br />
post-measures are kept separate for each of the two groups.<br />
and<br />
2. Determine the change in score from the pre-measure to the<br />
Step<br />
for each group. For example<br />
post-measure<br />
Hypothetical Scores for Two Groups of 10 Clients<br />
Group Contrast Group<br />
Experimental<br />
change<br />
Client pre-score post-score score<br />
1 2 9 +7<br />
2 4 10 +6<br />
3 3 8 +5<br />
4 1 9 +8<br />
5 3 10 +7<br />
6 2 8 +6<br />
7 4 9 +5<br />
8 1 7 +6<br />
9 3 8 +5<br />
10 5 10 +5<br />
-154-<br />
Clent pre-score post-score<br />
I 3 6<br />
2 2 5<br />
3 4 8<br />
4 1 4<br />
5 2 6<br />
6 3 5<br />
7 3 7<br />
8 1 3<br />
9 2 5<br />
10 5 7<br />
change<br />
score<br />
+3<br />
+3<br />
+4<br />
+3<br />
+4<br />
+2<br />
+4<br />
+2<br />
+3<br />
+2
WORK SHEET
3. Calculate the mean of the "change" scores for each group.<br />
Step<br />
the mean score from the scores for each subject and square this<br />
Subtract<br />
Then sum the squared values. Using the data from this example,<br />
value.<br />
calculations are as follows-<br />
the<br />
Change<br />
Score<br />
Cllent<br />
Experimental Groups<br />
X1<br />
1 7 +1 1<br />
2 6 0 0<br />
3 5 -1 1<br />
4 8 +2 4<br />
5 7 +1 1<br />
6 6 0 0<br />
7 5 -1 1<br />
8 6 0 0<br />
9 5 -1 1<br />
10 5 -1 1<br />
N 10<br />
.X 60<br />
X 6<br />
xq<br />
10<br />
Change<br />
lent Score<br />
C1<br />
Contrast Group<br />
X2<br />
1 3<br />
2 3<br />
3 4<br />
4 3<br />
5<br />
6<br />
7<br />
8<br />
9<br />
10<br />
I0<br />
30<br />
3<br />
0 0<br />
0 0<br />
+1 1<br />
0 0<br />
+1 1<br />
-I 1<br />
+1 1<br />
-1 1<br />
0 0<br />
-1 1<br />
4. Substltute the data obtalned from Step 3 into the t-test<br />
Step<br />
for independent samples 36 (Independent because the measures are<br />
formula<br />
J. P. Fundamental statistics l_D_n psycholog), and education.<br />
36Guilford,<br />
York McGraw-Hill, 1950, p. 228.<br />
New<br />
-156-
WORK SHEET
two dlfferent groups) and do the necessary calculatlons. The t-test<br />
for<br />
is as follows-<br />
formula<br />
degrees of freedom<br />
df NI + N2 2<br />
XI X2 directs us to subtract the mean change<br />
wherefor<br />
the experimental group from the<br />
score<br />
mean change score of the contrast group.<br />
for the data In ths example, the t-test formula would look<br />
Thus,<br />
this"<br />
like<br />
t=<br />
6-3<br />
+10-2<br />
3<br />
r/ 89 (.20)'<br />
-158-
WORK SHEET
3<br />
178<br />
df=N+N-2<br />
3<br />
10 + 10 2<br />
18<br />
a table of critlcal values for t-tests wlth 18 degrees of free-<br />
Uslng<br />
the necessary t-value for significance at the .05 probabllity level<br />
dom,<br />
2 10 This means only 5 times out of 100 could we expect a t of 2.10<br />
is<br />
greater to happen by chance alone.<br />
or<br />
IV Interpretatl on<br />
can be observed by the obtained t value, sgnlfcance was reached<br />
As<br />
the 05 probab111ty level because the obtained t value exceeded the<br />
at<br />
value of 2.10. The mean change from pretest to posttest for<br />
critical<br />
experimental group was +6, the mean change for the contrast group was<br />
the<br />
In thls fictitious example, this indicates that the clients who were<br />
+3.<br />
assigned to the special orientation program scored significantly<br />
randomly<br />
in knowledge of the rehabilitation process than cllents who were<br />
higher<br />
nvolved n the orientation program.<br />
not<br />
-160-
EXAMPLE 9<br />
General Procedure: Pretest posttest control<br />
A.<br />
study deslgn.<br />
group<br />
Measurement Approach: The Work Adjustment<br />
B.<br />
Form.<br />
Rat±ng<br />
C. Stat±st±cal Procedures: Analysls of covar±ance.
Ao<br />
Bo<br />
EXAMPLE 9<br />
Procedure: Pretest posttest control<br />
General<br />
study deslgn.<br />
group<br />
Approach: The Work Aduustment<br />
Measurement<br />
Form.<br />
Ratlng<br />
C. Statlstcal Procedures: Analyss of covarance.
I Sample Evaluation Objective<br />
compare the effectiveness of three different approaches to work<br />
To<br />
trainlng In a rehabilitation facility.<br />
adjustment<br />
I I Crl teria<br />
delivery effectiveness of three different approaches to work<br />
Service<br />
training as measured by the Work Adjustment Rating Form.<br />
adjustment<br />
A. General Procedure 37<br />
Ill Methodology<br />
Pretest posttest control group study design<br />
R random asslgnment of<br />
where<br />
to experimental<br />
clients<br />
and contrast groups<br />
X = service/treatment<br />
0 observation/measurement<br />
D. T. & Stanley, J. C. Experimental and quasi-experimental<br />
37Campbell,<br />
for research. Chicago. Rand McNally, 1966.<br />
designs<br />
-163-
Steps.<br />
Randomly asslgn, if possible, clients to one of three dif-<br />
1.<br />
combinations of work adjustment training methodologies using<br />
ferent<br />
of the random sampling methods described in the Introduction of<br />
one<br />
manual. Three examples might be (a) productlon workshop assign-<br />
thls<br />
wth supportive counsellng, (b_] job-site assignment with com-<br />
ment<br />
employers and supportive counseling, and (c) a program nvolv-<br />
munity<br />
the workshop, job-sites, classroom nstruction and supportive<br />
ing<br />
ing.<br />
counsel<br />
Obtain Work Adjustment Rating Form ratings from two to four<br />
faclliy/job-site 2<br />
who are In a pion to observe the clients<br />
staff<br />
a work situation. These ratings are obtained for all clients n<br />
in<br />
three work adjustment trainlng groups during the first week in<br />
the<br />
the program (pretest).<br />
3. Carry out the work adjustment training programs.<br />
Again, obtain Work Adjustment Rating Form ratings from two<br />
4.<br />
four facility/job-siTstaff for all clien Tn the three work<br />
to<br />
training groups during the last week n the program<br />
adjustment<br />
(posttest).<br />
Tabulate results and conduct the necessary statistics to<br />
5.<br />
the results meaningful. One possibility is an analysis of<br />
