16.01.2015 Views

Article 4: Mathematical Word Problem Solving in Third-Grade ...

Article 4: Mathematical Word Problem Solving in Third-Grade ...

Article 4: Mathematical Word Problem Solving in Third-Grade ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 283<br />

<strong>Mathematical</strong> <strong>Word</strong> <strong>Problem</strong> <strong>Solv<strong>in</strong>g</strong> <strong>in</strong><br />

<strong>Third</strong>-<strong>Grade</strong> Classrooms<br />

ASHA K. JITENDRA CYNTHIA C. GRIFFIN ANDRIA DEATLINE-BUCHMAN<br />

Lehigh University University of Florida Easton School District<br />

EDWARD SCZESNIAK<br />

Easton School District<br />

ABSTRACT The authors conducted design or classroom<br />

experiments (R. Gersten, S. Baker, & J. W. Lloyd, 2000) at<br />

2 sites (Pennsylvania and Florida) to test the effectiveness of<br />

schema-based <strong>in</strong>struction (SBI) prior to conduct<strong>in</strong>g formal<br />

experimental studies. Results of Study 1 conducted <strong>in</strong> 2 3rdgrade,<br />

low-ability classrooms and 1 special education classroom<br />

<strong>in</strong>dicated mean score improvements from pretest to<br />

posttest on word problem solv<strong>in</strong>g and computation fluency<br />

measures. In addition, student perceptions of SBI accord<strong>in</strong>g to<br />

a strategy satisfaction questionnaire revealed SBI as effective<br />

<strong>in</strong> help<strong>in</strong>g solve word problems. Results of Study 2, which<br />

<strong>in</strong>cluded a heterogeneous (high-, average-, and low-achiev<strong>in</strong>g)<br />

sample of 3rd graders, also revealed student improvement on<br />

the word problem solv<strong>in</strong>g and computation fluency measures.<br />

However, the outcomes were not as positive <strong>in</strong> Study 2 as <strong>in</strong><br />

Study 1. Lessons learned from the 2 studies are discussed<br />

with regard to teach<strong>in</strong>g and learn<strong>in</strong>g mathematical word problem<br />

solv<strong>in</strong>g for different groups of students.<br />

Keywords: elementary grade students, mathematics <strong>in</strong>struction,<br />

schema-based <strong>in</strong>struction, word problem solv<strong>in</strong>g<br />

Current education reforms support challeng<strong>in</strong>g<br />

learn<strong>in</strong>g standards and school accountability. In<br />

mathematics education, the emphasis is on the<br />

development of conceptual understand<strong>in</strong>g and reason<strong>in</strong>g<br />

over memorization and rote learn<strong>in</strong>g (Goldsmith & Mark,<br />

1999; Hiebert et al., 1996; National Research Council,<br />

2001). <strong>Mathematical</strong> problem solv<strong>in</strong>g is a central theme <strong>in</strong><br />

the Pr<strong>in</strong>ciples and Standards for School Mathematics (National<br />

Council of Teachers of Mathematics [NCTM], 2000).<br />

Although the current emphasis is on solv<strong>in</strong>g complex<br />

authentic problems situated <strong>in</strong> everyday contexts, story<br />

problems that range from simple to complex represent “the<br />

most common form of problem-solv<strong>in</strong>g” assignment <strong>in</strong><br />

school mathematics curricula (Jonassen, 2003, p. 267).<br />

Story problems pose difficulties for many elementary<br />

students because of the complexity of the solution process<br />

(Jonassen, 2003; Lucangeli, Tressoldi, & Cendron, 1998;<br />

Schurter, 2002). Mathematics tasks that <strong>in</strong>volve storycontext<br />

problems are much more challeng<strong>in</strong>g than are nocontext<br />

problems (Cumm<strong>in</strong>s, K<strong>in</strong>tsch, Reusser, &<br />

283<br />

Weimer, 1988; Mayer, Lewis, & Hegarty, 1992; Nathan,<br />

Long, & Alibali, 2002). Although children may know<br />

procedures for solv<strong>in</strong>g no-context problems, solv<strong>in</strong>g story<br />

problems requires them to <strong>in</strong>tegrate several cognitive<br />

processes that are difficult for children with an <strong>in</strong>sufficient<br />

knowledge base or limited work<strong>in</strong>g memory capacity.<br />

When solv<strong>in</strong>g story problems, for example, children<br />

need to (a) understand the language and factual <strong>in</strong>formation<br />

<strong>in</strong> the problem, (b) translate the problem with relevant<br />

<strong>in</strong>formation to create an adequate mental representation,<br />

(c) devise and monitor a solution plan, and (d)<br />

execute adequate procedural calculations (Desoete, Roeyers,<br />

& De Clercq, 2003; Mayer, 1999). In short, solv<strong>in</strong>g<br />

word problems relates closely to comprehension of the<br />

relations and goals with<strong>in</strong> the problem (e.g., Briars &<br />

Lark<strong>in</strong>, 1984; Cumm<strong>in</strong>s et al.; De Corte, Verschaffel, &<br />

De W<strong>in</strong>, 1985; K<strong>in</strong>tsch & Greeno, 1985; Riley, Greeno,<br />

& Heller, 1983).<br />

Story problems are critical for help<strong>in</strong>g children connect<br />

different mean<strong>in</strong>gs, <strong>in</strong>terpretations, and relationships to<br />

mathematics operations (Van de Walle, 2004). How then<br />

do educators enhance students’ mathematical word problem-solv<strong>in</strong>g<br />

skill Traditional textbook problem-solv<strong>in</strong>g<br />

<strong>in</strong>struction has not effectively improved the learn<strong>in</strong>g of students<br />

at risk for mathematics difficulties. Many mathematics<br />

textbooks are organized so that the same procedure<br />

(e.g., subtraction) is used to solve all problems on a page.<br />

As a result, students do not have the opportunity to discrim<strong>in</strong>ate<br />

among problems that require different solution<br />

strategies. Furthermore, traditional <strong>in</strong>struction teaches students<br />

to use keywords (e.g., <strong>in</strong> all suggests addition, left suggests<br />

subtraction, share suggests division; Lester, Garofalo,<br />

& Kroll, 1989, p. 84) that are limit<strong>in</strong>g. Accord<strong>in</strong>g to Van<br />

de Walle, “Key words are mislead<strong>in</strong>g,” “Many problems do<br />

not have key words,” and “Key words send a terribly wrong<br />

message about do<strong>in</strong>g math” (p. 152). Thus, this approach<br />

Address correspondence to Asha K. Jitendra, Lehigh University,<br />

College of Education, 315 A Iacocca Hall, 111 Research Drive, Bethlehem<br />

PA 18015. (E-mail: akj2@lehigh.edu)<br />

Copyright © 2007 Heldref Publications


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 284<br />

284 The Journal of Educational Research<br />

ignores the mean<strong>in</strong>g and structure of the problem and fails<br />

to develop reason<strong>in</strong>g and mak<strong>in</strong>g sense of problem situations<br />

(Van de Walle).<br />

In contrast, current mathematics reform efforts advocate<br />

that teachers act as facilitators to help students construct<br />

their own understand<strong>in</strong>gs of mathematical concepts and<br />

relationships (e.g., develop<strong>in</strong>g the abilities of <strong>in</strong>quiry, problem<br />

solv<strong>in</strong>g, and mathematics connections). However, the<br />

literature is <strong>in</strong>conclusive about the benefits of that approach<br />

for all learners (Baxter, Woodward, Voorhies, & Wong,<br />

2002). As the range of skills that students br<strong>in</strong>g to the classroom<br />

<strong>in</strong>creases, the challenge for teachers is synthesiz<strong>in</strong>g<br />

knowledge regard<strong>in</strong>g mathematics content and processes,<br />

student learn<strong>in</strong>g, effective <strong>in</strong>struction models, and appropriate<br />

classroom opportunities and experiences (Ma, 1999).<br />

Learners who lack sufficient prior knowledge may need more<br />

supportive <strong>in</strong>structional strategies <strong>in</strong> which the teacher scaffolds<br />

<strong>in</strong>formation process<strong>in</strong>g by supply<strong>in</strong>g a greater degree of<br />

<strong>in</strong>structional facilitation dur<strong>in</strong>g the learn<strong>in</strong>g process.<br />

To promote elementary students’ mathematical problemsolv<strong>in</strong>g<br />

skills, we relied on models for understand<strong>in</strong>g and<br />

assess<strong>in</strong>g children’s solutions for addition and subtraction<br />

word problems derived from schema theories of cognitive<br />

psychology (Briars & Lark<strong>in</strong>, 1984; Carpenter & Moser,<br />

1984; K<strong>in</strong>tsch & Greeno, 1985; Riley, Greeno, & Heller,<br />

1983). On the basis of those models, we designed an <strong>in</strong>tervention<br />

(i.e., schema-based <strong>in</strong>struction [SBI]) for students<br />

with high-<strong>in</strong>cidence disabilities and for those at risk for<br />

failure <strong>in</strong> mathematics (e.g., Jitendra et al., 1998; Jitendra,<br />

Hoff, & Beck, 1999). Contrary to earlier <strong>in</strong>vestigations of<br />

SBI that provided <strong>in</strong>dividual or small-group <strong>in</strong>struction by<br />

researchers <strong>in</strong> rooms adjacent to the students’ classrooms,<br />

we deemed it critical to broaden the learn<strong>in</strong>g environment<br />

to a typical classroom context where<strong>in</strong> the classroom<br />

teacher provided all <strong>in</strong>struction. In addition, we added selfmonitor<strong>in</strong>g<br />

to strategy use because teach<strong>in</strong>g students to<br />

self-regulate their learn<strong>in</strong>g has an added positive effect on<br />

their mathematics problem-solv<strong>in</strong>g performance (L. S.<br />

Fuchs et al., 2003; Schunk, 1998; Verschaffel et al., 1999).<br />

Theoretical Framework<br />

Results of a number of research studies on solv<strong>in</strong>g addition<br />

and subtraction word problems have shown that<br />

semantic or mathematical structure of problems (i.e., specific<br />

characteristics of the problem and the semantic relationships<br />

among the various problem features) is much more<br />

relevant than is syntax (i.e., how a problem is worded; Carpenter,<br />

Hiebert & Moser, 1983; Carpenter & Moser, 1984).<br />

Notwithstand<strong>in</strong>g the ease or difficulty of the syntactic<br />

structure of problems, students who lack a well-developed<br />

semantic structure use a “bottom-up or text-driven<br />

approach to comprehend a problem statement” (Kameenui<br />

& Griff<strong>in</strong>, 1989, p. 581). Given the critical role of semantic<br />

structure <strong>in</strong> problem solution, Carpenter and Moser postulated<br />

a classification of addition and subtraction wordproblem<br />

types that <strong>in</strong>clude change, comb<strong>in</strong>e, compare, and<br />

equalize. Of those problem types, change, comb<strong>in</strong>e, and compare<br />

problems are characteristic of most addition and subtraction<br />

word problems presented <strong>in</strong> elementary mathematics<br />

textbooks, <strong>in</strong>dicat<strong>in</strong>g the need for research focus<strong>in</strong>g<br />

on these problem types. Table 1 shows the three different<br />

problem types and their characteristics. Change problems<br />

usually beg<strong>in</strong> with an <strong>in</strong>itial quantity and a direct or<br />

implied action that causes either an <strong>in</strong>crease or decrease <strong>in</strong><br />

that quantity. The three sets of <strong>in</strong>formation <strong>in</strong> a change<br />

problem are the beg<strong>in</strong>n<strong>in</strong>g, change, and end<strong>in</strong>g. In the<br />

change situation, the object identities (e.g., video games)<br />

for beg<strong>in</strong>n<strong>in</strong>g, change, and end<strong>in</strong>g are the same (see Figure<br />

1). In contrast, comb<strong>in</strong>e or group problems <strong>in</strong>volve two dist<strong>in</strong>ct<br />

groups or subsets that comb<strong>in</strong>e to form a new group or<br />

set. Group problems require an understand<strong>in</strong>g of part-partwhole<br />

relations. The relation between a particular set and<br />

its two dist<strong>in</strong>ct subsets is static (i.e., no action is implied).<br />

Compare problems <strong>in</strong>volve the comparison of two disjo<strong>in</strong>t<br />

sets (compared and referent); the emphasis is on the static<br />

relation between the two sets (see Table 1). The three sets<br />

of <strong>in</strong>formation <strong>in</strong> a compare problem are the compared, referent,<br />

and difference. For each problem type, there are<br />

three items of <strong>in</strong>formation; the position of the unknown <strong>in</strong><br />

these problems may be any one of the three items, which<br />

can be found if the other two items are given.<br />

Also critical to successful mathematics problem solv<strong>in</strong>g<br />

is doma<strong>in</strong>-specific knowledge (conceptual and procedural;<br />

e.g., Hegarty, Mayer, & Monk, 1995). An important aspect<br />

of doma<strong>in</strong>-specific concept knowledge is problem comprehension/representation,<br />

which <strong>in</strong>volves translat<strong>in</strong>g the<br />

text of the problem <strong>in</strong>to a semantic representation on the<br />

TABLE 1. Semantic Structure of Addition and Subtraction <strong>Problem</strong>s<br />

<strong>Problem</strong> characteristics<br />

Relation<br />

<strong>Problem</strong> type Component Object identity between sets<br />

Change Beg<strong>in</strong>n<strong>in</strong>g, change, end<strong>in</strong>g Same Dynamic<br />

Comb<strong>in</strong>e or group Part, part, whole Different Static<br />

Compare Compared, referent, difference Different Static


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 285<br />

May/June 2007 [Vol. 100 (No. 5)] 285<br />

Change: Jane had 4 video games. Then her mother gave her 3 more video games for her birthday. Jane now has 7 video games.<br />

+ 3<br />

video games<br />

Change<br />

4<br />

video games<br />

7<br />

video games<br />

Beg<strong>in</strong>n<strong>in</strong>g<br />

End<strong>in</strong>g<br />

Group: Sixty-eight students at Hillcrest Elementary took part <strong>in</strong> the school play. There were 22 third graders, 19 fourth graders, and<br />

27 fifth graders <strong>in</strong> the school play.<br />

3rd<br />

graders<br />

4th<br />

graders<br />

5th<br />

graders<br />

All students (3rd,<br />

4th & 5th graders)<br />

22<br />

19<br />

27<br />

68<br />

Small Group<br />

or Parts<br />

Large Group<br />

or Whole (Sum)<br />

Compare: Joe is 8 years older than Jill. Jill is age 7 years and Joe is age 15 years.<br />

Joe<br />

15<br />

years<br />

Jill<br />

7<br />

years<br />

8<br />

years<br />

Bigger Smaller Difference<br />

FIGURE 1. Schematic diagrams for change, group, and compare problem situations. Change diagram adapted from “Schemas <strong>in</strong><br />

problem solv<strong>in</strong>g,” by S. P. Marshall, 1995, p. 133. Copyright 1995 by Cambridge University Press. Adapted with permission of<br />

the author.<br />

basis of an understand<strong>in</strong>g of the problem structure.<br />

Although procedural knowledge (e.g., know<strong>in</strong>g the series of<br />

sequential steps used to solve rout<strong>in</strong>e mathematical tasks,<br />

such as add<strong>in</strong>g 27 + 19 as well as know<strong>in</strong>g mathematical<br />

symbolism, such as +, =, and >) is also important, it “is<br />

extremely limited unless it is connected to a conceptual<br />

knowledge base” (Prawat, 1989, p. 10). In fact, a lack of<br />

understand<strong>in</strong>g of the problem situation may result <strong>in</strong> a solu-


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 286<br />

286 The Journal of Educational Research<br />

tion plan that is immaturely developed. Successful problem<br />

solvers can translate and <strong>in</strong>tegrate <strong>in</strong>formation <strong>in</strong> the problem<br />

<strong>in</strong>to a coherent mental representation that mediates<br />

problem solution (Mayer, 1999; Mayer & Hegarty, 1996).<br />

However, many students have difficulty with problem comprehension<br />

and solution and would benefit from <strong>in</strong>struction<br />

<strong>in</strong> construct<strong>in</strong>g a model to represent the situation <strong>in</strong><br />

the text, followed by solution plann<strong>in</strong>g based on the model<br />

(Hegarty et al., 1995). Evidently, schema-based <strong>in</strong>struction<br />

(SBI), with its emphasis on semantic structure and problem<br />

representation, is one solution to advanc<strong>in</strong>g students’<br />

mathematical problem-solv<strong>in</strong>g skills.<br />

The goal of SBI is to help students establish and expand<br />

on the doma<strong>in</strong> knowledge <strong>in</strong> which schemata are the central<br />

focus. A schema is a general description of a group of<br />

problems that share a common underly<strong>in</strong>g structure requir<strong>in</strong>g<br />

similar solutions (Chen, 1999; Gick & Holyoak, 1983).<br />

Accord<strong>in</strong>g to Marshall (1995), schemata “capture both the<br />

patterns of relationships as well as their l<strong>in</strong>kages to operations”<br />

(Marshall, 1995, p. 67). SBI analyzes explicitly the<br />

problem schema (e.g., the part-part whole) and the l<strong>in</strong>ks<br />

perta<strong>in</strong><strong>in</strong>g to how different elements of the schema are<br />

related (e.g., parts make up the whole). Understand<strong>in</strong>g<br />

these l<strong>in</strong>ks is crucial <strong>in</strong> select<strong>in</strong>g appropriate operations<br />

needed for problem solution. For example, if the whole <strong>in</strong><br />

the problem is unknown, add<strong>in</strong>g the parts is necessary to<br />

solve for the whole; if one of the parts is unknown, subtract<strong>in</strong>g<br />

the part(s) from the whole is needed to solve for<br />

the unknown part. An important difference between SBI<br />

and other <strong>in</strong>structional approaches is that only SBI emphasizes<br />

<strong>in</strong>tegrat<strong>in</strong>g the various pieces of factual <strong>in</strong>formation<br />

essential for problem solv<strong>in</strong>g. Although factual details are<br />

important, they should not be the central focus of <strong>in</strong>struction<br />

and learn<strong>in</strong>g. In short, SBI allows students to approach<br />

the problem by focus<strong>in</strong>g on the underly<strong>in</strong>g problem structure,<br />

thus facilitat<strong>in</strong>g conceptual understand<strong>in</strong>g and adequate<br />

word-problem-solv<strong>in</strong>g skills (Marshall).<br />

Purpose<br />

We conducted design or classroom experiments concurrently<br />

at two different geographical sites. Our research team<br />

collaborated with teachers to conduct a series of word problem<br />

solv<strong>in</strong>g teach<strong>in</strong>g sessions us<strong>in</strong>g SBI with a small sample<br />

of students to understand what works, prior to conduct<strong>in</strong>g a<br />

formal experiment (Cobb, Confrey, diSessa, Lehrer, &<br />

Schauble, 2003). Design experiments <strong>in</strong>volve <strong>in</strong>-depth<br />

study of an <strong>in</strong>novation <strong>in</strong> a small scale to understand not<br />

only how responsive an <strong>in</strong>tervention is to the dynamics of<br />

the classroom but also to understand “the nature of effective<br />

adaptations for students with disabilities” (Gersten et al.,<br />

2000, p. 29). Specifically, a design experiment is aimed at<br />

improv<strong>in</strong>g the connections between teach<strong>in</strong>g and learn<strong>in</strong>g<br />

through the process of cont<strong>in</strong>uous cycles of “design, enactment,<br />

analysis, and redesign” of lessons <strong>in</strong> authentic classrooms<br />

(Cobb et al., 2003; The Design-Based Research Collective,<br />

2003). Researchers use the processes of iteration<br />

and feedback loops to empirically ref<strong>in</strong>e the <strong>in</strong>novation<br />

(e.g., curriculum product, tasks, scaffolds) and design an<br />

effective and efficient <strong>in</strong>tervention for use <strong>in</strong> experimental<br />

or quasi-experimental studies (Gersten et al., 2000).<br />

The learn<strong>in</strong>g environment <strong>in</strong> our design experiments<br />

<strong>in</strong>cluded a typical classroom context rather than an exceptional<br />

<strong>in</strong>structional sett<strong>in</strong>g (i.e., special rooms outside the<br />

regular classroom), where<strong>in</strong> the teacher conducted all<br />

<strong>in</strong>struction. The purposes of the design experiments were<br />

twofold. First, we <strong>in</strong>vestigated the effects of SBI on the<br />

acquisition of skills for solv<strong>in</strong>g mathematical word problems<br />

by third graders and explored the differential effects of<br />

SBI for different groups of children (e.g., children with<br />

learn<strong>in</strong>g disabilities; low, average, and high achievers). Second,<br />

we exam<strong>in</strong>ed the <strong>in</strong>fluence of word problem-solv<strong>in</strong>g<br />

<strong>in</strong>struction on the acquisition of computational skills given<br />

the role that word problems play <strong>in</strong> the development of<br />

number operations (Van de Walle, 2004).<br />

The basic methods were similar across both studies. In<br />

Study 1, we addressed the issue of the efficacy of SBI for a<br />

homogeneous sample of students from low-ability mathematics<br />

classrooms and potential differences between students<br />

with learn<strong>in</strong>g disabilities and students without disabilities<br />

who were low achievers. Study 2 focused on the<br />

efficacy of SBI for a heterogeneous sample of third graders,<br />

<strong>in</strong>clud<strong>in</strong>g students differ<strong>in</strong>g <strong>in</strong> <strong>in</strong>itial mathematics gradelevel<br />

achievement.<br />

DESIGN STUDY 1<br />

Method<br />

Participants<br />

Participants <strong>in</strong>cluded 40 third-grade students who<br />

attended an elementary school <strong>in</strong> a suburban school district<br />

<strong>in</strong> Pennsylvania serv<strong>in</strong>g 472 students <strong>in</strong> <strong>Grade</strong>s K–5.<br />

Approximately 15% of the student population was African<br />

American, Hispanic, or Asian; 17% were economically disadvantaged;<br />

and 5% were English language learners (ELLs).<br />

<strong>Third</strong>-grade students were grouped accord<strong>in</strong>g to ability levels<br />

(two low- and two high-ability classrooms). Teachers <strong>in</strong><br />

the two low-ability third-grade classrooms (Classrooms 1<br />

and 2), as well as the special education teacher (Classroom<br />

3), participated to learn about <strong>in</strong>novative approaches to<br />

improv<strong>in</strong>g students’ mathematical problem-solv<strong>in</strong>g performance.<br />

The f<strong>in</strong>al sample, however, comprised 38 students<br />

(20 boys and 18 girls) because 2 students moved out of the<br />

school district before complet<strong>in</strong>g the study. The mean<br />

chronological age of students was 102.60 months (range =<br />

91 to 119 months; SD = 5.54). Twenty-eight students (74%)<br />

were Caucasian, 3 (8%) were African American, 6 (16%)<br />

were Hispanic, and 1 was Middle Eastern (3%).<br />

The sample <strong>in</strong>cluded 9 students (6 boys and 3 girls) with<br />

learn<strong>in</strong>g disabilities (LD). While 3 of the students with LD<br />

received <strong>in</strong>struction <strong>in</strong> the general education classrooms,


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 287<br />

May/June 2007 [Vol. 100 (No. 5)] 287<br />

the rema<strong>in</strong><strong>in</strong>g 6 students were <strong>in</strong>structed <strong>in</strong> the learn<strong>in</strong>g<br />

support classroom. In addition, we designated 9 of the 38<br />

students as low achievers (LA) on the basis of their score<br />

(below the 35th percentile) on the Computation and Concepts<br />

and Applications subtests of the TerraNova mathematics<br />

achievement test (CTB/McGraw-Hill, 2001).<br />

Tables 2 and 3 show a summary of participant demographic<br />

<strong>in</strong>formation by classroom and group status.<br />

Separate one-way analyses of variance (ANOVAs) <strong>in</strong>dicated<br />

no significant differences between classrooms on the<br />

mathematics subtests of the TerraNova: Concepts and<br />

Application, F(2, 35) = 0.38, ns, and Computation, F(2,<br />

35) = 0.80, ns; Total Read<strong>in</strong>g subtest, F(2, 35) = 1.03, ns.<br />

However, there was a significant difference between classrooms<br />

on chronological age, F(2, 35) = 15.99, p < .000.<br />

The mean age of special education students <strong>in</strong> Classroom 3<br />

was greater than that of students <strong>in</strong> Classrooms 1 and 2,<br />

which we expected given that the special education students<br />

were <strong>in</strong> the third and fourth grades. Chi-square<br />

analyses revealed no significant between-classroom differences<br />

on gender, χ 2 (2, N = 38) = 2.57, ns, and race, χ 2 (6,<br />

N = 38) = 8.68, ns.<br />

Similarly, separate one-way ANOVAs <strong>in</strong>dicated no significant<br />

between-group (LD and LA) differences on the Concepts<br />

and Application subtest of the TerraNova, F(1, 16) =<br />

2.84, ns, and on the Total Read<strong>in</strong>g subtest, F(1, 16) = 1.96,<br />

ns. However, there was a significant between-group difference<br />

on the Computation subtest, F(1, 16) = 5.53, p < .05.<br />

The mean score for students with LD was greater than that<br />

for LA students. In addition, there was a significant betweengroup<br />

difference on chronological age, F(1, 16) = 10.73, p <<br />

.01. The mean age of LD students was greater than was that<br />

of LA students. Chi-square analyses revealed no significant<br />

between-group differences on gender, χ 2 (1, N = 18) = .90,<br />

ns, and race, χ 2 (2, N = 18) = 1.62, ns.<br />

Three female teachers participated <strong>in</strong> Study 1, each<br />

teach<strong>in</strong>g <strong>in</strong> one classroom. The two general education<br />

teachers were certified <strong>in</strong> elementary education and each<br />

had more than 25 years of teach<strong>in</strong>g experience; the special<br />

education teacher was certified <strong>in</strong> special education and<br />

had 19 years of teach<strong>in</strong>g experience. Teachers attended 1<br />

hr of <strong>in</strong>service tra<strong>in</strong><strong>in</strong>g on ways to implement the <strong>in</strong>tervention;<br />

they received ongo<strong>in</strong>g support from two research<br />

assistants (doctoral students <strong>in</strong> special education) throughout<br />

the study.<br />

Procedures<br />

In all classrooms, teachers taught mathematics five times<br />

a week for 50 m<strong>in</strong> with the district-adopted basal text,<br />

Heath Mathematics Connections (Manfre, Moser, Lobato, &<br />

Morrow, 1994). Teachers taught the word problem-solv<strong>in</strong>g<br />

unit dur<strong>in</strong>g regularly scheduled mathematics <strong>in</strong>struction for<br />

15 weeks. They taught students how to solve one-step addition<br />

and subtraction problems <strong>in</strong>volv<strong>in</strong>g change, group,<br />

and compare problems for 30 m<strong>in</strong> daily, 3 days per week.<br />

Instruction was scripted to ensure consistency of <strong>in</strong>formation<br />

and <strong>in</strong>cluded an <strong>in</strong>structional paradigm of teacher<br />

model<strong>in</strong>g with th<strong>in</strong>k-alouds, followed by guided practice,<br />

paired partner work, <strong>in</strong>dependent practice, and homework.<br />

In addition, <strong>in</strong>struction dur<strong>in</strong>g guided practice emphasized<br />

frequent teacher–student exchanges to facilitate problem<br />

solv<strong>in</strong>g. <strong>Word</strong> problem-solv<strong>in</strong>g <strong>in</strong>struction <strong>in</strong>volved two<br />

phases: (a) problem schema and (b) problem solution. We<br />

developed a strategy checklist to help scaffold student<br />

learn<strong>in</strong>g. The checklist <strong>in</strong>cluded the follow<strong>in</strong>g six steps: (a)<br />

read and retell the problem to discover the problem type:<br />

(b) underl<strong>in</strong>e and map important <strong>in</strong>formation <strong>in</strong> the word<br />

problem onto the schematic diagram, (c) decide whether to<br />

add or subtract to solve the problem, (d) write the mathematics<br />

sentence and solve it, (e) write the complete answer,<br />

and (f) check the answer.<br />

Schema for <strong>Problem</strong>-<strong>Solv<strong>in</strong>g</strong> Instruction<br />

Dur<strong>in</strong>g the problem-solv<strong>in</strong>g phase, students received<br />

story situations that did not conta<strong>in</strong> any unknown <strong>in</strong>formation.<br />

Instruction focused on identify<strong>in</strong>g the problem<br />

schema for each of the three problem types (change, group,<br />

compare) and on represent<strong>in</strong>g the features of the story situation<br />

with schematic diagrams (see Figure 1). That is, students<br />

learned to <strong>in</strong>terpret and elaborate on the ma<strong>in</strong> features<br />

of the story situation and map the details of the story<br />

onto the schema diagram. Teachers <strong>in</strong>troduced each problem<br />

type successively and cumulatively reviewed them to<br />

help students discern the three problem types.<br />

Change story situation. Students learned to identify the<br />

problem type by us<strong>in</strong>g story situations such as, “Jane had<br />

four video games. Then her mother gave her three more<br />

video games for her birthday. Jane now has seven video<br />

games.” Us<strong>in</strong>g Step 1 of the strategy checklist, students<br />

identified the story situation as change because it <strong>in</strong>itially<br />