make<br />
covariance 3B<br />
B Measurement Approach Work Adjustment Rating Form (WARF____) 3<br />
Work Adjustment Rating Form is a rating scale constructed pri-<br />
The<br />
for use by rehabllitation facility staff to assess work adjustment<br />
marily<br />
and limitations. It can be used to develop work adjustment<br />
strengths<br />
plans and assess progress toward job readiness.<br />
training<br />
WARF contains eight subscales, each having five Items, making a<br />
The<br />
of 40 i----tems The subscales are (a_) Amount of Supervision Requlred,<br />
total<br />
Realism of (b) Jo (c) Teamwork, (d_---) Acceptance of Rules/Authority,<br />
(e_---)<br />
Goals,<br />
Tolerance, (f__) PerSeverance n Wrk, (g) Extent Trainee Seeks<br />
Work<br />
Assistance, and (h) Importance Attached to Job Tralnlng.<br />
E. F. Deslgn and analysis of experiments in psychology<br />
3eLindquist,<br />
education. Boston Houghton Mifflin, '1953, pp. 317-339T--<br />
and<br />
J. A. & Bolanovich, D. J. WARF. A scale for measuring<br />
gBitter,<br />
behaviors. American Journal of Mental Deficiency, 1970,<br />
job-readiness<br />
616-621.<br />
74,<br />
-164-
WORK SHEET
of the subscales is represented by items descrlbing flve df-<br />
Each<br />
levels of performance from low to high. The following five items<br />
ferent<br />
which represent the subscale "Teamwork" are an example of this-<br />
1. Trainee is unable to work effectively wth any others.<br />
Trainee can work wlth only one or two others whom he particularly<br />
2.<br />
kes.<br />
1<br />
3. Trainee can usually work with a few others without conflct.<br />
4. Trainee works effectively in small (2-3) groups.<br />
5. Trainee is a good team worker in any group sltuatlon.<br />
items are rated by checklng "yes" or "no" to each and they are<br />
The<br />
(Guttman, 1944) so that a posltve response to an item at any level<br />
scaled<br />
also give a positive response to all items below that level 4.<br />
should<br />
are scrambled and the level and scale of items are not known by<br />
They<br />
raters.<br />
WAR____F can be completed n 3 to 7 minutes depending on the rater<br />
The<br />
instructons for completien of the WARF are as follows-<br />
The<br />
Work Adjustment Rating __Form s a rating scale consisting<br />
"The<br />
40 items relating to observable work behaviors You are<br />
of<br />
to answer each item either 'yes' or 'no' for the trainee<br />
asked<br />
on the form. Some Items appear very slmllar. An-<br />
identified<br />
swer each item. Please do not omit any items.<br />
Mark your response with an 'X' over your choice."<br />
for the WARF consists of cumulative positive responses. A<br />
Scoring<br />
s a positive response for the items A3, A4, A5, B3, B4, B5, C3,<br />
"yes"<br />
C5, D3, D4, D5, E3, E4, E5, F3, F4, F5, G4, G5, H4, and H5. A "no"<br />
C4,<br />
a positive response for tems A1, A2, BI, B2, C1, C2, D1, D2, El, E2,<br />
s<br />
F1, F2, G1, G2, G3, H1, H2, and H3.<br />
if a key template s constructed, takes approximately fve<br />
Scoring,<br />
The range of total scores on the WARF is zero to 40. A profile<br />
mnutes.<br />
trainee strengths and llmtatons can be obtained by nspection of<br />
of<br />
subscale scores. A scorlng form wlth dentfication of scrambled<br />
WARF<br />
L A. A basis for scallng qualtatve data. American<br />
4OGuttman,<br />
Review, 1944, 9, 139-150.<br />
Sociological<br />
-166-
WORK SHEET
items in parentheses Is included wth the scale (a WARF appears at<br />
WARF<br />
back of this example) It is advlsable to obtaln more than one rater's<br />
the<br />
ratings for each client to effect control for rater bias 41<br />
C. Statistical Procedures<br />
i__. Score each WARF making sure the pre-measures and post-<br />
Step<br />
are kept separate for each of the three groups<br />
measures<br />
2. Tabulate each rater's total WARF score for pre and post<br />
Step<br />
for the three different approaches (methods) to work adjustment<br />
ratings<br />
Use the average score of the two to four raters for each cli-<br />
training<br />
Hypothetical average total scores for 9 clients are glven below<br />
ent.<br />
1.<br />
Method<br />
on Workshop<br />
Product<br />
2<br />
Method<br />
tes<br />
Job-si<br />
3<br />
Method<br />
Job-sl tes,<br />
Workshop,<br />
C l as s room<br />
Pre (X) pos.t. Pre (x) Pos.t Pre.(X) Post<br />
17 24 18 22 17 38<br />
20 25 21 29 18 38<br />
20 23 21 27 19 35<br />
Calculate the prellmlnary nformaton that Is needed to<br />
st__9_E3.<br />
out the analysis of covariance. Analysis of covarance s a method<br />
carry<br />
statistically controlling for differences on some variable (n ths<br />
of<br />
the pretest) by adjusting the criterion measure (the posttest n<br />
case<br />
example) Ths technique is partlcularly useful when subjects can-<br />
this<br />
be randomly assigned to groups so that differences between groups<br />
not<br />
be minimized In this fictitious example, subjects were randomly<br />
may<br />
to groups, thus, analysis of covarlance wll further ncrease<br />
assigned<br />
precision of the study<br />
the<br />
Calculations and explanations follow<br />
J. A. Bas effect on valldlty and relability of a rating<br />
41Bitter,<br />
Measurement and Evaluation _n Guidance, 1970, 3(2), 70-75<br />
scale.<br />
-168-
WORK SHEET
a<br />
TX<br />
Ty#<br />
TxaTY a<br />
.X#Y a<br />
Method 1 Method 2 Method 3<br />
X Y X Y X Y<br />
17 24 18 22 17 38<br />
20 25 21 29 18 38<br />
20 23 21 27 19 35<br />
3 3 3<br />
57 60 54<br />
19<br />
1083<br />
1089<br />
1368<br />
1368<br />
24<br />
1728<br />
1730<br />
72 78 111<br />
20 18<br />
26<br />
1200 972<br />
1560<br />
2028<br />
1206 974<br />
1572<br />
2054<br />
-170-<br />
1998<br />
1995<br />
37<br />
4107<br />
4113<br />
N=9<br />
T X<br />
171<br />
261<br />
Ty<br />
3249<br />
Tx2/N<br />
7569<br />
Ty2/N<br />
4959<br />
TxTy/N<br />
X X<br />
19<br />
Xy = 29<br />
.<br />
s<br />
na<br />
TxaTy a<br />
3255<br />
7863<br />
2 3269<br />
zzXj<br />
7897<br />
zzYa2<br />
r.r.X#Y,] 4935<br />
4926
WORK SHEET
where the number of cllents in each column<br />
N total number of cllents in the evaluatlon<br />
total of the WARF scores in each pretest column (e g<br />
for Method 1 T"# 17 + 20 + 20 57)<br />
TX#<br />
X total of the WARF pretest scores for all three methods<br />
T<br />
T X 570 + 54 171)<br />
(e.g.<br />
total of the WARF scores in each posttest column (e.g.<br />
for Method 1 ---# 24 + 25 + 23 72)<br />
Ty#<br />
total of the WARF posttest scores for all three methods<br />
Ty<br />
Ty 728 + 111 261).<br />
(e.g.<br />
the prewously calculated T X value squared dlvlded by<br />
Tx2/N<br />
total number of clents<br />
the<br />
Tx2/N 171 x 171 29,241<br />
9 9 3249)<br />
(e.g.<br />
the prewously calculated Ty value squared dlvided by<br />
Ty2/N<br />
total number of clents<br />
the<br />
g. Ty2/N 261 x 261 68,121<br />
9 9 7569)<br />
(e<br />
the prevlously calculated TX value multlplied by the<br />
TxTy/N<br />
calculated Ty value dvided by the total<br />
prewously<br />
of clients.<br />
number<br />
TxTy/N 171 x9 261 44,6319 4959)<br />
(e.g.<br />
TX<br />
XX n#<br />
or the average of the WARF pretest scores for each<br />
method (e.g. 57<br />
Method 1. XX = 19)<br />
for<br />
172-
WORK SHEET
Xy =<br />
average of the WARF pretest scores for all methods<br />
the<br />
+ 20<br />
(e.g. X<br />
19<br />
3 = 19)<br />
X<br />
or the average of the WARF posttest scores for<br />
each method (e.g. for Method 1 _vj 7--23 24)<br />
average of the WARF posttest scores for all methods<br />
the<br />
X- 24 + 26<br />
(e.g<br />
3 = 29)<br />
= take the value of TXj for each method (pretest) and<br />
multiply it by its own value (square it). Then dlwde<br />
by nj for each column.<br />
(e.g. for Method 1<br />
57 x 57 3249<br />
TX<br />
3 = 1083)<br />
n<br />
. n---_ take the three values calculated for -- and add them<br />
together.<br />
the value of Tyj for each method (posttest) and<br />
take<br />
it by Its own value (square t). The dvde<br />
multiply<br />
for each column.<br />
by<br />
2 T<br />
(e.g. for Method I 72 x 72 5184<br />
n 3 1728)<br />
T 2<br />
TY<br />
take the three values calculated for and add them<br />
nj<br />
z<br />
nj<br />
-174-
WORK SHEET
.<br />
Tx"Ty<br />
total of the WARF scores in each pretest column<br />
the<br />
by its corresponding total for WARF scores<br />
multlplled<br />
the posttest column and diwded by the er of<br />
in<br />
for that method.<br />
clients<br />
(e.g. for Method 1<br />
TxjTy#. 57 x 72 4104<br />
nj 3 - 1368)<br />
the three values calculated for<br />
TXJTY-#<br />
and add<br />
take<br />
nj<br />
nj<br />
together.<br />
them<br />
.XY =<br />
each WARF pretest score and sum for each column<br />
square<br />
for Me-- 1- sX 2 (17)(17) + (20)(20) +<br />
(e.g.<br />
(20)(20) = 1089)<br />
the sum of all three .X# 2 (e.g. 1089 + 1206 + 974 3269).<br />
square each WARF posttest score and sum for each column.<br />
sY2<br />
for Met----1 sY2 (24)(24) + (25)(25) +<br />
(e.g.<br />
= 1730)<br />
(23)(23)<br />
the sum of all three sYj (e.g. 1730 + 2054 + 4113 7897).<br />
X (pretest) score multlplied by Its corresponding<br />
each<br />
(posttest) score and summed for each method column.<br />
Y<br />
for Method 1" XY (17)(24) + (20)(25) +<br />
(e.g.<br />
1368)<br />
(20)(23)<br />
..XjY# the sum of all three sX#Y (e g. 1368 + 1572 + 1995 4935)<br />
Using the values calculated in Step 3, It is posslble to<br />
St__p_4<br />
sum of squares (a) total (SST), (b_) between (SSB) and (c)<br />
determine<br />
(SS W) for both pretest (X) and posttest (Y) scores Also, the (a)<br />
wthn<br />
of products (SP), (b_) adjusted sum of squares (SSV), and (c_) adjusted<br />
sum<br />
squares (MSV} can be computed. It is the adjusted mean squares for<br />
mean<br />
(MS#B) and within groups methods (MS which are mportant for calculating<br />
B)<br />
F value. The F-ratio tests the significance of the differences among<br />
the<br />
method means.<br />
the<br />
-176-
WORK SHEET
detemine the sum of squares total, between, and within, apply<br />
To<br />
following formulas-<br />
the<br />
of Squares Total (SSTx) XzXa2 PTx.._<br />
Sum<br />
N<br />
Degrees of freedom = N 1<br />
:9-1<br />
=8<br />
Sum of Squares Between (SSBx) x:<br />
Degrees of freedom k- 1<br />
=3-1<br />
=2<br />
-178-<br />
= 3269 3249<br />
N total number<br />
wheresubjects<br />
of<br />
TX2<br />
TX<br />
N<br />
a<br />
3255- 3249<br />
where number of groups
WORK SHEET
Sum of Squares With1"n (SSWx) =<br />
Degrees of freedom N /<br />
=9-3<br />
:6<br />
SSTx<br />
SSBx<br />
=20-6<br />
SCmlarly, for posttest WARF scores (Y):<br />
of Squares Total (SSTy) ssyj2 TY__<br />
Sum<br />
N<br />
Sum of Squares Between (SSBy) s<br />
-180-<br />
7897- 7569<br />
N = total number<br />
wheresubjects<br />
of<br />
Ty____<br />
Tv2--<br />
N<br />
nj<br />
7863- 7569<br />
number of groups
WORK SHEET
Sum of Squares Within (SSWy) SSTy SSBy<br />
Then, calculate the sum of products-<br />
Sum of Products Total (SP T) zzXjYj<br />
Sum of Products Between (SP B) :<br />
328 294<br />
4935 4959<br />
z<br />
TxYy<br />
N<br />
T X Ty TxT Y<br />
= 4926 4959<br />
Sum of Products W1thln (SP W) SP T SP B<br />
-182-<br />
(-24) (-33)<br />
N
WORK SHEET
using the prevlous values, calculate the adjusted sum of squares<br />
Now,<br />
posttest WARF scores (SS).<br />
for<br />
Sum of Squares Withln (SSy) SSWy<br />
Degrees of freedom N /¢ 1<br />
=9-3-1<br />
=5<br />
Sum of Squares Total (SSTy) =<br />
Degrees of freedom N 2<br />
=9 -2<br />
=7<br />
-184-<br />
SSTy<br />
(SPw)2<br />
X<br />
SSw<br />
34- (9)(9)<br />
14<br />
34<br />
81<br />
14<br />
34- 5.8<br />
12 8-2<br />
N total number<br />
wheresubjects<br />
of<br />
(SPT)2<br />
.STy<br />
328<br />
328<br />
(-24)(-24)<br />
20<br />
576<br />
20<br />
328 28.8<br />
i299.2<br />
k = number of groups<br />
N total number<br />
where:<br />
subjects<br />
of
WORK SHEET
Sum of Squares Between (SSy) SSy- SSy<br />