<strong>in</strong>volved four video games, then an action occurred that<br />

<strong>in</strong>creased this quantity by three, which resulted <strong>in</strong> a total<br />

of seven video games. Also, the object identity of the<br />

beg<strong>in</strong>n<strong>in</strong>g, change, and end<strong>in</strong>g is video games. As such,<br />

this story situation is considered a change problem<br />

schema. For Step 2, teachers prompted students to use the<br />

correspond<strong>in</strong>g diagram (see Figure 1) to organize or represent<br />

the <strong>in</strong>formation. That step <strong>in</strong>volved identify<strong>in</strong>g and<br />

writ<strong>in</strong>g the object identity or label (e.g., video games) for<br />

the three items of <strong>in</strong>formation (i.e., beg<strong>in</strong>n<strong>in</strong>g, change,<br />

and end<strong>in</strong>g) <strong>in</strong> the change diagram. Students then read<br />

the story to f<strong>in</strong>d the quantities associated with the beg<strong>in</strong>n<strong>in</strong>g,<br />

change, and end<strong>in</strong>g and wrote them <strong>in</strong> the diagram.<br />

Next, students summarized the <strong>in</strong>formation <strong>in</strong> the story<br />

with the completed diagram and checked the accuracy of<br />

the representation.<br />

Group story situation. Students identified the group story<br />

problem type by us<strong>in</strong>g story situations such as, “Sixty-eight<br />

students at Hillcrest Elementary took part <strong>in</strong> the school<br />

play. There were 22 third graders, 19 fourth graders, and 27


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 288<br />

288 The Journal of Educational Research<br />

TABLE 2. Student Demographic Information and Performance by Classroom: Site 1<br />

Classroom 1 Classroom 2 Classroom 3<br />

Variable n % M SD n % M SD n % M SD ES<br />

Age (<strong>in</strong> months) 99.69 3.39 102.44 3.74 110.83 5.40<br />

Gender<br />

Male 6 37.50 10 62.50 4 66.67<br />

Female 10 62.50 6 37.50 2 33.33<br />

Race<br />

EA 10 62.50 13 81.25 5 83.33<br />

AA 2 12.50 1 6.25 0 0.00<br />

Hispanic 4 25.00 2 12.50 0 0.00<br />

Other 0 0.00 0 0.00 1 16.67<br />

Spec. ed. student 1 2 6<br />

TerraNova<br />

C & A 567.62 21.23 575.79 29.31 569.50 20.86<br />

Computation 540.38 10.78 531.21 28.16 541.33 19.39<br />

Total read<strong>in</strong>g 612.15 15.64 607.79 27.55 594.17 36.96<br />

WPS-CRT (25/50)<br />

Pretest 21.38 7.62 16.69 7.17 16.92 5.18 C1 vs. C2: +0.63<br />

C1 vs. C3: +0.63<br />

C2 vs. C3: −0.03<br />

Posttest 29.03 7.02 25.47 7.66 24.83 5.03 C1 vs. C2: +0.49<br />

C1 vs. C3: +0.64<br />

C2 vs. C3: +0.09<br />

Growth 7.66 6.95 8.78 7.49 7.92 5.91 C1 vs. C2: −0.16<br />

C1 vs. C3: −0.04<br />

C2 vs. C3: +0.12<br />

WPS-F (16/32)<br />

Pretest 12.31 5.26 9.38 5.73 11.00 5.36 C1 vs. C2: +0.53<br />

C1 vs. C3: +0.25<br />

C2 vs. C3: −0.29<br />

Posttest 19.91 5.47 16.34 6.00 17.84 5.73 C1 vs. C2: +0.62<br />

C1 vs. C3: +0.37<br />

C2 vs. C3: −0.25


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 289<br />

May/June 2007 [Vol. 100 (No. 5)] 289<br />

Growth 7.59 4.92 6.97 4.73 7.13 5.07 C1 vs. C2: +0.13<br />

C1 vs. C3: +0.09<br />

C2 vs. C3: −0.03<br />

Computation a (25/43)<br />

Pretest 7.81 4.90 6.50 3.69 12.83 3.76 C1 vs. C2: +0.30<br />

C1 vs. C3: −1.08<br />

C2 vs. C3: −1.71<br />

Posttest 35.69 4.87 23.63 8.05 26.83 8.89 C1 vs. C2: +1.81<br />

C1 vs. C3: +1.45<br />

C2 vs. C3: −0.39<br />

Growth 27.88 6.89 17.13 7.14 14.00 6.88 C1 vs. C2: +1.49<br />

C1 vs. C3: +1.96<br />

C2 vs. C3: +0.43<br />

Strategy satisfaction<br />

questionnaire (5/5)<br />

Enjoyed 3.94 0.68 3.31 1.35 4.67 0.52 C1 vs. C2: +0.59<br />

C1 vs. C3: −1.13<br />

C2 vs. C3: −1.14<br />

Diagram 3.81 0.83 3.87 1.20 4.67 0.82 C1 vs. C2: −0.06<br />

C1 vs. C3: −1.04<br />

C2 vs. C3: −0.72<br />

Help solve 4.20 1.03 3.67 1.03 5.00 0.00 C1 vs. C2: +0.52<br />

C1 vs. C3: −0.90<br />

C2 vs. C3: −1.49<br />

Recommend 3.56 1.32 3.88 1.31 4.83 0.41 C1 vs. C2: −0.24<br />

C1 vs. C3: −1.09<br />

C2 vs. C3: −0.82<br />

Cont<strong>in</strong>ue 3.81 1.47 3.69 0.95 4.33 0.82 C1 vs. C2: +0.10<br />

C1 vs. C3: −0.39<br />

C2 vs. C3: −0.70<br />

Total 19.13 0.87 18.75 0.87 23.50 1.41 C1 vs. C2: +0.44<br />

C1 vs. C3: −4.24<br />

C2 vs. C3: −4.60<br />

Note. Parenthetical numbers separated by slash are total number of test items and total possible score for each pretest and posttest. EA = European American; AA = African American; Spec. ed.= special<br />

education; C & A = Concepts and Application mathematics subtest; WPS-CRT = word problem solv<strong>in</strong>g criterion test; WPS-F = word problem solv<strong>in</strong>g fluency measure; C = classroom; ES = effect size.<br />

a Digits correct.


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 290<br />

290 The Journal of Educational Research<br />

TABLE 3. Student Demographic Characteristics and Mathematics Performance by Group: Site 1<br />

Group<br />

LD (n = 9) LA (n = 9)<br />

Variable n % M SD n % M SD ES<br />

Age (<strong>in</strong> months) 108.44 6.50 100.22 3.80<br />

Gender<br />

Male 6 66.67 4 44.44<br />

Female 3 33.33 5 55.56<br />

Race<br />

EA 6 66.67 8 88.89<br />

AA 0 0.00 0 0.00<br />

Hispanic 2 22.22 1 11.11<br />

Other 1 11.11 0 0.00<br />

IQ<br />

Full scale 94.78 6.92<br />

Verbal 92.17 6.82<br />

Performance 95.00 7.75<br />

TerraNova<br />

C & A 567.00 18.46 552.29 14.81<br />

Comp. 537.50 18.24 511.71 24.18<br />

Total read<strong>in</strong>g 588.63 33.40 607.14 10.61<br />

WPS-CRT (25/50)<br />

Pretest 14.83 5.40 18.22 8.34 −0.48<br />

Posttest 22.61 5.61 25.83 7.64 −0.48<br />

Growth 3.57 4.20 4.83 7.43 0.03<br />

WPS-F (16/32)<br />

Pretest 14.29 5.71 13.94 5.19 0.06<br />

Posttest 17.50 4.58 18.78 6.91 −0.22<br />

Growth 8.89 2.40 9.72 4.01 −0.21<br />

Computation a (25/43)<br />

Pretest 10.44 4.67 4.00 1.73 +1.83<br />

Posttest 25.56 8.09 25.78 11.95 −0.02<br />

Growth 15.12 6.85 21.78 11.95 −0.56<br />

Strat. Quest. (5/5)<br />

Enjoyed 4.56 0.73 3.44 1.13 +1.18<br />

Diagram helps 4.44 1.13 4.00 0.87 +0.44<br />

Help solve 4.89 0.33 3.44 1.01 +1.93<br />

Recommend 4.56 1.01 3.89 1.05 +0.65<br />

Cont<strong>in</strong>ue 4.44 0.73 3.00 1.32 +1.35<br />

Total 22.89 3.26 17.78 2.73 +1.70<br />

Note. Parenthetical numbers separated by slashes are total number of test items and total possible score. LD = learn<strong>in</strong>g disability; LA = low achiev<strong>in</strong>g;<br />

EA = European American; AA = African American; C & A = Concepts and Application mathematics subtest; Comp. = Computation mathematics<br />

subtest; WPS-CRT = word problem solv<strong>in</strong>g criterion test; WPS-F = word problem solv<strong>in</strong>g fluency measure; ES = effect size.<br />

a<br />

Digits correct.<br />

fifth graders <strong>in</strong> the school play.” Us<strong>in</strong>g Step 1 of the strategy<br />

checklist, students learned that because the story<br />

described a situation <strong>in</strong> which three small groups (third-,<br />

fourth-, and fifth-grade students) comb<strong>in</strong>e to form a large<br />

group (all students <strong>in</strong> the play), it is a story situation of the<br />

group problem type. For Step 2, teachers prompted students<br />

to use the group diagram (see Figure 1) to represent the<br />

<strong>in</strong>formation. Instruction <strong>in</strong>volved identify<strong>in</strong>g the three<br />

small groups and the large group and writ<strong>in</strong>g the group<br />

names <strong>in</strong> the diagram. Students then read the story to f<strong>in</strong>d<br />

the quantities associated with each group and wrote them<br />

<strong>in</strong> the diagram. Next, students summarized the <strong>in</strong>formation<br />

<strong>in</strong> the story with the completed diagram and checked the<br />

accuracy of the representation.<br />

Compare story situation. Students identified the story<br />

comparison problem type by us<strong>in</strong>g story situations such as,<br />

“Joe is 15 years old. He is 8 years older than Jill. Jill is 7<br />

years old.” Us<strong>in</strong>g Step 1 of the strategy checklist, students<br />

identified the story situation as compare because it required<br />

a comparison of Joe’s age to Jill’s age. For Step 2, students<br />

used the correspond<strong>in</strong>g diagram (see Figure 1) to organize<br />

or represent the <strong>in</strong>formation. That diagram <strong>in</strong>volved read<strong>in</strong>g<br />

the comparison sentence (He is 8 years older than Jill)<br />

to identify the two sets that were compared <strong>in</strong> the story—


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 291<br />

May/June 2007 [Vol. 100 (No. 5)] 291<br />

determ<strong>in</strong><strong>in</strong>g the identity of the large (Joe’s age) and small<br />

(Jill’s age) sets—label<strong>in</strong>g them <strong>in</strong> the diagram, and writ<strong>in</strong>g<br />

the difference amount <strong>in</strong> the diagram. Students then read<br />

the story to f<strong>in</strong>d the quantities associated with the two sets<br />

and wrote them <strong>in</strong> the diagram. Next, students summarized<br />

the <strong>in</strong>formation <strong>in</strong> the story with the completed diagram<br />

and checked the accuracy of the representation.<br />

<strong>Problem</strong>-Solution Instruction<br />

Dur<strong>in</strong>g the problem solution phase that followed problem<br />

schema <strong>in</strong>struction, students solved problems with<br />

unknowns by us<strong>in</strong>g either addition or subtraction; SBI <strong>in</strong><br />

this phase <strong>in</strong>cluded six steps.<br />

Change problems. Teachers prompted students to identify<br />

and represent the problem (i.e., Steps 1 and 2) with the<br />

change schematic diagram similar to the problem schema<br />

<strong>in</strong>struction phase. The only difference was that students<br />

used a question mark to represent the unknown quantity <strong>in</strong><br />

the diagram. Step 3 <strong>in</strong>volved select<strong>in</strong>g the appropriate<br />

operation and transform<strong>in</strong>g the <strong>in</strong>formation <strong>in</strong> the diagram<br />

<strong>in</strong>to a number sentence. That is, students learned that they<br />

needed to add the parts if the whole was unknown and subtract<br />

for the part when the whole was known. Instruction<br />

emphasized that when the change action caused an<br />

<strong>in</strong>crease, the end<strong>in</strong>g quantity represented the whole; when<br />

the change action <strong>in</strong>volved a decrease, the beg<strong>in</strong>n<strong>in</strong>g<br />

quantity was the whole. For Step 4, students had to write<br />

the mathematics sentence and solve for the unknown us<strong>in</strong>g<br />

the operation identified <strong>in</strong> the previous step. Step 5<br />

prompted students to write a complete answer. F<strong>in</strong>ally, for<br />

Step 6, students had to check the reasonableness of their<br />

answer and ensure the accuracy of the representation and<br />

computation.<br />

Group problems. Students learned to identify and represent<br />

the problem (i.e., Steps 1 and 2) with the group<br />

schematic diagram as <strong>in</strong> the problem schema <strong>in</strong>struction<br />

phase. For Step 3, when select<strong>in</strong>g the operation to solve for<br />

the unknown quantity <strong>in</strong> group problems, students learned<br />

that the large group represents the whole and the small<br />

groups are the parts that make up the whole. Steps 4<br />

through 6 were identical to those described for the change<br />

problem.<br />

Compare problems. Students identified and represented<br />

the problem (i.e., Steps 1 and 2) us<strong>in</strong>g the compare<br />

schematic diagram as <strong>in</strong> the problem schema <strong>in</strong>struction<br />

phase. For Step 3, when select<strong>in</strong>g the operation to solve for<br />

the unknown quantity <strong>in</strong> compare problems, students<br />

learned that the larger set is the big number or whole,<br />

whereas the smaller set and difference are the parts that<br />

make up the larger set. Steps 4 through 6 were identical to<br />

those described for the change and group problems.<br />

When <strong>in</strong>structors taught students to solve word problems<br />

us<strong>in</strong>g SBI, only one type of story situation or word<br />

problem with the correspond<strong>in</strong>g schema diagram <strong>in</strong>itially<br />

appeared on student worksheets follow<strong>in</strong>g the <strong>in</strong>struction<br />

of that problem type (e.g., change problem). After students<br />

learned how to map the features onto schematic diagrams<br />

or solve change and group problems, teachers presented<br />

story situations or word problems with both types, along<br />

with a discussion of the samenesses and differences between<br />

the change and group problems. Later, when students completed<br />

<strong>in</strong>struction for change, group, and compare problem<br />

types, teachers distributed worksheets with word problems<br />

that <strong>in</strong>cluded all problem types.<br />

We measured fidelity of treatment with a checklist of<br />

critical <strong>in</strong>structional steps. The checklist <strong>in</strong>cluded salient<br />