Degrees of freedom k 1<br />
=3-1<br />
=2<br />
299.2 28.2<br />
271.0<br />
where, k number of groups<br />
results of these computations can be summarlzed in a table as<br />
The<br />
lows.<br />
fol<br />
Sources df SS x SSy SP SSy df MSy<br />
Between<br />
Within<br />
2 6 294 -33 271.0 2 135.50<br />
6 14 34 9 28.2 5 5.64<br />
Total 8 20 328 -24 299.2 7<br />
adjusted mean squares (MS) for treatment and wthln groups are<br />
The<br />
by dividing the adjusted sum of squares (SSy) by the correspond-<br />
determined<br />
Ing degrees of freedom (dr) Thus,<br />
B<br />
SSy<br />
df<br />
MSYB<br />
-186-<br />
271.0<br />
2<br />
1135.50
WORK SHEET
W SSy<br />
MSYw_ and<br />
df<br />
=I5,.641<br />
Step 5_. Calculate the F-ratlo for analysis of covamance.<br />
F<br />
Square between 42<br />
Mean<br />
Square within<br />
Mean<br />
135.50<br />
5.64<br />
I'24 02<br />
obtained F value is 24.02 Using a table of crltcal values for<br />
The<br />
tests (found in statistics books) for 2 and 5 degrees of freedom, the<br />
F<br />
F value is 5.79 at the .05 probabllity level. This means that<br />
necessary<br />
5 times out of 100 would we expect a F of 5 79 or greater to happen<br />
only<br />
by chance alone.<br />
Is now known from the sgnlficant F value (24.02) that there are<br />
It<br />
differences between the three methods. However, a significant<br />
sgnlfcant<br />
doesn't indicate where the differences are among the three methods.<br />
F<br />
it is necessary to conduct tests between each par of groups. Ths<br />
Thus,<br />
would not be necessary, of course, if a significant F was not obtained<br />
6. To test dfferences for ndlvdual pairs of groups, It is<br />
Step<br />
to compute the adjusted means for methods. This adjustment Is<br />
necessary<br />
pretest differences between groups First, determine the wlthin-group<br />
for<br />
coefficient (b) as follows<br />
regression<br />
E. F. Design and analysis of experiments in psychology<br />
42Lindquist,<br />
education. Boston Houghton Mifflin, 1953, pp. 317-339'-<br />
and<br />
-188-
WORK SHEET
determine the adjusted posttest means for each group uslng the<br />
Next,<br />
formula-<br />
following<br />
Xy Xy (x Tx)<br />
Thus, for Method 1 the adjusted posttest mean is<br />
Y;a 24- .64(19 19)<br />
9<br />
24- 64(0)<br />
=24-0<br />
For Method 2 the adjusted posttest mean is<br />
Xya 26 -.64(20 19)<br />
26 .64(1)<br />
26 .64<br />
25.36<br />
-190-
WORK SHEET
Slmllarly, for Method 3 the adjusted posttest mean Is.<br />
yj 37 64(18- 19)<br />
37 64(-1)<br />
1'37.6'4<br />
detemne the error variance (o 2) for each of the comparisons<br />
Then,<br />
two groups. The fomula s-<br />
between<br />
2 :<br />
1<br />
+ +<br />
(<br />
S -wx )<br />
The error varlance for comparing Method 1 wth Method 2 is as follows<br />
: [.73]<br />
.64<br />
-192-<br />
MSWy
WORK SHEET
Next, test the dlfference by<br />
24- 25.36<br />
I,..67<br />
other two comparlsons are conducted in the same manner wlth the<br />
The<br />
ng results-<br />
follow<br />
t<br />
Comparison<br />
II -.67<br />
I,<br />
I, III 6.72<br />
II, III 5 32<br />
a t table of critical values (found in statstlcs book) ndl-<br />
Uslng<br />
that a t value of 2.78 or greater Is necessary with 4 degrees of<br />
cates<br />
freedom (df n I<br />
+ n2 2) at the .05 probabillty level Thus, the df-<br />
between Methods 1 and 3 (t = 6.72) and between Methods 2 and 3<br />
ferences<br />
5.32) are sgnfcant.<br />
(t<br />
-194-
WORK SHEET
IV Interpretatl on<br />
can be observed by the obtalned F value of 24 02 for three dlf-<br />
As<br />
approaches to work adjustment tralnng in a rehabilitation facility,<br />
ferent<br />
was reached at the .05 probability level because the obtained<br />
sgnlflcance<br />
value exceeded 5.79. This indicates that differences exist among the<br />
F<br />
posttest means for the three methods after adjustment for pretest<br />
WARF<br />
To determine where these differences are among the three<br />
dfferences.<br />
the evaluator must compare n pars the adjusted WARF posttest<br />
methods,<br />
using a t-test. These comparisons resulted in signlficant differ-<br />
means<br />
between Methods 1 and 3, and between Methods 2 and 3, but no sig-<br />
ences<br />
difference between Methods 1 and 2. Again, the adjusted WARF<br />
nificant<br />
posttest means were<br />
1 (productlon workshop assignment wlth 24.00<br />
Method<br />
ve counsel ng)<br />
supportl<br />
2 (job-slte assignments with co,,nunity 25 36<br />
Method<br />
and supportive counseling)<br />
employers<br />
3 (workshop, job-sltes, classroom in- 37.64<br />
Method<br />
and supportive counseli}g)<br />
structon<br />
Method 3 involwng production workshop assignment, job-ste assign-<br />
Thus,<br />
with community employers, classroom nstructon and supportive coun-<br />
ments<br />
was sCgnificantly more effective in mprowng work adjustment of<br />
sellng<br />
as measured by the WARF than Methods 1 and 2<br />
clients<br />
-196-
TRAINEE<br />
RATER<br />
WORK ADJUSTMENT RATIN8 FORM<br />
Sub-scale<br />
Amount of Supervasaon<br />
A.<br />
Requared<br />
B. Realasm of Job Goals<br />
C. Teamwork<br />
Acceptance of<br />
D.<br />
Rules/Authoraty<br />
E. Work Tolerance<br />
F. Perseverance an Work<br />
Extent Traanee Seeks<br />
G.<br />
Assastance<br />
Importance Attached<br />
H.<br />
Job Tralnang<br />
to<br />
James A. Batter and D. J. Bolanovach<br />
SCORING FORM<br />
(25)<br />
(34)<br />
(19)<br />
(12)<br />
Level<br />
DATE<br />
1 2 3 5<br />
(5)<br />
(38)<br />
(23)<br />
(32)<br />
(33)<br />
(10)<br />
(3)<br />
(36)<br />
(21)<br />
(6)<br />
(31)<br />
(16)<br />
C1)<br />
(18)<br />
(35)<br />
(28)<br />
(13)<br />
(30)<br />
(39)<br />
(8)<br />
(17)<br />
(26)<br />
(ii)<br />
C4)<br />
(29)<br />
Note: Numbers an parentheses andcate atem numbers.<br />
(22)<br />
(15)<br />
(24)<br />
(9)<br />
(2)<br />
(27)<br />
(20)<br />
(37)<br />
(14)<br />
(7)<br />
(40)<br />
Total<br />
Total
INSTRUCT IONS :<br />
Work Adlustment Ratanq Form as a rating scale consastang<br />
The<br />
40 atems relatang to observable work behavlors. You are asked<br />
of<br />
answer each atem eather "yes" or "no" for the traanee adenta-<br />
to<br />
on the form. Some atems appear very samlar. Answer each<br />
fled<br />
atem. Please do not omlt any atems.<br />
l•<br />
i0.<br />
13.<br />
Mark your response wath an "X" over your choace.<br />
traanang and darectaon, traanee can work<br />
Wath<br />
under occasaonal supervasaon.<br />
andependently<br />
has developed realastac 3ob goals and<br />
Traanee<br />
seeks counsel an plannang.<br />
readaly<br />
can work wath only one or two others<br />
Traanee<br />
he partacularly lakes.<br />
whom<br />
knows and usually follows rules<br />
Traanee<br />
remlnder.<br />
wlthout<br />
becomes frustrated and gaves up<br />
Traanee<br />
on almost any 3ob-<br />
easaly<br />
generally stays at work but as<br />
Traanee<br />
dastracted and loses anterest.<br />
easaly<br />
seeks assastance only on bonafade<br />
Traanee<br />
and after attemptang to solve<br />
problems<br />
them hamself.<br />
desares work but does not do anythang<br />
Traanee<br />
to land at.<br />
hamself<br />
catches on easaly and does has work<br />
Traanee<br />
practacally no supervsaon.<br />
wlth<br />
consaders 3ob plans but they are not<br />
Traanee<br />
wth has abalataes.<br />
compatable<br />
Tralnee works effectively an small (2-3) groups.<br />
shows open hostalaty to authoraty<br />
Tralnee<br />
rules.<br />
and<br />
generally works at routine 3obs<br />
Trainee<br />
wthout resistance.<br />
readaly<br />
Y N<br />
Y N<br />
Y N<br />
Y N<br />
Y N<br />
Y N<br />
Y N<br />
Y<br />
Y<br />
Y<br />
N<br />
N
Traanee as a persastent worker on all assagn-<br />
14.<br />
even under adverse carcumstances.<br />
ments,<br />
Traanee generally handles own problems wlth<br />
15.<br />
occasaonal help.<br />
only<br />
Traanee expresses anterest an future work,<br />
16.<br />
not an a 3ob now.<br />
but<br />
Once shown what he must do, traanee applaes<br />
17.<br />
dalagently wlthout much supervlsaon.<br />
hamself<br />
Traanee has begun to thank about possible<br />
18.<br />
for hamself that are wathan has<br />
occupataons<br />
capabalatles.<br />
Trainee as unable to work effectavely wth<br />
19.<br />
others.<br />
any<br />
Traanee understands rules and regulations<br />
20.<br />
adheres to them conslstently.<br />
and<br />
Traanee traes sampler 3obs but usually becomes<br />
21.<br />
when he encounters changes an<br />
dascouraged<br />
routlne.<br />
Traanee applles hamself dalagently to almost<br />
22.<br />
kands of work.<br />
all<br />
Trainee frequently seeks help and attentaon<br />
23.<br />
personal as well as work-related problems.<br />
for<br />
Trainee wants a 3ob and seeks asslstance an<br />
24.<br />
to prepare for one.<br />
tryang<br />
Traanang works wath daffaculty, even under<br />
25.<br />
supervasaon and after gettang<br />
constant<br />
considerable traanang.<br />
Traanee accepts guadance n plannang for<br />
26.<br />
3obs, reallzng has lamltataons.<br />
future<br />
Traanee as a good team worker an any group<br />
27.<br />
a tuata on.<br />
s<br />
Traanee generally accepts anstructaons and<br />
28.<br />
but requares occasaonal rent%nder.<br />
rules<br />
Trainee gets personal satsfactaon from most<br />
29.<br />
gaven hlm.<br />
3obs<br />
Y<br />
Y<br />
Y<br />
Y<br />
Y<br />
Y<br />
Y<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N
32.<br />
33.<br />
34.<br />
35.<br />
36.<br />
37.<br />
38.<br />
39.<br />
40.<br />
Tralnee usually works steadaly unless dasturbed.<br />
seeks assastance In manor problems<br />
Tralnee<br />
could solve hamself.<br />
he<br />
is not nterested an worklng now or<br />
Trainee<br />
n the future.<br />
•<br />
can work on hs own after thorough<br />
Trainee<br />
f hs work s frequently observed<br />
tranlng<br />
and checked.<br />
does not have plans for a 3ob and as<br />
Tralnee<br />
ready to consider any.<br />
not<br />
can usually work wlth a few others<br />
Traanee<br />
conflact.<br />
wlthout<br />
delaberately and knowangly vlolates<br />
Tralnee<br />
and instructaons on occasaons.<br />
rules<br />
s quate flexlble and enjoys any knd<br />
Tralnee<br />
work.<br />
of<br />
needs frequent relaef "breaks", cannot<br />
Tralnee<br />
at even sample tasks.<br />
keep<br />
Trainee often needs help and asks for t.<br />
makes active effort to seek a 3ob<br />
Trainee<br />
hmself.<br />
for<br />
Work Ad]ustment Ratang Form is covered by Copyright 1969 by<br />
The<br />
A. Btter.<br />
James<br />
Y<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N<br />
N
PART FOUR<br />
Solomon Four-Group Study Deslgn<br />
0 X 0<br />
1<br />
R<br />
2<br />
0 0<br />
3<br />
R<br />
4<br />
R X 0 5<br />
R 0 6<br />
R random ass±gnment<br />
where:<br />
sub3ects<br />
of<br />
X<br />
service/treatment<br />
observation/measure-<br />
01<br />
before treat-<br />
ment<br />
to one experl-<br />
ment<br />
0 2<br />
0 3<br />
mental group<br />
observation/measureafter<br />
treatment<br />
ment<br />
one exper±mental<br />
to<br />
group<br />
observation/measurebefore<br />
treatment<br />
ment<br />
one contrast group<br />
for<br />
treatment)<br />
(no<br />
0 4<br />
observation<br />
treatment for one<br />
after<br />
group (no treat-<br />
contrast<br />
ment)<br />
observat±on/measurement<br />
0<br />
after<br />
5<br />
to one<br />
treatment<br />
0 6<br />
experlmental group<br />
observation/measurement<br />
treatment for one<br />
after<br />
group (no treat-<br />
contrast<br />
ment)<br />
des±gn calls for a group of sub3ects that are randomly<br />
Th±s<br />
to one of four groups. These groups are then examined<br />
ass±gned<br />
four dlfferent cond±tlons. The post measures of the four<br />
under<br />
are used to determ±ne the effects of pretests and treat-<br />
groups<br />
ment upon the sub3ects.<br />
The evaluator dentfes a sample of sub3ects. From<br />