<strong>in</strong>structional features (e.g., provid<strong>in</strong>g clear <strong>in</strong>structions,<br />

read<strong>in</strong>g word problems aloud, model<strong>in</strong>g the strategy application,<br />

provid<strong>in</strong>g guided practice). Two graduate students<br />

<strong>in</strong> special education collected fidelity data as they observed<br />

one of the <strong>in</strong>structors for approximately 30% of the teach<strong>in</strong>g<br />

sessions. Treatment fidelity, estimated as the percentage<br />

of steps completed correctly by the <strong>in</strong>structor, was 93%<br />

(range = 85–100%) across the 3 teachers.<br />

Measures and Data Collection<br />

Two research assistants adm<strong>in</strong>istered and scored the<br />

mathematical word problem-solv<strong>in</strong>g tests and computation<br />

tests with scripted directions and answer keys. All data<br />

were collected <strong>in</strong> a whole-class arrangement.<br />

<strong>Word</strong> problem-solv<strong>in</strong>g criterion referenced test (WPS-CRT).<br />

To assess student growth on third-grade addition and subtraction<br />

word problems, students completed the WPS-CRT<br />

prior to and at the end of the <strong>in</strong>tervention. The test consisted<br />

of 25 one-step and two-step addition and subtraction<br />

word problems. Part I of the CRT comprised 9 one-step<br />

word problems derived from the Test of <strong>Mathematical</strong><br />

Achievement (Brown, Cron<strong>in</strong>, & McEntire, 1994). Of the<br />

9 problems, 5 items <strong>in</strong>cluded distracters. Part II of the WPS-<br />

CRT <strong>in</strong>cluded 16 addition and subtraction word problems<br />

selected from five commonly used third-grade mathematics<br />

textbooks. The items consisted of 12 one-step and 4 twostep<br />

problems that met the semantic criteria for change,<br />

group, and compare problem types. In addition, 2 of the<br />

problems <strong>in</strong>cluded distracters. The design of the measure<br />

<strong>in</strong>cluded the three problem types and both operations.<br />

<strong>Word</strong> problems on Parts I and II of the WPS-CRT<br />

required apply<strong>in</strong>g simple (e.g., s<strong>in</strong>gle-digit numbers) to<br />

complex computation skills (e.g., three- and four-digit<br />

numbers; regroup<strong>in</strong>g). Students had 50 m<strong>in</strong> to complete<br />

the test. Directions for adm<strong>in</strong>ister<strong>in</strong>g the word problemsolv<strong>in</strong>g<br />

test required students to show their complete work<br />

and to write the answer and label. Scor<strong>in</strong>g <strong>in</strong>volved assign<strong>in</strong>g<br />

one po<strong>in</strong>t for the correct number model and one po<strong>in</strong>t<br />

for correct answer and label for a possible total score of 2<br />

po<strong>in</strong>ts for each item. Cronbach’s alpha for each of the<br />

pretest and posttest measures was 0.84. Interscorer agreement<br />

assessed by two research assistants <strong>in</strong>dependently<br />

scor<strong>in</strong>g 30% of the protocols was .92 at pretreatment and<br />

.97 at postreatment.


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 292<br />

292 The Journal of Educational Research<br />

<strong>Word</strong> problem-solv<strong>in</strong>g fluency (WPS-F) measure. We<br />

developed six forms to represent items <strong>in</strong> Part II of the<br />

WPS-CRT but modified them to <strong>in</strong>clude fewer problems,<br />

less advanced computation (one- and two-digit numbers<br />

only), and no problems with distracters to address the<br />

timed nature of the task. The problems for each WPS-fluency<br />

probe differed with respect to numbers, context, and<br />

position of problems, which were random. Given that each<br />

probe <strong>in</strong>cluded only half the number of problems on the<br />

WPS-CRT, the probes were not identical and differed with<br />

respect to the unknown quantity to be solved (e.g., result,<br />

beg<strong>in</strong>n<strong>in</strong>g, compared). We covered all possible comb<strong>in</strong>ations<br />

of problem types and operations <strong>in</strong> two probes rather<br />

than one probe. To control for the difficulty of the probes,<br />

we designed odd- and even-numbered probes to be parallel<br />

forms. Teachers adm<strong>in</strong>istered the probes every 3 weeks to<br />

monitor students’ progress <strong>in</strong> solv<strong>in</strong>g word problems. Students<br />

had 10 m<strong>in</strong> to complete eight problems. For the purpose<br />

of this study, we used the comb<strong>in</strong>ed score of the first<br />

two probes and the comb<strong>in</strong>ed score of the last two probes<br />

<strong>in</strong> the data analysis. Cronbach’s alphas for the two aggregated<br />

probes were 0.83 and 0.80, respectively. Interscorer<br />

agreement was .93 at pretreatment and .98 at postreatment.<br />

Basic mathematics computation fluency measure (Fuchs,<br />

Hamlett, & Fuchs, 1998). We monitored student progress<br />

toward proficiency on third-grade mathematics computation<br />

curriculum prior to and at the completion of the <strong>in</strong>tervention<br />

with basic mathematics computation probes. Students<br />

had 3 m<strong>in</strong> to complete 25 problems, with a maximum<br />

score of 43 correct digits. We scored the performance on the<br />

computation probes as the total number of correct digits,<br />

which provided credit for correct segments of responses.<br />

That assessment system is known to have adequate reliability<br />

and validity (see Fuchs, Fuchs, Hamlett, & All<strong>in</strong>der,<br />

1989). Interscorer agreement was .99 at pretreatment and<br />

1.00 at postreatment.<br />

Strategy satisfaction questionnaire. We developed and<br />

adm<strong>in</strong>istered a strategy-satisfaction questionnaire follow<strong>in</strong>g<br />

the <strong>in</strong>tervention to provide <strong>in</strong>formation about student perceptions<br />

regard<strong>in</strong>g the problem-solv<strong>in</strong>g strategy <strong>in</strong>tervention<br />

(i.e., SBI). The questionnaire <strong>in</strong>cluded five items that<br />

required students to rate whether they (a) enjoyed the<br />

strategy, (b) found the diagrams helpful <strong>in</strong> understand<strong>in</strong>g<br />

and solv<strong>in</strong>g problems, (c) improved their problem-solv<strong>in</strong>g<br />

skills, (d) would recommend us<strong>in</strong>g the strategy with other<br />

students, and (e) would cont<strong>in</strong>ue to use it to solve word<br />

problems <strong>in</strong> the classroom. Rat<strong>in</strong>gs for the Likert-type<br />

items on the questionnaire ranged from a high score of 5<br />

(strongly agree) to a low score of 1 (strongly disagree). Cronbach’s<br />

alpha for the questionnaire was 70.<br />

Data Analysis<br />

The unit of analysis was each student’s <strong>in</strong>dividual score<br />

(N = 16 for the general education classrooms; n = 6 for the<br />

special education classroom) rather than the classroom<br />

because of sample limitations. On the WPS-CRT, WPS-F,<br />

and computation pretest measures, we conducted separate<br />

one-way, between-subjects (classroom or group) analysis of<br />

variance (ANOVA) to exam<strong>in</strong>e <strong>in</strong>itial classroom or group<br />

(LD and LA) comparability. On the WPS-CRT, WPS-F,<br />

and computation pretest and posttest scores, we conducted<br />

a one between-subjects (classroom or group), one with<strong>in</strong>subjects<br />

(time: pretest vs. posttest) ANOVA. We used the<br />

Fisher least significance difference (LSD) post-hoc procedure<br />

to evaluate any pairwise comparisons for significant<br />

effects for the full sample. If a lack of classroom or group<br />

comparability on pretreatment measures occurred, we conducted<br />

a one-factor ANOVA on the change scores, with<br />

classroom or group as the between-subjects factor. (We<br />

used change scores for analyz<strong>in</strong>g the two-wave data because<br />

their <strong>in</strong>terpretation is straightforward.) In addition, we<br />

analyzed scores from the strategy questionnaire with multivariate<br />

analysis of variance (MANOVA) with teacher or<br />

group as the between-subjects factor. To estimate the practical<br />

significance of effects for classroom and group, we<br />

computed effect sizes (ESs) by subtract<strong>in</strong>g the difference<br />

between the posttest means, then divid<strong>in</strong>g by the pooled<br />

standard deviation of the posttest. On the computation<br />

measure, we calculated improvement effects by divid<strong>in</strong>g<br />

the difference between the improvement means by the<br />

pooled standard deviation of the improvement divided by<br />

the square root of 2(1– rxy; Glass, McGraw, & Smith,<br />

1981).<br />

Results<br />

Table 2 shows the means and standard deviations for all<br />

measures for the full sample; Table 3 shows means and standard<br />

deviations for the measures by group (LD and LA).<br />

Pretreatment Differences Among Classrooms on<br />

Mathematics Performance<br />

Results <strong>in</strong>dicated that differences between classrooms on<br />

the WPS-CRT, F(2), 35) = 1.95, ns, and WPS-F, F(2, 35 =<br />

0.99, ns, were not significant prior to the study.<br />

On the computation pretest, however, there was a significant<br />

effect for classrooms, F(2, 35) = 4.87, p < .05. LSD<br />

follow-up tests <strong>in</strong>dicated that mean scores for Classroom 3<br />

were significantly different than those for Classroom 1 (p <<br />

.05) and Classroom 2 (p < .01); Classroom 3 > Classroom<br />

1 > Classroom 2. However, the difference between Classrooms<br />

1 and 2 was not significant. We found large effect<br />

sizes of 1.08 and 1.71 for Classroom 3 when compared with<br />

Classrooms 1 and 2, respectively.<br />

Posttreatment Differences Among Classrooms on<br />

Mathematics Performance<br />

Results of the repeated-measures ANOVA applied to the<br />

pretest and posttest scores demonstrated a significant ma<strong>in</strong>


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 293<br />

May/June 2007 [Vol. 100 (No. 5)] 293<br />

effect for time on the WPS-CRT, F(1, 35) = 40.90, p < .000,<br />

and WPS-F, F(1, 35) = 12.56, p < .01. Effect sizes for time<br />

at posttest when compared with pretest were moderate for<br />

the WPS-CRT (ES = 0.77) and large for the WPS-F (ES =<br />

0.83). There was no significant ma<strong>in</strong> effect for classroom or<br />

<strong>in</strong>teraction between classroom and time. ANOVA values<br />

for the classroom ma<strong>in</strong> effect on WPS-CRT and WPS-F<br />

were, F(2, 35) = 2.14, ns, and F(2, 35) = 0.11, ns, respectively.<br />

The correspond<strong>in</strong>g values for the <strong>in</strong>teraction effect<br />

on the two measures were, F(2, 35) = 0.11, ns, and F(2, 35)<br />

= 0.64, ns.<br />

On the computation measure, results of an ANOVA on<br />

change scores <strong>in</strong>dicated a significant effect for classrooms,<br />

F(2, 35) = 13.19, p < .000. Follow-up tests <strong>in</strong>dicated that<br />

the differences between Classrooms 1 and 2 and between<br />

Classrooms 1 and 3 only were significant (p < .01). (Mean<br />

posttest scores for Classroom 1 > Classroom 2 > Classroom<br />

3.) We found large effects of 1.49 and 1.96 for Classroom 1<br />

when compared with Classrooms 2 and 3, respectively.<br />

Posttreatment Differences Between Classrooms on<br />

the Strategy Satisfaction Questionnaire<br />

Results of the MANOVA applied to the Strategy Satisfaction<br />

Questionnaire posttreatment scores revealed no significant<br />

differences among classrooms, Wilks’s lambda =<br />

.69, approximate F(2, 35) = 1.29, ns.<br />

Pretreatment Differences Between Groups (LD and LA) on<br />

Mathematics Performance<br />

Results <strong>in</strong>dicated that differences between groups on the<br />

WPS-CRT, F(1, 16) = 1.05, ns, and WPS-F, F(1, 16) =<br />

0.36, ns, were not significant prior to the study.<br />

On the computation pretest, however, there was a significant<br />

ma<strong>in</strong> effect for group, F(1, 16) = 15.09, p < .01.<br />

The mean computation score for LD students was higher<br />

than that for LA students, with a large effect size of 1.83 for<br />

LD students.<br />

Posttreatment Differences Between Groups on<br />

Mathematics Performance LD and LA Sample<br />

Results of repeated-measures ANOVA applied to the<br />

pretest and posttest mathematics test scores demonstrated a<br />

significant ma<strong>in</strong> effect for time on the WPS-CRT, F(1, 16)<br />

= 26.94, p < .000, and WPS-F, F(1, 16) = 8.56, p < .01.<br />

Effect sizes for time at posttest when compared with pretest<br />

were large for the WPS-CRT (ES = 1.12) and moderate for<br />

the WPS-F (ES = 0.76). There were no significant ma<strong>in</strong><br />

effects for group or <strong>in</strong>teraction between group and time.<br />

ANOVA values for the ma<strong>in</strong> effect of group on WPS-CRT<br />

and WPS-F were F(1, 16) = 1.32, ns, and F(1, 16) 0.96, ns,<br />

respectively. The correspond<strong>in</strong>g values for the <strong>in</strong>teraction<br />

effect on the two measures were, F(1, 16) = 0.00, ns, and<br />

F(1, 16) = 0.89, ns. On the computation measure, results of<br />

an ANOVA on change scores <strong>in</strong>dicated the lack of a significant<br />

ma<strong>in</strong> effect for group, F(1, 16) = 2.11, ns.<br />

Posttreatment Differences Between Groups on<br />

the Strategy Satisfaction Questionnaire<br />

On the strategy satisfaction posttreatment scores, we<br />

found significant differences between groups, Wilks’s lambda<br />

= .43, approximate F(1, 16) = 3.16, p < .05. Results of<br />

separate univariate ANOVAs <strong>in</strong>dicated significant differences<br />

between groups regard<strong>in</strong>g total strategy satisfaction<br />

scores, F(1, 16) = 13.02, p < .01, enjoy<strong>in</strong>g the strategy, F(1,<br />

16) = 6.15, p < .05, usefulness of the strategy <strong>in</strong> solv<strong>in</strong>g<br />

word problems, F(1, 16) = 16.49, p < .01, and cont<strong>in</strong>u<strong>in</strong>g<br />

to use the strategy <strong>in</strong> solv<strong>in</strong>g word problems, F(1, 16) =<br />