Method.<br />
sample he randomly assigns the members to one of four groups.<br />
ths<br />
groups recelve a posttest. Groups one and two are also<br />
All<br />
a pretest (0). Groups one and three are the only ones to<br />
g±ven<br />
the treatment (X). After the treatment has been con-<br />
receive<br />
and the posttest g±ven to the four groups, the posttest<br />
cluded<br />
scores are compared.<br />
Ths des±gn, lke the pretest posttest control group<br />
Value.<br />
design, involves random ass±gnment of sub3ects to groups;<br />
study<br />
permitting the assumption of equal groups on all possible<br />
thus,<br />
It also has experimental and contrast groups<br />
characteristics.<br />
comparison. The advantage of the Solomon Four-Group deslgn<br />
for<br />
its control for the effects of pretesting.<br />
is
EXAMPLE 10<br />
A. General Procedures: Solomon four-group study des±gn.<br />
Measurement Approach: Test of counselor knowledge<br />
B.<br />
placement methods and resources.<br />
of<br />
Statlstlcal Procedures: 2 x 2 factorlal analyss<br />
C.<br />
varlance.<br />
of
EXAMPLE i0<br />
A. General Procedures: Solomon four-group study design.<br />
Measurement Approach: Test of counselor knowledge<br />
B.<br />
placement methods and resources.<br />
of<br />
Statlstcal Procedures: 2 x 2 factorial analyss<br />
C.<br />
variance.<br />
of
Sample Evaluatlon Objectlve<br />
determine the effectiveness of in-servlce tralning on counselors'<br />
To<br />
of placement methods and resources.<br />
knowledge<br />
II Criteria<br />
in posttest scores relative to counselor knowledge of<br />
Difference<br />
methods and resources by counselors involved in in-service traln-<br />
placement<br />
ing and counselors not involved in the training.<br />
A. General Procedure 3<br />
Solomon four-group study design<br />
III Methodology<br />
RO I<br />
XO 2<br />
R 03 04<br />
R XO s<br />
R OB<br />
R = random assignment of<br />
where<br />
to two exper-<br />
subjects<br />
and two contrast<br />
mental<br />
groups<br />
X service/treatment<br />
0 observatlon/measurement<br />
D. T. & Stanley, J. C. Experimental and quasi-experimental<br />
W3Campbell,<br />
for research. Chicago: Rand McNally, 196'6.<br />
designs<br />
-205-
Steps.<br />
Identify counselors to be included in the study<br />
1.<br />
them to one of the four groups.<br />
assign<br />
2. Administer a pretest to groups 1 and 2.<br />
Randomly<br />
3. Give in-service training (treatment) to groups 1 and 3.<br />
Administer posttest to all four groups when in-service traln-<br />
4<br />
has been completed.<br />
ing<br />
Tabulate the results for posttest analysis and apply statls-<br />
5.<br />
procedures. For this example, a 2 x 2 factorial analysls of<br />
tical<br />
is illustrated , and only the effects of treatment and pre-<br />
variance<br />
(main effects) are studied.<br />
testing<br />
Measurement Approach- Test of counselor knowledge of placement methods<br />
B.<br />
resources<br />
and<br />
way to measure the counselor's knowledge of placement methods<br />
One<br />
resources is to construct a test. For this example, a fictitious set<br />
and<br />
scores for four groups of 5 counselors each is used. Two groups re-<br />
of<br />
in-service training in placement methods and resources and two<br />
ceived<br />
did not recieve the in-service training. A knowledge test con-<br />
groups<br />
of 10 items (allowing scores from 0 10) was given to two groups<br />
slstlng<br />
a pretest and all four groups were given the posttest. As the pretest<br />
as<br />
are not necessary to determine the affect of pretestlng, they are<br />
scores<br />
included in this example. The scores on the posttest for the 20 coun-<br />
not<br />
involved n the study are as follows:<br />
selors<br />
1 Group 2.. Gr..oup 3 Group .4.<br />
Group<br />
6 9 4<br />
10<br />
9 5 8 7<br />
9 5 10 6<br />
I0 6 I0 6<br />
8 7 8 5<br />
E. F. Design and a.nalxsis of experiments in psychology<br />
4Lindquist,<br />
education. Boston: Houghton Mifflin, 1953, pp. 108-120.<br />
and<br />
-206-
WORK SHEET
C. Statistical Procedures<br />
1_. Tabulate the results of the posttest scores in an analysls<br />
Step<br />
variance double-entry table and calculate the mean of each group, the<br />
of<br />
means, column means, and general mean. The mean for each group (cell)<br />
row<br />
determined by summing the scores and divldlng by the number of subjects,<br />
is<br />
group 1 sum is 46, the number of subjects is 5, therefore, the mean<br />
e.x.<br />
9.2. The means for rows and columns are done in the same manner. For<br />
is<br />
amp I e<br />
ex<br />
PRETEST<br />
Yes<br />
No<br />
I0<br />
9<br />
9<br />
10<br />
8<br />
9<br />
8<br />
10<br />
10<br />
8<br />
TREATMENT<br />
(In-serw ce<br />
Training)<br />
Row<br />
No Means<br />
Yes<br />
Group 1<br />
Xi=92<br />
sXI 46<br />
Group 3<br />
X3 9<br />
zX3 = 45<br />
Group 2<br />
X2 5.8<br />
.X2 29<br />
Group 4<br />
X4 = 5.6<br />
sX 28<br />
o I umn<br />
C<br />
X 9.1 X 5 7<br />
Means<br />
': mean (average)<br />
7.5<br />
row 1 group'<br />
of<br />
= mean (average<br />
7.3<br />
row 2 group,<br />
of<br />
4 General Mean for<br />
7<br />
scores<br />
all<br />
2. Calculate the preliminary information that is needed to<br />
Step<br />
out the analyss of data. This would look llke the following when<br />
carry<br />
up (explanations for each of the symbols and sample calculations ap-<br />
set<br />
on pages 210 214).<br />
pear<br />
-208-
WORK SHEET
Row 1 T. 46 29<br />
Col 1 Col 2 T<br />
426 171<br />
423.2 168.2<br />
Row 2 T
WORK SHEET
T# Is the sum of scores for each group, e g. for group<br />
where<br />
(column one, row one) T# 10 + 9 + 9 + 10 +<br />
one<br />
8:46<br />
is the sum across rows, e.g. for row one T 46 +<br />
TC<br />
75<br />
29<br />
instructs us to square the T value prevlously cal-<br />
T2/m<br />
and divide by the number of subjects across<br />
culated<br />
rows, e.g. for row one T2/m; 75 x 75 10 562 5<br />
2 instructs us to take each X in a given group, square<br />
zX<br />
value, and then sum the results, e.g. 10 2 + 92<br />
that<br />
+ 92 + 10 2 + 82 : 426<br />
us to square the Tj value already cal-<br />
nstructs<br />
and then divide by mi# which is the number<br />
culated<br />
subjects in a given group, e.g for group one<br />
T2/mj of<br />
x 46 5 423.2<br />
46<br />
Is the sum of scores for columns, e g. for column<br />
T#<br />
T 46 + 45 91<br />
one<br />
us to square the prewously obtained T<br />
nstructs<br />
and dwde by the number of subjects n the<br />
value<br />
e.g. for column one T2/m 91 x 91 10<br />
column,<br />
8281 10 828.1<br />
2 instructs us to add all of the zX previously cal-<br />
zz.X<br />
values for the four groups together, e.g.<br />
culated<br />
426 + 171 + 409 + 162 1168.0<br />
us to add the four values for zT..?2/m<br />
instructs<br />
e.g. 423.2 + 168.2 + 405 0 + 156 8 1153 2<br />
together,<br />
is the value obtained from adding the two values<br />
zT2/m<br />
TC2/m together, e.g 562.5 + 532 9 1095 4<br />
for<br />
-212-
WORK SHEET
is the value obtained from adding the two values for<br />
F#2/m#<br />
zTjZ/nj<br />
e.g 828.1 + 324.9 1153 0<br />
together,<br />
instructs us to flnd the total value of all scores<br />
T2/N<br />
together, square that value, and divlde by<br />
added<br />
the total number of subjects, e.g. 1482 20 = 1095.2<br />
3_. Substitute the information obtalned in the previous calcula-<br />
Step<br />
nto the following sum of squares formulas<br />
tlons<br />
SSTotal =<br />
SSR°ws<br />
r c ni<br />
. X 2 T2/N<br />
z<br />
#=I<br />
i=1<br />
r Ti/n i T2/N<br />
Z<br />
i=1<br />
T2/n# T2/N<br />
SSCol z<br />
#=1<br />
umns<br />
SSCell s<br />
r<br />
s Ti2/ni T2/N<br />
s<br />
#=1<br />
i=1<br />
To determlne the followlng, use the above sum of squares informatlon<br />
SSRows x Columns SSCelIs SSRows SSColumns<br />
SSwithln SSTota I SSCell s<br />
-214-
WORK SHEET
For thls example, the results would appear as follows-<br />
SS T<br />
SS R<br />
SS c<br />
1168.0- 1095.2 72.8<br />
1095.4 1095.2 .2<br />
1153.0 1095.2 57.8<br />
SSCell s =1153.2 1095.2 58.0<br />
SSRc 58.0 .2 57.8 0<br />
SS W<br />
72.8 58.0 14.8<br />
Complete a Summary Table by substituting the above Informa-<br />
St__p_4<br />
In the followlng format and conduct the necessary calculations to<br />
ton<br />
arrive at F values.<br />
of<br />
Source<br />
at on<br />
Var<br />
Columns (C)<br />
Rows<br />
(Cells)<br />
(R)<br />
x Columns<br />
Rows<br />
(RC)<br />
thn Cells<br />
W<br />
(w)<br />
Total<br />
df<br />
c-1<br />
r-1<br />
(rc-1)<br />
(r-l)(c-1)<br />
N-rc<br />
N-1<br />
Summary Table<br />
of Squares<br />
Sum<br />
(SS)<br />
SS c z T#21n# T21N<br />
R Z Ti2/n i T2/N<br />
SS<br />
,=1<br />
o<br />
SSCells<br />
1<br />
s T#2/m# T2/N<br />
s<br />
i=1 j=l<br />
SSRc SSCell s<br />
SS W<br />
SS T<br />
SS R<br />
SSCell s<br />
SS C<br />
z z z X 2 T2/N<br />
SST<br />
;=1 #=1<br />
-216-<br />
Square<br />
Mean<br />
MS<br />
C<br />
SS<br />
c-1<br />
R<br />
SS<br />
r-1<br />
SSRc<br />
(r-1)(c-1)<br />
W<br />
SS<br />
N-rc<br />
C<br />
MS<br />
W MS<br />
MS<br />
R<br />
MSw
WORK SHEET
The calculations for the degrees of freedom are as follows.<br />
c-1<br />
r-1<br />
rc-1<br />
represents the number of columns less 1 (2 I 1)<br />
represents the number of rows less I (2 1 1)<br />
the number of rows tlmes the number of columns<br />
represents<br />
1 (2 x 2 i 3)<br />
less<br />
represents the number of rows less 1 tlmes the number<br />
(r-1)(c-1)<br />
columns less 1 I(2 1)(2 I) 11<br />
of<br />
represents the total number of subjects less rows tlmes<br />
N-rc<br />
120 -_(2 x 2) 161<br />
columns<br />
represents the total number of subjects less 1<br />
N-1<br />
1 19)<br />
(20-<br />
Completed summary table using example nformaton-<br />
of<br />
Source<br />
Variation<br />
Col umns<br />
Rows<br />
Cells<br />
x"Col umns<br />
Rows'<br />
( I nteractl ons )<br />
df<br />
1<br />
of Squares<br />
Sum<br />
(SS)<br />
57.8<br />
2<br />
58 0<br />
MS<br />
57 8<br />
Wthn Cells 16 14 8 .9<br />
Total 19 72.8<br />
*9_
WORK SHEET
degrees of freedom for uslng a table of crltlcal values for F is<br />
The<br />
by the corresponding Sun=nary Table df used to calculate F. For<br />
determlned<br />
the F 64.22 was calculated from MS c (df 1) and MS W (df 16),<br />
example,<br />
the degrees of freedom used for looking up the critical value of F<br />
thus,<br />
1 and 16. The necessary F value for slgnificance at the .05 probability<br />
are<br />
is 4.49 This means that only 5 times out of 100 would we expect a<br />
level<br />
F of 4.49 or greater to happen by chance alone.<br />
IV Interpretation<br />
can be observed by the obtained F values, sgnlficance was reached<br />
As<br />
better than the .05 probability level for treatment (column) effects<br />
at<br />
F value for pretest (row) effects was not sgnficant at the 05 prob-<br />
The<br />
level. Recall that the mean posttest scores for counselors re-<br />
abllity<br />
training was 9.1, and for those not receiwng training was 5.7.<br />
ceivlng<br />
in this fictitious example, the n-service training significantly<br />
Thus,<br />
performance on the test of counselor knowledge of placement<br />
Improved<br />
and resources. Giving a pretest did not significantly influence<br />
methods<br />
scores. The mean posttest score for those receiving a pretest<br />
posttest<br />
7.5, for those not pretested it was 7.3.<br />
was<br />
-220-
REFERENCES<br />
J. R., Brown, F. L., & Mink, O. J. Statistical concepts: A basic<br />
Amos,<br />
New York- Harper and Row, 1965.<br />
program.<br />
E. C. & Weisinger, M. Program evaluation- A resource handbook<br />
Bennett,<br />
vocational rehabilitation. New York: ICD Rehabilitation and<br />
for<br />
Research Center, 1974.<br />
J. A. Bias effect on validity and reliabClity of a rating scale.<br />
Bitter,<br />
and Evaluation in Guidance, 1970, 3_(2), 70-75.<br />
Measurement<br />