8.24, p < .01. Specifically, LD students rated the strategy<br />

higher than did LA students.<br />

Discussion<br />

Results must be <strong>in</strong>terpreted <strong>in</strong> light of two serious limitations.<br />

First, the full sample (N = 38) and group sample<br />

sizes (n = 9 each for LD and LA students) were small. Second,<br />

the design was unbalanced; classroom sizes ranged<br />

from 16 to 6 students. As such, the f<strong>in</strong>d<strong>in</strong>gs <strong>in</strong>dicate only<br />

prelim<strong>in</strong>ary evidence regard<strong>in</strong>g the effectiveness of SBI.<br />

With<strong>in</strong> the constra<strong>in</strong>ts of the limitations, results from<br />

Study 1 provide evidence that SBI led to improvements <strong>in</strong><br />

word problem-solv<strong>in</strong>g performance for the three classrooms.<br />

Although the overall treatment fidelity was high,<br />

teach<strong>in</strong>g styles varied concern<strong>in</strong>g teachers’ adherence to<br />

the scripted curriculum (i.e., read verbatim or used their<br />

own explanations). The high level of treatment fidelity<br />

f<strong>in</strong>d<strong>in</strong>g suggests that SBI accounted for improved student<br />

learn<strong>in</strong>g, which is encourag<strong>in</strong>g.<br />

Furthermore, SBI was effective <strong>in</strong> enhanc<strong>in</strong>g the word<br />

problem-solv<strong>in</strong>g performance of students with LD, whether<br />

they received <strong>in</strong>struction <strong>in</strong> general education mathematics<br />

classrooms or <strong>in</strong> a special education classroom. When<br />

we separated outcomes on the word problem-solv<strong>in</strong>g measures<br />

for students with LD and their LA peers, the effects of<br />

the word problem-solv<strong>in</strong>g curriculum on students’ performance<br />

was comparable for both groups. Those results are<br />

notable because teachers implemented the treatment <strong>in</strong> a<br />

whole-class format and taught students to solve only onestep<br />

problems, although students were tested on one-step<br />

and two-step problems. The f<strong>in</strong>d<strong>in</strong>gs support and extend<br />

previous research regard<strong>in</strong>g the effectiveness of SBI <strong>in</strong> solv<strong>in</strong>g<br />

arithmetic word problems (e.g., Fuchs, Fuchs, F<strong>in</strong>elli,<br />

Courey, & Hamlett, 2004; Jitendra et al., 1998; Zawaiza &<br />

Gerber, 1993).<br />

Our results <strong>in</strong>dicated that classrooms and groups were<br />

not comparable on their computational skills prior to the<br />

study and that computational skill improvement varied as<br />

a function of classroom and group. In general, students with<br />

LD at pretreatment demonstrated better computational


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 294<br />

294 The Journal of Educational Research<br />

skills than did the other students at pretreatment. We<br />

expected that result given that many of those students were<br />

older students receiv<strong>in</strong>g <strong>in</strong>struction <strong>in</strong> a special education<br />

classroom <strong>in</strong> which the focus of <strong>in</strong>struction was on the<br />

acquisition of basic skills. However, the general f<strong>in</strong>d<strong>in</strong>g<br />

that computation improvement was evident for all students<br />

was encourag<strong>in</strong>g because opportunities for word problem<br />

solv<strong>in</strong>g facilitate computation skills. The effect size for<br />

posttest over pretest was large for the entire sample (ES =<br />

2.98) and for the LD and LA students (ES = 2.16). Evidently,<br />

hav<strong>in</strong>g students solve story problems enhanced the<br />

development of number operations or computational skills.<br />

The use of schematic diagrams, which provided mean<strong>in</strong>g to<br />

number sentences, had an added value <strong>in</strong> promot<strong>in</strong>g computation<br />

(Van De Walle, 2004).<br />

F<strong>in</strong>ally, the positive evaluation of SBI by students <strong>in</strong> the<br />

study seemed to play a role <strong>in</strong> enhanc<strong>in</strong>g mathematics performance<br />

as <strong>in</strong> several previous <strong>in</strong>vestigations (e.g., Case,<br />

Harris, & Graham, 1992; Jitendra et al., 1999). Students<br />

with learn<strong>in</strong>g disabilities especially were more enthused<br />

about SBI than were their peers with regard to treatment<br />

acceptability and benefits. Wood, Frank, and Wacker<br />

(1998) stated that “Student preference is an important factor,<br />

because students are not as likely to exhibit effort over<br />

time with strategies that they do not like or do not feel are<br />

helpful” (p. 336).<br />

Overall, the study helped us learn about effective ways to<br />

enhance the problem-solv<strong>in</strong>g curriculum as well as facilitate<br />

teacher implementation and student learn<strong>in</strong>g. Teachers<br />

became conversant about modify<strong>in</strong>g the curriculum<br />

only when they had completed the problem-solv<strong>in</strong>g unit<br />

on teach<strong>in</strong>g change and group problems. The lessons that<br />

we learned from our observations and teacher <strong>in</strong>put (prior<br />

to <strong>in</strong>struction on compare problems) dur<strong>in</strong>g SBI implementation<br />

allowed us to make several modifications that<br />

we organized <strong>in</strong> the follow<strong>in</strong>g paragraphs accord<strong>in</strong>g to curriculum,<br />

teacher, and student enhancements.<br />

Curriculum<br />

Although Marshall (1995) discussed the importance of<br />

present<strong>in</strong>g problem schema <strong>in</strong>struction <strong>in</strong> concert with the<br />

three problem types followed by problem-solution <strong>in</strong>struction,<br />

teachers raised concerns that problem solution was<br />

seem<strong>in</strong>gly removed from problem schema <strong>in</strong>struction for<br />

each problem type. Therefore, we revised the curriculum<br />

such that <strong>in</strong>struction for the change problem type began<br />

with problem schema <strong>in</strong>struction, followed by problem<br />

solution <strong>in</strong>struction. We used that same sequence for group<br />

and compare problem types.<br />

We redesigned the self-monitor<strong>in</strong>g strategy checklist to<br />

<strong>in</strong>clude four steps, and used an acronym, FOPS (F<strong>in</strong>d the<br />

problem type, Organize <strong>in</strong>formation <strong>in</strong> the problem us<strong>in</strong>g<br />

the schema diagram, Plan to solve the problem, and Solve<br />

the <strong>Problem</strong>), to help students remember the steps. Furthermore,<br />

we elaborated on each step to align with<br />

doma<strong>in</strong>-specific knowledge consistent with SBI. For example,<br />

to f<strong>in</strong>d the problem type (Step 1), we prompted students<br />

to exam<strong>in</strong>e <strong>in</strong>formation perta<strong>in</strong><strong>in</strong>g to each set<br />

(beg<strong>in</strong>n<strong>in</strong>g, change, and end<strong>in</strong>g) <strong>in</strong> the problem (e.g.,<br />

change).<br />

Given that the compare problem type was difficult for<br />

many students dur<strong>in</strong>g the problem schema phase, we collaborated<br />

with teachers to develop a compare structure that<br />

was coherent to third-grade students. For example, we<br />

focused on three sets (bigger, smaller, and difference) of<br />

<strong>in</strong>formation and elim<strong>in</strong>ated reference to difficult terms<br />

(e.g., compared, referent) when discuss<strong>in</strong>g the problem. In<br />

addition, we added oral exercises prior to written work to<br />

emphasize the critical features and relations <strong>in</strong> the problem<br />

to ensure that students were familiar with the <strong>in</strong>formation<br />

needed for problem comprehension, a key aspect of SBI.<br />

We found that teachers were spend<strong>in</strong>g considerable<br />

amounts of <strong>in</strong>structional time expla<strong>in</strong><strong>in</strong>g unfamiliar terms<br />

to students who did not have the necessary experiential<br />

background, which detracted from the focus on problem<br />

solv<strong>in</strong>g. Therefore, we modified the textbook problems to<br />

meet the needs of the students. For example, unknown<br />

words such as “alpaca” were replaced by more familiar terms<br />

(sheep).<br />

An important issue that emerged was teacher accountability<br />

for student performance on statewide test<strong>in</strong>g. Teachers<br />

believed that the use of word problems presented only<br />

<strong>in</strong> text format would not adequately prepare students to<br />

generalize to word problems on the state mathematics test.<br />

Therefore, revisions to the problem-solv<strong>in</strong>g curriculum<br />

content entailed <strong>in</strong>clusion of items that presented <strong>in</strong>formation<br />

<strong>in</strong> tables, graphs, and pictographs. Also, because the<br />

state test required the use of mathematical vocabulary (e.g.,<br />

addend) and emphasized written communication, we modified<br />

our <strong>in</strong>struction to <strong>in</strong>clude key mathematical terms and<br />

provided students with practice <strong>in</strong> writ<strong>in</strong>g explanations for<br />

how the problem was solved.<br />

Furthermore, teachers noted that the time required to<br />

implement SBI was unrealistic given the need to cover<br />

other topics <strong>in</strong> the school curriculum. In the redesign of the<br />

problem-solv<strong>in</strong>g curriculum, we reduced problem schema<br />

<strong>in</strong>struction for each problem type from three 30-m<strong>in</strong><br />

lessons to one 50-m<strong>in</strong> lesson and developed four lessons for<br />

each problem type that addressed problem solution <strong>in</strong>struction.<br />

In addition, we <strong>in</strong>corporated fad<strong>in</strong>g of schematic diagrams<br />

to ensure that students were able to apply learned<br />

solution procedures <strong>in</strong>dependent of diagrams. Also, we<br />

elim<strong>in</strong>ated homework problems from the curriculum<br />

because of teacher concerns regard<strong>in</strong>g students’ <strong>in</strong>consistent<br />

completion of homework.<br />

Teachers<br />

Our observations revealed that teachers need ongo<strong>in</strong>g<br />

support dur<strong>in</strong>g the <strong>in</strong>itial implementation of a newly developed<br />

<strong>in</strong>tervention. Although we provided teach<strong>in</strong>g scripts


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 295<br />

May/June 2007 [Vol. 100 (No. 5)] 295<br />

to ensure consistency <strong>in</strong> implement<strong>in</strong>g the critical content,<br />

we suggested that teachers use them as a framework for<br />

<strong>in</strong>structional implementation. However, our observations<br />

<strong>in</strong>dicated that Teacher 1 read the script verbatim and followed<br />

it <strong>in</strong> its entirety, whereas Teacher 2 followed the<br />

script <strong>in</strong>consistently and required rem<strong>in</strong>ders to adhere to<br />

the relevant <strong>in</strong>formation needed to promote problem solv<strong>in</strong>g.<br />

In contrast, Teacher 3 (special education teacher)<br />

familiarized herself with the script and used her own explanations<br />

and elaborations to implement the <strong>in</strong>tervention<br />

with ease. Obviously, general education teachers <strong>in</strong> this<br />

study needed more support to implement the <strong>in</strong>tervention.<br />

Also, we noticed that some students <strong>in</strong> one classroom were<br />

struggl<strong>in</strong>g dur<strong>in</strong>g partner work and had to wait until wholeclass<br />

discussion to get corrective feedback, <strong>in</strong>dicat<strong>in</strong>g the<br />

need for provid<strong>in</strong>g teachers with explicit guidel<strong>in</strong>es about<br />

direct monitor<strong>in</strong>g student work.<br />

In addition, we found that the general education teachers<br />

seemed more at ease <strong>in</strong> communicat<strong>in</strong>g their concerns<br />

to the two research assistants who supported them dur<strong>in</strong>g<br />

the course of the project than to the primary researcher,<br />

which raised the issue of how and with whom communication<br />

should be facilitated.<br />

Students<br />

Our observations <strong>in</strong>dicated that for many students, especially<br />

those with LD, scaffold<strong>in</strong>g <strong>in</strong>struction (model<strong>in</strong>g, use<br />

of schematic diagrams and checklists) was critical as they<br />

learned to apply the strategy. Explicit model<strong>in</strong>g and explanations<br />

us<strong>in</strong>g several examples enhanced their problemsolv<strong>in</strong>g<br />

skills. Follow<strong>in</strong>g teacher-led <strong>in</strong>struction, we noticed<br />

that several students with LD, as well as low-achiev<strong>in</strong>g students,<br />

were enthused and participated actively, as evidenced<br />

by their raised hands <strong>in</strong> response to teacher question<strong>in</strong>g.<br />

DESIGN STUDY 2<br />

Method<br />

Participants<br />

Students <strong>in</strong> Study 2 were 56 third-grade students <strong>in</strong> two<br />

heterogeneous classrooms <strong>in</strong> a parochial school located <strong>in</strong><br />

a small city <strong>in</strong> Florida that served 570 students <strong>in</strong> <strong>Grade</strong>s<br />

pre-K–8. Approximately 20% of the student population<br />

was African American, Hispanic, or Asian. The total sample<br />

<strong>in</strong>cluded 27 boys and 29 girls. The mean chronological<br />

age of the students was 108 months (range = 96 to 113<br />

months). Forty-four of the students (78%) were Caucasian,<br />

5 (8%) were African American, 6 (10%) were Hispanic,<br />

and 1 was Asian American (2%). Although not formally<br />

identified at the school, the two third-grade classroom<br />

teachers reported that 9 of the 56 third-grade students had<br />

either LD, attention deficit disorder, or were considered<br />

low achievers. On the basis of scores from the <strong>Problem</strong><br />

<strong>Solv<strong>in</strong>g</strong> and Data Interpretation (PSDI) mathematics subtest<br />

of the Iowa Test of Basic Skills (ITBS), we designated<br />

each student’s <strong>in</strong>itial mathematics achievement status as<br />

low perform<strong>in</strong>g (LO; below the 34th percentile), average<br />

perform<strong>in</strong>g (AV; between the 35th and 66th percentiles),<br />

or high perform<strong>in</strong>g (HI; above the 66th percentile). Tables<br />

4 and 5 summarize participant demographic <strong>in</strong>formation by<br />

the two classrooms and by student type (LO, AV, and HI<br />

students).<br />

Separate one-way analysis of variances (ANOVAs) <strong>in</strong>dicated<br />

no significant differences between classrooms on the<br />

mathematics and read<strong>in</strong>g subtests of the ITBS, <strong>Problem</strong><br />

<strong>Solv<strong>in</strong>g</strong> and Data Interpretation, F(1, 54) = 0.11, ns; Computation,<br />