J. A. & Bolanovich, D. J. WARF- A scale for measuring job-<br />
Bitter,<br />
behaviors. American Journal of Mental Deficlenc, 74(5),<br />
readiness<br />
616-621.<br />
W. R. & Merideth, D. G. Educational research. An introduction.<br />
Borg,<br />
Ed.) New York: Davl'd McKay Company, Inc., 63.<br />
(2nd<br />
J. L. & Keintz, B. L. Computational handbook of statistics.<br />
Bruning,<br />
Scott, Foresman and Co., 1968.<br />
Dallas-<br />
D. T. & Stanley, J. C. Experimental and quasi-experimental<br />
Campbell,<br />
for research. Chlcago: Rand McNal--, 1966,<br />
designs<br />
F., Zawada, A., Thompson, B., & Markowitz, J. Guidelines and<br />
Collignon,<br />
for evaluating vocational rehabilitation programs: A dis-<br />
criteria<br />
paper for the prime study group on program evaluatlon, Tenth<br />
cussion<br />
on Rehabilitation Services, Berkeley: Institute of Urban<br />
Institute<br />
& Regional Development, University of Californla, 1972.<br />
O. L., Rosenstock, I. M., & Gettlng, V. A. Evaluation of pro-<br />
Deniston,<br />
effectiveness. Public Health Reports, 1968, 83(4), 323-335.<br />
gram<br />
G. A. Statistical analysls n psychology and education. New<br />
Ferguson,<br />
McGraw-Hi I l, 1966.<br />
York:<br />
J. P. Fundamental statstlcs n psychology and educatlon. New<br />
Guilford,<br />
McGraw-Hill, 19'0.<br />
York"<br />
L. A. A bass for scaling qualitative data. American Sociological<br />
Guttman,<br />
1944, 9, 139-150.<br />
Review,<br />
Hatry, H. P., Wnnie, R. E., & Fisk, D. M. Practical program evaluation<br />
state and local overnment officials'. Washington, D.C.: Urban<br />
for<br />
n s----ti tute ,73.<br />
I<br />
W. G. & V1alle, H. V. Consumer's measurement of vocational reha-<br />
H111s,<br />
Norman, OK. University of Oklahoma Pr--ess, 1973.<br />
bilitation.<br />
-221-
T. J. Goal attalnment scaling at a county mental health servlce.<br />
Kiresuk,<br />
1973, 1, 12-18.<br />
Evaluation,<br />
T. J. & Sherman, R. E. Goal attainment scallng. A general method<br />
Kresuk,<br />
evaluating comprehensive community mental health programs. Com-<br />
for<br />
munity Mental Health Journal, 1968, 4_(6), 443-453.<br />
F. N. Foundations of behavioral research. New York- Holt,<br />
Kerlinger,<br />
& Winston, 1964--<br />
Rnehart,<br />
E. F. Design and anal)'sis of experiments In psychology and<br />
Lindquist,<br />
Boston Houghton-Mifflin, 1953<br />
education.<br />
R. F. Preparing instructlonal objectives. Palo Alto, CA- Feardon<br />
Mager,<br />
1962.<br />
Publishers,<br />
A. H. Motivation and personality. (2nd Ed.) Evanston, IL- Harper<br />
Maslow,<br />
Row, 1970.<br />
and<br />
R. B. Fundamental statlstics for psychology. New York. Harcourt,<br />
McCall,<br />
and World, Inc., 1970.<br />
Brace<br />
K. W Kravetz, S., Wright, G. N., & Butler, A. J. A better<br />
Reagles,<br />
The Human Service Scale. Madison. The University of<br />
yardstick"<br />
Wisconsin Regional Rehabilitation Research Institute, 1974.<br />
K. W., Wright, G. N.,& Butler, A.J. Human Serwce Scale. Madison,<br />
Reagles,<br />
Human Service Systems, Inc., 1973.<br />
WI:<br />
I. M. & Guttman, L. Smallest Space Analysis of Intellgence<br />
Schlesinger,<br />
achievement test. Psychological Bulletin, 1969, 71, 95-100.<br />
and<br />
D. T. The difficult), ndex an expanded measure of counselor<br />
Sermon,<br />
St Paul. Minnesota Divslon of Vocatiol' Rehabi'i'"ta-<br />
performance.<br />
tion, Research Monograph 1, 1972.<br />
S. N..onparametric statstlcs for the behavioral sclences. New<br />
Segel,<br />
McGraw-Hll, 1956.<br />
York-<br />
E. A. Action for what A crtlque of evaluation research. In<br />
Suchman,<br />
R. (Ed.). The organlzaton, management, and tactics of soclal<br />
O'Toole,<br />
research. Cambridgeassachusetts Schenkman Publishing Company, 1971.<br />
Institute on Rehab11tation Services. Program evaluation A begin-<br />
Tenth<br />
statement. Washington, D.C. Rehabilitation Servlces A-minlstra-<br />
ning<br />
U. S. Department of Health, Education, and Welfare, 1972.<br />
tion,<br />
D. J. An introduction to evaluatlon Program effectiveness and<br />
Trantow,<br />
need. Rehabilitation Literature, 1970, 31(1), 2-9, 12.<br />
connunity<br />
-222-
df<br />
SP<br />
SS<br />
SW2<br />
T<br />
0<br />
P<br />
r<br />
D<br />
F<br />
GLOSSARY OF SYMBOLS<br />
in a regression equation Indlcatlng the orlgln or inter-<br />
constant<br />
of a regression 11he with the Y axls<br />
cept<br />
regression welght reflecting the slope of a regression line.<br />
the number of columns<br />
a statistic whose value Is computed fr observed<br />
cht-square,<br />
e.g. from frequency data.<br />
data,<br />
of freedom.<br />
degrees<br />
between paired measurements.<br />
difference<br />
frequency in the calculation of X 2<br />
expected<br />
frequency<br />
the F test- analysis of variance.<br />
a subscript denoting rows.<br />
a subscript denoting columns<br />
number of groups or subclasses<br />
square (in analysis of variance)<br />
mean<br />
of observations in a subclass<br />
number<br />
number of observations n a sample,<br />
total<br />
frequency in the calculation of X 2.<br />
observed<br />
11ty estimate.<br />
probabl<br />
I n=o, o, row,, o,<br />
l,<br />
the Pearson product-moment correlation coefflcleet.<br />
of products (in analyss of covarlance)<br />
sum<br />
of squares (in analysis of variance).<br />
sum<br />
mean square wlthln (In a $cheffd test)<br />
t-test.<br />
the<br />
standardt°tal (inscore.analysis of variance), or<br />
})<br />
indicating a level of mportance<br />
weight,<br />
expressed as scores<br />
varlables<br />
the mean (or average) of a sample of observations<br />
pred cted score.<br />
standard error or deviation of a population or sample<br />
sum of, symbols above and below deflne lmits of the sum-<br />
the<br />
on. mati<br />
a superscript denoting Instructlons to square a value, e.g. 4 2 16<br />
root' sign.<br />
'square<br />
s gn.<br />
equal<br />
than' s gn.<br />
greater<br />
than' sign.<br />
'less