F(1, 54) = 0.04, ns; Vocabulary, F(1, 54) = 0.35,<br />

ns; and Comprehension, F(1, 54) = 0.01, ns. In addition,<br />

there was a lack of a significant difference between classrooms<br />

on chronological age, F(1, 54) = 0.31, ns. Chi-square<br />

analyses revealed no significant between-classroom differences<br />

on gender, χ 2 (1, N = 56) = .27, ns, and ethnicity,<br />

χ 2 (3, N = 56) = 1.96, ns.<br />

Similarly, separate one-way ANOVAs <strong>in</strong>dicated significant<br />

differences between student type (i.e., LO, AV, and<br />

HI) on the PSDI, F(2, 53) = 239.81, p < .000; computation,<br />

F(2, 53) = 9.498, p < .000; vocabulary, F(2, 53) =<br />

17.80, p < .000; and comprehension, F(2, 53) = 25.54, p <<br />

.000. For all analyses, LSD post-hoc tests revealed that the<br />

differences between the three groups were significant.<br />

(Performance of HI students > AV students > LO students.)<br />

In addition, there was no significant between-student<br />

type difference on chronological age, F(2, 53) = 0.95,<br />

ns. Chi-square analyses revealed no significant betweenstudent<br />

type differences on gender, χ 2 (2, N = 56) = 0.27,<br />

ns, or race χ 2 (6, N = 56) = 1.96, ns.<br />

The two general education teachers who participated <strong>in</strong><br />

Study 2 were women; one teacher held a master’s degree<br />

with 10 years’ teach<strong>in</strong>g experience, and the other teacher,<br />

a bachelor’s degree with 30 years’ experience. The teachers<br />

attended 1 hr of <strong>in</strong>service tra<strong>in</strong><strong>in</strong>g on ways to implement<br />

the <strong>in</strong>tervention; they received support from one of the<br />

researchers periodically throughout the study.<br />

Procedures and Measures<br />

Procedures and measures were similar to those used <strong>in</strong><br />

Study 1. In this section, we report only how components of<br />

the two studies differed. First, although students received the<br />

same amount of <strong>in</strong>structional time as those <strong>in</strong> Study 1,<br />

teachers began Study 2 later <strong>in</strong> the school year. Second,<br />

teachers <strong>in</strong> Study 2 used the textbook series, “Mathematics—The<br />

Path to Math Success! <strong>Grade</strong> 3” by Silver, Burdett,<br />

and G<strong>in</strong>n (as cited <strong>in</strong> Fennell et al., 1998) for daily mathematics<br />

<strong>in</strong>struction. <strong>Third</strong>, <strong>in</strong>structional procedures differed<br />

from Study 1 with regard to compare <strong>in</strong>struction. On the<br />

basis of student difficulty with the compare problem type,<br />

teachers <strong>in</strong> Study 1 used the revised compare diagram and<br />

<strong>in</strong>struction. However, teachers <strong>in</strong> Study 2 decided to implement<br />

the orig<strong>in</strong>al compare <strong>in</strong>struction and determ<strong>in</strong>e how


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 296<br />

296 The Journal of Educational Research<br />

TABLE 4. Student Demographic Information and Mathematics Performance by Classroom: Site 2<br />

Classroom 1 Classroom 2<br />

Variable n % M SD n % M SD ES<br />

Age (<strong>in</strong> months) 106.69 4.35 107.26 3.21<br />

Gender<br />

Male 13 44.82 14 51.85<br />

Female 16 55.17 13 48.15<br />

Race<br />

EA 23 79.31 21 77.77<br />

AA 3 10.34 2 7.40<br />

Hispanic 2 6.90 4 14.81<br />

Asian 1 0.03 0 0.00<br />

ITBS<br />

Mathematics subtests<br />

PSDI 554.01 114.55 565.99 100.64<br />

Computation 550.07 99.79 543.67 84.14<br />

Read<strong>in</strong>g subtests<br />

Vocabulary 568.22 95.34 582.00 85.95<br />

Comprehension 558.04 95.86 563.14 91.86<br />

WPS-CRT (25/50)<br />

Pretest 30.05 10.61 25.50 7.65 +0.24<br />

Posttest 29.98 12.01 28.15 7.18 +0.09<br />

Growth −0.07 7.96 2.65 6.82 −0.18<br />

WPS-F (16/32)<br />

Pretest 19.76 8.45 19.67 6.40 +0.01<br />

Posttest 22.59 8.50 21.17 3.97 +0.10<br />

Growth 2.83 5.03 1.50 4.74 +0.13<br />

Computation a (25/43)<br />

Pretest 31.79 5.55 33.48 7.50 −0.12<br />

Posttest 33.45 5.88 36.22 6.57 −0.21<br />

Growth 1.66 4.04 2.74 2.65 −0.15<br />

Note. EA = European American; AA = African American; ITBS = Iowa Test of Basic Skills (Houghton Miffl<strong>in</strong>, 1999); PSDI = problem solv<strong>in</strong>g<br />

and data <strong>in</strong>terpretation; WPS-CRT = word problem solv<strong>in</strong>g criterion test; WPS-F = word problem solv<strong>in</strong>g fluency measure; ES = effect size. Parenthetical<br />

numbers separated by slashes are total number of test items and total possible score.<br />

a<br />

Digits correct.<br />

their students, who were predom<strong>in</strong>ately average and high<br />

perform<strong>in</strong>g, would respond. Fourth, <strong>in</strong> Study 2, we did not<br />

use the student strategy satisfaction questionnaire adm<strong>in</strong>istered<br />

<strong>in</strong> Study 1 because of time constra<strong>in</strong>ts at the end of the<br />

school year. Treatment fidelity was 97.5% (range =<br />

91–100%) across the two teachers.<br />

Results<br />

Table 4 shows the means and standard deviations for all<br />

measures for the full sample. Table 5 displays means and<br />

standard deviations for the measures by group (LD and LA).<br />

Pretreatment Differences Between Classrooms on<br />

Mathematics Performance<br />

Results <strong>in</strong>dicated that differences between classrooms on<br />

the WPS-CRT, F(1, 54) = 3.35, ns, WPS-F, F(1, 54) = 0.00,<br />

ns, and computation pretest scores, F(1, 54) = 0.92 ns, were<br />

not significant prior to the study.<br />

Posttreatment Differences Between Classrooms on<br />

Mathematics Performance<br />

Results of ANOVA applied to the pretest and posttest<br />

data yielded no significant classroom by time of test<strong>in</strong>g<br />

<strong>in</strong>teraction for WPS-CRT, F(1, 54) = 1.39, ns, <strong>in</strong>dicat<strong>in</strong>g<br />

that classroom had no effect on changes <strong>in</strong> correct responses<br />

from pretest to posttest.<br />

However, the analyses yielded a significant ma<strong>in</strong> effect<br />

for time of test<strong>in</strong>g for WPS-F, F(1, 54) = 10.93, p < .00, and<br />

for computation measures, F(1, 54) = 22.81, p < .000, <strong>in</strong>dicat<strong>in</strong>g<br />

improvement from pretest to posttest for the entire<br />

sample of students regardless of classroom condition (average<br />

for WPS-F = 2.19; SD = 4.90; computation = 2.18, SD<br />

= 3.45). Effect sizes for time at posttest when compared<br />

with pretest were small to moderate for the WPS-F (ES =<br />

0.31) and computation measures (ES = 0.34). The ma<strong>in</strong><br />

effect for time of test<strong>in</strong>g on the WPS-CRT was not significant,<br />

F(1, 54) = 1.68, p = .20. We found no significant<br />

ma<strong>in</strong> effect for classroom for all three tests: WPS-CRT, F(1,


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 297<br />

May/June 2007 [Vol. 100 (No. 5)] 297<br />

54) = 1.80, ns, for WPS-F, F(1, 54) = .18, ns, and for computation,<br />

F(1, 54) = 1.84, ns.<br />

Pretreatment Differences Between Groups (LO, AV, and HI)<br />

on Mathematics Performance<br />

Results <strong>in</strong>dicated that, as expected, differences between<br />

groups on the WPS-CRT, F(2, 53) = 5.68, p < .000 and computation,<br />

F(2, 53) = 5.68, p < .01, were significant prior to<br />

the study.<br />

LSD follow-up tests <strong>in</strong>dicated that mean scores for HI<br />

were significantly different than were those for AV (p <<br />

.00) and LO (p < .00) on the WPS-CRT (HI > AV > LO).<br />

However, the difference between LO and AV was not significant.<br />

We found large effect sizes of 1.14 and 1.15 for HI<br />

when compared with AV and LO, respectively. For the<br />

WPS-F measure, follow-up tests <strong>in</strong>dicated that the mean<br />

scores for HI were significantly different than those for AV<br />

(p < .000) and LO (p < .000; HI > AV > LO). In addition,<br />

the difference between LO and AV was significant (p <<br />

.05), favor<strong>in</strong>g the AV group. On the WPS-F, we found large<br />

effect sizes of 1.26 for HI compared with AV, 2.51 for HI<br />

compared with LO, and 0.94 for AV compared with LO.<br />

F<strong>in</strong>ally, follow-up tests on the computation measure<br />

revealed that mean scores for HI were significantly different<br />

than were those for LO (p < .00; (HI > LO), but not<br />

between HI and AV (p = .15). The difference between LO<br />

and AV on the computation measure was also significant (p<br />

= .05); the AV group scored higher. We found large effect<br />

sizes of 1.67 and 0.91 for HI compared with the LO, and for<br />

AV compared with LO, respectively.<br />

Results of ANOVA applied to the WPS-CRT pretest<br />

and posttest data yielded no significant group by time of<br />

test<strong>in</strong>g <strong>in</strong>teraction, F(2, 53) = 1.19, ns. That f<strong>in</strong>d<strong>in</strong>g <strong>in</strong>dicated<br />

that group status (HI, AV, LO) had no effect on<br />

changes <strong>in</strong> correct scores from pretest to posttest. In addition,<br />

there was no significant ma<strong>in</strong> effect for time, F(1, 53)<br />

= 0.26, ns, <strong>in</strong>dicat<strong>in</strong>g lack of improvement from pretest to<br />

posttest for the entire sample of students, regardless of<br />

group status. However, the analysis yielded a significant<br />

ma<strong>in</strong> effect for group, F(2, 53) = 12.45, p < .000. The LSD<br />

follow-up test <strong>in</strong>dicated that WPS-CRT mean scores for HI<br />

students were significantly different from AV (p < .000)<br />

and LO (p < .000). We found effect sizes of 1.05 and 1.48<br />

for HI students when compared with AV and LO students,<br />

respectively, on the WPS-CRT. The difference between LO<br />

and AV, however, was not significant (p = .12).<br />

Results for WPS-F <strong>in</strong>dicated the follow<strong>in</strong>g significant<br />

effects: group by time of test<strong>in</strong>g <strong>in</strong>teraction, F(2, 53) =<br />

3.19, p < .05, time of test<strong>in</strong>g, F(1, 53) = 17.37, p < .000,<br />

and group status, F(2, 53) = 16.72 p < .000. Because the<br />

group by time of test<strong>in</strong>g <strong>in</strong>teraction supercedes ma<strong>in</strong><br />

effects, we conducted follow-up tests only for the significant<br />

<strong>in</strong>teraction. Follow-up tests <strong>in</strong>dicated that from<br />

pretest to posttest, averaged across all groups, the LO students<br />

improved reliably more than did the HI students (p =<br />

.02); the trend of improvement for LO students compared<br />

with AV students approached significance (p = .07); the<br />

growth of the AV and HI students was comparable (average<br />

growth for LO students = 5.61; SD = 5.81; AV students =<br />

2.13, SD = 4.04; HI students = 0.96, SD = 4.88). The effect<br />

size for LA growth compared with AV growth was 0.76; for<br />

LA compared with HI, 0.91; for AV compared with HI,<br />

0.26.<br />

The analysis of computation data yielded no significant<br />

group by time of test<strong>in</strong>g <strong>in</strong>teraction, F(1, 54) = 0.36, ns,<br />

<strong>in</strong>dicat<strong>in</strong>g that group status (HI, AV, LO) had no effect on<br />

changes <strong>in</strong> correct scores from pretest to posttest. However,<br />

results <strong>in</strong>dicated a significant ma<strong>in</strong> effect for time, F(1,<br />

53) = 18.84, p < .000, <strong>in</strong>dicat<strong>in</strong>g improvement from pretest<br />

to posttest for the entire sample of students, regardless of<br />

group status (M = 2.19; SD = 4.90). The effect size for time<br />

from posttest to pretest was 0.34. In addition, there was a<br />

significant ma<strong>in</strong> effect for group, F(2, 53) = 5.74, p < .01.<br />

The follow-up test for group <strong>in</strong>dicated that mean scores for<br />

HI students were significantly different than that for LO<br />

students (p < .00) but comparable to those for AV students<br />

(p = .20). In addition, the difference between LO and AV<br />

students was significant (p < .05). We found effect sizes of<br />

1.39 and 0.29 for HI students when compared with LO students,<br />

and for AV students compared with LO students,<br />

respectively.<br />

Discussion<br />

Students <strong>in</strong> Study 2 made small-to-moderate improvements<br />

<strong>in</strong> their word problem-solv<strong>in</strong>g and computation performance<br />

despite unforeseen problems that occurred dur<strong>in</strong>g<br />

this field-based research. The benefits of a SBI were particularly<br />

apparent for the low-perform<strong>in</strong>g students <strong>in</strong> the two<br />

heterogeneously grouped third-grade classrooms. As <strong>in</strong><br />

Study 1, third-grade teachers delivered <strong>in</strong>struction <strong>in</strong><br />

whole-class arrangement without special <strong>in</strong>structional<br />

adaptations for low-perform<strong>in</strong>g students. Despite the lack<br />

of <strong>in</strong>dividualized <strong>in</strong>struction, those students showed the<br />

greatest amount of growth from pretest to posttest on two<br />

of the three measures. Conversely, the results <strong>in</strong>dicated<br />

that the high-perform<strong>in</strong>g students made the least amount<br />

of improvement. The f<strong>in</strong>d<strong>in</strong>gs support previous studies<br />

regard<strong>in</strong>g the importance of explicit <strong>in</strong>struction for struggl<strong>in</strong>g<br />

learners (e.g., Swanson, 1999) and its triviality for<br />

high performers. Even young children with high <strong>in</strong>telligence<br />

are capable of us<strong>in</strong>g strategies spontaneously to perform<br />

given tasks without explicit <strong>in</strong>struction (Cho & Ahn,<br />

2003).<br />

Results for the WPS-CRT, <strong>in</strong> particular, were disappo<strong>in</strong>t<strong>in</strong>g.<br />

Our hypothesis that there would be a significant<br />

effect for time from pretest to the posttest was not realized.<br />

The lack of improvement over time may be attributed to<br />

several reasons. First, the length of the WPS-CRT task<br />

(i.e., 25 one-step and two-step word problems) <strong>in</strong> conjunction<br />

with adm<strong>in</strong>istration of the posttest only 2 weeks before


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 298<br />

298 The Journal of Educational Research<br />

TABLE 5. Student Demographic Characteristics and Mathematics Performance by Mathematics <strong>Grade</strong>-Level Status: Site 2<br />

Mathematics grade-level status<br />

High (n = 24) Average (n = 23) Low (n = 9)<br />

Variable n % M SD n % M SD n % M SD ES<br />

Age (<strong>in</strong> months) 107.14 2.88 107.23 4.88 106.00 3.67<br />

Gender<br />

Male 14 58.33 8 34.78 6 66.67<br />

Female 10 41.67 15 65.22 3 33.33<br />

Race<br />

EA 20 83.33 15 65.22 8 88.89<br />

AA 1 4.17 4 17.39 1 11.11<br />

Hispanic 2 8.33 4 17.39 0 0.00<br />

Asian 1 4.17 0 0.00 0 0.00<br />

ITBS<br />

Mathematics<br />

PSDI 660.47 62.74 517.47 21.68 466.92 67.67<br />

Computation 598.40 96.41 524.66 60.49 399.45 44.81<br />

Read<strong>in</strong>g<br />

Vocabulary 639.00 57.42 542.54 70.77 486.54 93.62<br />

Comprehension 631.13 67.90 524.23 70.48 464.86 60.66<br />

WPS-CRT (25/50)<br />

Pretest 33.27 8.17 24.50 7.26 22.00 11.18 LO vs. AV: −0.29<br />

LO vs. HI: −1.25<br />

AV vs. HI: −1.13<br />

Posttest 34.67 5.64 26.89 8.90 19.89 12.97 LO vs. AV: −0.69<br />

LO vs. HI: −1.81<br />

AV vs. HI: −1.05


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 299<br />

May/June 2007 [Vol. 100 (No. 5)] 299<br />

Growth 1.40 6.10 2.39 8.52 −2.11 7.93 LO vs. AV: −0.54<br />

LO vs. HI: −0.53<br />

AV vs. HI: +0.13<br />

WPS-F (16/32)<br />

Pretest 24.85 5.49 17.35 6.42 12.06 4.66 LO vs. AV: −0.88<br />

LO vs. HI: −2.42<br />

AV vs. HI: −1.26<br />

Posttest 25.81 3.89 19.48 6.09 17.67 8.91 LO vs. AV: −0.26<br />

LO vs. HI: −1.45<br />

AV vs. HI: −1.25<br />

Growth 0.96 4.88 2.13 4.04 5.61 5.81 LO vs. AV: +0.76<br />

LO vs. HI: +0.91<br />

AV vs. HI: +0.26<br />

Computation a (25/43)<br />

Pretest 34.96 5.35 32.35 7.25 27.00 4.06 LO vs. AV: −0.82<br />

LO vs. HI: −1.58<br />

AV vs. HI: −0.41<br />

Posttest 36.67 5.40 34.91 6.82 29.44 4.50 LO vs. AV: −0.87<br />

LO vs. HI: −1.39<br />

AV vs. HI: −0.29<br />

Growth 1.71 2.94 2.57 4.02 2.44 3.40 LO vs. AV: −0.03<br />

LO vs. HI: +0.24<br />

AV vs. HI: +0.24<br />

Note. Parenthetical numbers separated by slashes are total number of test items and total possible score. EA = European American; AA = African American; ITBS = Iowa Test of Basic Skills (Houghton<br />

Miffl<strong>in</strong>, 1999); PSDI = <strong>Problem</strong> solv<strong>in</strong>g and data <strong>in</strong>terpretation; HI = high; AV-average; LO = low; WPS-CRT = word problem solv<strong>in</strong>g criterion test; WPS-F = word problem solv<strong>in</strong>g fluency measure;<br />

ES = effect size.<br />

a Digits correct.<br />

* p = .05;<br />

** p = .01;<br />

*** p = .001.


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 300<br />

300 The Journal of Educational Research<br />

school ended for the summer may not have been motivat<strong>in</strong>g<br />

for students, <strong>in</strong> spite of explicit <strong>in</strong>structions to do their<br />

best. Teachers reported students’ less-than-determ<strong>in</strong>ed<br />

attempts to complete the test. As such, that problem could<br />

have negatively affected student performance on the WPS-<br />

CRT posttest, which is not uncommon to <strong>in</strong>tervention<br />

research (Dunst & Trivette, 1994).<br />

Another plausible explanation for a lack of improvement<br />

was that the high-perform<strong>in</strong>g students may have<br />

known how to solve third-grade word problems before the<br />

study began and did not require extensive strategy <strong>in</strong>struction.<br />

A closer look at the data suggests that that may be the<br />

case. Fifteen of the high-perform<strong>in</strong>g students scored close<br />

to an accepted criterion level of 80% correct before the<br />

study began, <strong>in</strong>dicat<strong>in</strong>g possible ceil<strong>in</strong>g effects (M = 39.91,<br />

SD = 3.99; range = 34 to 47 on the 50-po<strong>in</strong>t WPS-CRT).<br />

As such, pretest to posttest improvement was not evident<br />

for the students, and ceil<strong>in</strong>g effects limited the ability to<br />

ferret out differences among higher scor<strong>in</strong>g students.<br />

Results on the WPS-F measures were more encourag<strong>in</strong>g.<br />

All students improved their word problem-solv<strong>in</strong>g performance<br />

on the shorter, 8-item, probes from pretest to<br />

posttest. The LO group made the most progress on this<br />

measure, as evidenced by the large effect size of 0.79 when<br />

compared with effect sizes of 0.34 and 0.20 for the AV and<br />

HI groups, respectively. Given that the measure <strong>in</strong>cluded<br />

easier items than did the WPS-CRT, was shorter <strong>in</strong> length,<br />

and required only 10 m<strong>in</strong> to complete each probe, thereby<br />

reduc<strong>in</strong>g fatigue, it may have positively <strong>in</strong>fluenced the<br />

results of the WPS-F.<br />

The pretest-to-posttest results for the computation measure<br />

revealed a pattern similar to those for the WPS-F. All<br />

groups improved significantly over time; however, the LO<br />

group demonstrated the largest ga<strong>in</strong>s from pretest to<br />

posttest, with a medium effect size of .57 when compared<br />

with small-to-moderate effect sizes of .36 and .32 for the<br />

AV and HI students, respectively. Aga<strong>in</strong>, word problem<br />

solv<strong>in</strong>g seems to play a role <strong>in</strong> the development of number<br />

operations, such as computational skills (Van de Walle,<br />

2004), particularly for low-perform<strong>in</strong>g students.<br />

As <strong>in</strong> Study 1, we learned ways to enhance curriculum,<br />

<strong>in</strong>struction, and measures. For example, whereas teachers<br />

from both sites were concerned with the length of the<br />

<strong>in</strong>tervention, Study 2 teachers were less concerned about<br />

coverage of other topics. Instead, teachers from Study 2<br />

reported the need to reduce what they considered redundancy<br />

<strong>in</strong> the curriculum as well as highly directive <strong>in</strong>struction<br />

because the majority of their students (i.e., average<br />

and high performers) occasionally appeared unmotivated<br />

dur<strong>in</strong>g the <strong>in</strong>structional sessions. Considerable evidence<br />

shows that pac<strong>in</strong>g curriculum and <strong>in</strong>struction to match the<br />

needs of the student is one way to ensure that the needs of<br />

highly able students are addressed (Rogers, 2002).<br />

Attempts to replicate SBI <strong>in</strong> heterogeneous classrooms may<br />

necessitate changes to differentiate the <strong>in</strong>struction for<br />

mixed-ability group<strong>in</strong>gs (e.g., Toml<strong>in</strong>son, 1995).<br />

Our observations <strong>in</strong> Study 2, like those <strong>in</strong> Study 1, <strong>in</strong>dicated<br />

that LO students benefited from the explicit model<strong>in</strong>g,<br />

guided practice, use of schematic diagrams, and checklists<br />

that are characterized by SBI. We observed that LO<br />

students tended to participate more than was typical dur<strong>in</strong>g<br />

SBI <strong>in</strong>struction, and their teachers stated that <strong>in</strong>struction<br />

had clear benefits for them. However, as a group, the performance<br />

of LO students on the three measures was <strong>in</strong>consistent<br />

and highly variable. They performed well on the<br />

WPS-F and computation measure but not on the WPS-<br />

CRT. On the WPS-CRT (pretest: M = 22.00, SD = 11.18;<br />

posttest: M = 19.89, SD = 12.97), performance of LO students<br />

was highly variable, as <strong>in</strong>dicated by large standard<br />

deviations. That result did not occur for the WPS-F<br />

(pretest: M = 12.06, SD = 4.66; posttest: M = 17.67, SD =<br />

8.91) or computation test (pretest: M = 27.00, SD = 4.06;<br />

posttest: M = 29.44, SD = 4.50). The length and difficulty<br />

of the WPS-CRT appeared to affect the performance of the<br />

LO group most directly; variability decreased when the<br />

measures were shorter or less challeng<strong>in</strong>g.<br />

F<strong>in</strong>ally, although the two teachers who participated <strong>in</strong><br />

Study 2 reported that they had few problems implement<strong>in</strong>g<br />

SBI <strong>in</strong>struction and had high levels of treatment fidelity<br />

(M = 98%), <strong>in</strong> h<strong>in</strong>dsight, the teachers may have had more<br />

to reveal had they received our consistent support. Provid<strong>in</strong>g<br />

teachers with more frequent opportunities to converse<br />

with us may have allowed for <strong>in</strong>structional adaptations that<br />

addressed vary<strong>in</strong>g student needs. Instruction that meets the<br />

needs of heterogeneous groups of students is not easily<br />

designed, and, consequently, may require more structured,<br />

collaborative efforts between teachers and researchers.<br />

General Discussion<br />

Design studies have the potential to allow researchers to<br />

develop <strong>in</strong>-depth understand<strong>in</strong>gs of teach<strong>in</strong>g and learn<strong>in</strong>g<br />

<strong>in</strong> classroom environments, that <strong>in</strong> turn, can assist them <strong>in</strong><br />

the design of experimental or quasi-experimental studies<br />

(Gersten et al., 2000). Specific examples of those benefits<br />

<strong>in</strong>clude enhanced knowledge about (a) <strong>in</strong>structional design<br />

features of an <strong>in</strong>tervention, (b) student and teacher<br />

responses to the <strong>in</strong>tervention, (c) measures used, and (d)<br />

the role of the researcher <strong>in</strong> facilitat<strong>in</strong>g the study. Like pilot<br />

studies, design experiments allow researchers to determ<strong>in</strong>e<br />

the feasibility of an experiment before implement<strong>in</strong>g it on<br />

a larger scale and under more carefully controlled conditions<br />

(Gersten, <strong>in</strong> press).<br />

We undoubtedly learned much about SBI for promot<strong>in</strong>g<br />

mathematics word problem solv<strong>in</strong>g by conduct<strong>in</strong>g the two<br />

design studies reported here. Although we may have a few<br />

unanswered questions, neither the knowledge ga<strong>in</strong>ed nor<br />

the uncerta<strong>in</strong>ties revealed would have occurred without<br />

carry<strong>in</strong>g out the <strong>in</strong>vestigative work of these studies. The<br />

most notable <strong>in</strong>sights revealed to us from conversations<br />

with the teachers and from our own observations <strong>in</strong>cluded<br />

the benefits of strategy fad<strong>in</strong>g to promote <strong>in</strong>dependent stu-


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 301<br />

May/June 2007 [Vol. 100 (No. 5)] 301<br />

dent use of SBI. Encourag<strong>in</strong>g <strong>in</strong>dependent, self-regulated<br />

use of strategies has been supported <strong>in</strong> the read<strong>in</strong>g literature<br />

for years and is perceived as an important long-term goal of<br />

strategy <strong>in</strong>struction (Pressley et al., 1992). We suggest that<br />

support<strong>in</strong>g the <strong>in</strong>dependent use of strategies is an important<br />

aspect of mathematics word problem-solv<strong>in</strong>g <strong>in</strong>struction<br />

as well.<br />

We also realized the need for changes <strong>in</strong> the compare<br />

problem <strong>in</strong>struction by collaborat<strong>in</strong>g first with the teachers<br />

at Site 1, and later with teachers at Site 2. We observed<br />

that the vocabulary words and compare schema diagram<br />

were difficult and confus<strong>in</strong>g for students at all achievement<br />

levels. Hav<strong>in</strong>g the opportunity to alter academic <strong>in</strong>terventions<br />

on the basis of student and teacher <strong>in</strong>put is a powerful<br />

outcome of design studies and will strengthen the<br />

<strong>in</strong>struction when exam<strong>in</strong>ed <strong>in</strong> subsequent studies. We also<br />

learned about the importance of provid<strong>in</strong>g extensive support<br />

to teachers dur<strong>in</strong>g the study implementation. We are<br />

conv<strong>in</strong>ced that the frequent and consistent support provided<br />

to teachers <strong>in</strong> Study 1 allowed their students to perform<br />

better. Alternatively, if we had offered more support <strong>in</strong><br />

Study 2, the teachers may have shared additional comments<br />

with us about the implementation of the study.<br />

Those comments could have prompted changes <strong>in</strong> the<br />

<strong>in</strong>struction, which may have better addressed the academic<br />

needs of the students and improved their performance.<br />

Furthermore, Study 2 helped reveal the complexities of<br />

work<strong>in</strong>g <strong>in</strong> heterogeneously grouped classrooms. With the<br />

strengthened Least Restrictive Environment (LRE) provision<br />

<strong>in</strong> 1997 mandat<strong>in</strong>g that students with disabilities have<br />

access to, and make progress <strong>in</strong>, the general education curriculum<br />

(U.S. Department of Education, 1997), more students<br />

with disabilities may be part of general education<br />

classrooms. Consequently, teachers will be <strong>in</strong>creas<strong>in</strong>gly<br />

challenged to create <strong>in</strong>struction that addresses the needs of<br />

students <strong>in</strong> academically heterogeneous classrooms. The<br />

method of do<strong>in</strong>g that well is still unclear to us. However,<br />

we conjecture that some aspects of SBI will likely benefit<br />

high-, average-, and low-perform<strong>in</strong>g students differentially,<br />

and others may improve their performance as a group.<br />

Low-perform<strong>in</strong>g students <strong>in</strong> both studies most likely benefited<br />

from direct, explicit, and systematic <strong>in</strong>struction.<br />

High-perform<strong>in</strong>g students may have enjoyed the pairedlearn<strong>in</strong>g<br />

opportunities provided dur<strong>in</strong>g SBI. Research that<br />

spans more than a decade suggests that the learn<strong>in</strong>g of<br />

high-achiev<strong>in</strong>g students, such as those <strong>in</strong> Study 2, is facilitated<br />

through peer <strong>in</strong>teraction <strong>in</strong> learn<strong>in</strong>g (e.g., Sw<strong>in</strong>g &<br />

Peterson, 1982; Webb, 1991; Webb, Troper, & Fall, 1995).<br />

The use of schema-based diagrams may have contributed to<br />

the word problem-solv<strong>in</strong>g and computation improvement<br />

among many of the participat<strong>in</strong>g students, regardless of<br />

ability level. Us<strong>in</strong>g what we learned from the design studies<br />

will allow us to strengthen the <strong>in</strong>struction before it is<br />

exam<strong>in</strong>ed <strong>in</strong> future research studies. In the <strong>in</strong>terim, prelim<strong>in</strong>ary<br />

f<strong>in</strong>d<strong>in</strong>gs reveal the potential of SBI <strong>in</strong> mathematics<br />

word problem solv<strong>in</strong>g with third-grade students.<br />

NOTES<br />

This research was supported by Grant H324D010024 from the U.S.<br />

Department of Education, Office of Special Education Programs.<br />

The authors thank Dr. Russell Gersten for his <strong>in</strong>sightful feedback dur<strong>in</strong>g<br />

the study. A version of this article was presented at the 2004 annual<br />

meet<strong>in</strong>g of the Council for Exceptional Children <strong>in</strong> New Orleans,<br />

Louisiana.<br />

REFERENCES<br />

Baxter, J., Woodward, J., Voorhies, J., & Wong, J. (2002). We talk about it,<br />

but do they get it Learn<strong>in</strong>g Disabilities: Research & Practice, 17, 173–185.<br />

Briars, D. J., & Lark<strong>in</strong>, J. H. (1984). An <strong>in</strong>tegrated model of skill <strong>in</strong> solv<strong>in</strong>g<br />

elementary word problems. Cognition and Instruction, 1, 245–296.<br />

Brown, V., Cron<strong>in</strong>, M. E., & McEntire, E. (1994). TOMA-2 Test of <strong>Mathematical</strong><br />

Abilities. Aust<strong>in</strong>, TX: Pro-Ed.<br />

Carpenter, T. P., Hiebert, J., & Moser, J. M. (1983). The effect of <strong>in</strong>struction<br />

on children’s solutions of addition and subtraction word problems.<br />

Educational Studies <strong>in</strong> Mathematics, 14, 55–72.<br />

Carpenter, T. P., & Moser, J. M. (1984). The acquisition of addition and<br />

subtraction concepts <strong>in</strong> grades one through three. Journal for Research <strong>in</strong><br />

Mathematics Education, 15, 179–202.<br />

Case, L. P., Harris, K. R., & Graham, S. (1992). Improv<strong>in</strong>g the mathematical<br />

problem-solv<strong>in</strong>g skills of students with learn<strong>in</strong>g disabilities:<br />

Self-regulated strategy development. The Journal of Special Education,<br />

26, 1–19.<br />

Chen, Z. (1999). Schema <strong>in</strong>duction <strong>in</strong> children’s analogical problem solv<strong>in</strong>g.<br />

Journal of Educational Psychology, 91, 703–715.<br />

Cho, S., & Ahn, D. (2003). Strategy acquisition and ma<strong>in</strong>tenance of gifted<br />

and nongifted young children. Exceptional Children, 69, 497–505.<br />

Cobb, P., Confrey, J., diSessa, A., Lehrer, R., & Schauble, L. (2003).<br />

Design experiments <strong>in</strong> educational research. Educational Researcher,<br />

32(1), 9–13.<br />

CTB/McGraw-Hill. (2001). TerraNova. Monterey, CA: Author.<br />

Cumm<strong>in</strong>s, D. D., K<strong>in</strong>tsch, W., Reusser, K., & Weimer, R. (1988). The role<br />

of understand<strong>in</strong>g <strong>in</strong> solv<strong>in</strong>g word problems. Cognitive Pyschology, 20,<br />

405–438.<br />

De Corte, E., Verschaffel, L., & De W<strong>in</strong>, L. (1985). Influence of reword<strong>in</strong>g<br />

verbal problems on children’s problem representations and solutions.<br />

Journal of Educational Psychology, 77, 460–470.<br />

The Design-Based Research Collective. (2003). Design-based research:<br />

An emerg<strong>in</strong>g paradigm for educational <strong>in</strong>quiry. Educational Researcher,<br />

32, 5–8.<br />

Desoete, A., Roeyers, H., & De Clercq, A. (2003). Can offl<strong>in</strong>e metacognition<br />

enhance mathematical problem solv<strong>in</strong>g Journal of Educational<br />

Psychology, 95(1), 188–200.<br />

Dunst, C. J., & Trivette, C. M. (1994). Methodological considerations<br />

and strategies for study<strong>in</strong>g the long-term effects of early <strong>in</strong>tervention. In<br />

S. L. Friedman & H. Carl Haywood (Eds.), Developmental follow-up:<br />

Concepts, doma<strong>in</strong>s, and methods (pp. 277–313). New York: Academic<br />

Press.<br />

Fennell, F., Ferr<strong>in</strong>i-Mundy, J., G<strong>in</strong>sburg, H. O., Greenes, C., Murphy, S. J.,<br />

& Tate, W. (1998). Mathematics: The path to math success! <strong>Grade</strong> 3.<br />

Needham, MA: Silver Burdett G<strong>in</strong>n.<br />

Fuchs, L. S., Fuchs, D., F<strong>in</strong>elli, R., Courey, S. J., & Hamlett, C. L. (2004).<br />

Expand<strong>in</strong>g schema-based transfer <strong>in</strong>struction to help third graders solve<br />

real-life mathematical problems. American Educational Research Journal,<br />

41, 419–445.<br />

Fuchs, L. S., Fuchs, D., Hamlett, C. L., & All<strong>in</strong>der, R. (1989). The reliability<br />

and validity of skills analysis with<strong>in</strong> curriculum-based measurement.<br />

Diagnostique, 14, 203–221.<br />

Fuchs, L. S., Fuchs, D., Prentice, K., Burch, M., Hamlett, C. L., & Owen,<br />

R., et al. (2003). Enhanc<strong>in</strong>g third-grade students’ mathematical problem<br />

solv<strong>in</strong>g with self-regulated learn<strong>in</strong>g strategies. Journal of Educational<br />

Psychology, 95, 306–315.<br />

Fuchs, L. S., Hamlett, C. L., & Fuchs, D. (1998). Monitor<strong>in</strong>g basic skills<br />

progress: Basic math computational manual (2nd ed.). Aust<strong>in</strong>, TX: Pro-Ed.<br />

Gersten, R. (<strong>in</strong> press). Beh<strong>in</strong>d the scenes of an <strong>in</strong>tervention research<br />

study. Learn<strong>in</strong>g Disabilities Research and Practice.<br />

Gersten, R., Baker, S., & Lloyd, J. W. (2000). Design<strong>in</strong>g high-quality<br />

research <strong>in</strong> special education: Group experimental designs. Journal of<br />

Special Education, 34, 2–19.<br />

Gick, M. L., & Holyoak, K. J. (1983). Schema <strong>in</strong>duction and analogical


283-302 Jitendra M_J 07 5/8/07 10:57 AM Page 302<br />

302 The Journal of Educational Research<br />

transfer. Cognitive Psychology, 15, 1–38.<br />

Glass, G. V., McGraw, B., & Smith, M. L. (1981). Meta-analysis <strong>in</strong> social<br />

research. Beverly Hills, CA: Sage.<br />

Goldsmith, L. T., & Mark, J. (1999). What is a standards-based mathematics<br />

curriculum Educational Leadership, 57, 40–44.<br />

Hegarty, M., Mayer, R. E., & Monk, C. A. (1995). Comprehension of<br />

arithmetic word problems: A comparison of successful and unsuccessful<br />

problem solvers. Journal of Educational Psychology, 87, 18–32.<br />

Hiebert, J., Carpenter, T. P., Fennema, E., Fuson, K., Human, P., Murray,<br />

H., et al. (1996). <strong>Problem</strong> solv<strong>in</strong>g as a basis for reform <strong>in</strong> curriculum and<br />

<strong>in</strong>struction: The case of mathematics. Educational Researcher, 25(4),<br />

12–21.<br />

Houghton Miffl<strong>in</strong>/Riverside Publish<strong>in</strong>g. (1999). Iowa Test of Basic Skills<br />

(ITBS). Itasca, IL: Author.<br />

Jitendra, A. K., Griff<strong>in</strong>, C. C., McGoey, K., Gardill, M. C., Bhat, P., &<br />

Riley, T. (1998). Effects of mathematical word problem solv<strong>in</strong>g by students<br />

at risk or with mild disabilities. The Journal of Educational<br />

Research, 91, 345–355.<br />

Jitendra, A. K., Hoff, K., & Beck, M. M. (1999). Teach<strong>in</strong>g middle school<br />

students with learn<strong>in</strong>g disabilities to solve word problems us<strong>in</strong>g a<br />

schema-based approach. Remedial and Special Education, 20, 50–64.<br />

Jonassen, D. H. (2003). Design<strong>in</strong>g research-based <strong>in</strong>struction for story<br />

problems. Educational Psychology Review, 15, 267–296.<br />

Kameenui, E. J., & Griff<strong>in</strong>, C. C. (1989). The national crisis <strong>in</strong> verbal<br />

problem solv<strong>in</strong>g <strong>in</strong> mathematics: A proposal for exam<strong>in</strong><strong>in</strong>g the role of<br />

basal mathematics programs. Elementary School Journal, 89, 575–593.<br />

K<strong>in</strong>tsch, W., & Greeno, J. G. (1985). Understand<strong>in</strong>g and solv<strong>in</strong>g word<br />

arithmetic problems. Psychological Review, 92, 109–129.<br />

Lester, F. K., Garofalo, J., & Kroll, D. L. (1989). Self-confidence, <strong>in</strong>terest,<br />

beliefs, and metacognition: Key <strong>in</strong>fluences on problem-solv<strong>in</strong>g behavior.<br />

In D. B. McLeod & V. M. Adams (Eds.), Affect and mathematical<br />

problem solv<strong>in</strong>g (pp. 75–88). New York: Spr<strong>in</strong>ger-Verlag.<br />

Lucangeli, D., Tressoldi, P. E., & Cendron, M. (1998). Cognitive and<br />

metacognitive abilities <strong>in</strong>volved <strong>in</strong> the solution of mathematical word<br />

problems: Validation of a comprehensive model. Contemporary Educational<br />

Psychology, 23, 257–275.<br />

Ma, L. (1999). Know<strong>in</strong>g and teach<strong>in</strong>g elementary mathematics: Teacher’s<br />

understand<strong>in</strong>g of fundamental mathematics <strong>in</strong> Ch<strong>in</strong>a and United States.<br />

Mahwah, NJ: Erlbaum.<br />

Manfre, E., Moser, J. M., Lobato, J. E., & Morrow, L. (1994). Heath mathematics<br />

connections. Wash<strong>in</strong>gton, DC: Heath.<br />

Marshall, S. P. (1995). Schemas <strong>in</strong> problem solv<strong>in</strong>g. New York: Cambridge<br />

University Press.<br />

Mayer, R. E. (1999). The promise of educational psychology: Vol. I. Learn<strong>in</strong>g<br />

<strong>in</strong> the content areas. Upper Saddle River, NJ: Merrill Prentice Hall.<br />

Mayer, R. E., & Hegarty, M. (1996). The process of understand<strong>in</strong>g mathematics<br />

problems. In R. J. Sternberg & T. Ben-Zeev (Eds.), The nature<br />

of mathematical th<strong>in</strong>k<strong>in</strong>g (pp. 29–53). Mahwah, NJ: Erlbaum.<br />

Mayer, R. E., Lewis, A. B., & Hegarty, M. (1992). <strong>Mathematical</strong> misunderstand<strong>in</strong>gs:<br />

Qualitative reason<strong>in</strong>g about quantitative problems. In J. I.<br />

D. Campbell (Ed.), The nature and orig<strong>in</strong>s of mathematical skills (pp.<br />

137–154). Amsterdam: Elsevier.<br />

Nathan, M. J., Long, S. D., & Alibali, M. W. (2002). The symbolic precedence<br />

view of mathematical development: A corpus analysis of the<br />

rhetorical structure of textbooks. Discourse Processes, 33, 1–21.<br />

National Council of Teachers of Mathematics. (2000). Pr<strong>in</strong>ciples and standards<br />

for school mathematics. Reston, VA: Author.<br />

National Research Council. (2001). Add<strong>in</strong>g it up: Help<strong>in</strong>g children learn<br />

mathematics. J. Kilpatrick, J. Swafford, & B. F<strong>in</strong>dell (Eds.), Mathematics<br />

Learn<strong>in</strong>g Study Committee, Center for Education, Division of<br />

Behavioral and Social Sciences and Education. Wash<strong>in</strong>gton, DC:<br />

National Academy Press.<br />

Prawat, R. S. (1989). Promot<strong>in</strong>g access to knowledge, strategy, and disposition<br />

<strong>in</strong> students: A research synthesis. Review of Educational Research,<br />

59, 1–41.<br />

Pressley, M., El-D<strong>in</strong>ary, P., Gask<strong>in</strong>s, I., Schuder, T., Bergman, J., & Almasi,<br />

J., et al. (1992). Beyond direct explanation: Transactional <strong>in</strong>struction of<br />

read<strong>in</strong>g comprehension strategies. Elementary School Journal, 92,<br />

513–556.<br />

Riley, M. S., Greeno, J. G., & Heller, J. I. (1983). Development of children’s<br />

problem–solv<strong>in</strong>g ability <strong>in</strong> arithmetic. In H. P. G<strong>in</strong>sburg (Ed.),<br />

The development of mathematical th<strong>in</strong>k<strong>in</strong>g (pp. 153–196). New York: Academic<br />

Press.<br />

Rogers, K. (2002). Re-form<strong>in</strong>g gifted education: Match<strong>in</strong>g the program to the<br />

child. Scottsdale, AZ: Great Potential Press.<br />

Schunk, D. H., (1998). Teach<strong>in</strong>g elementary students to self-regulate<br />

practice of mathematical skills with model<strong>in</strong>g. In D. H. Schunk & B. J.<br />

Zimmerman (Eds.), Self-regulated learn<strong>in</strong>g: From teach<strong>in</strong>g to self-reflective<br />

practice (pp. 137–159). New York: Guilford.<br />

Schurter, W. A. (2002). Comprehension monitor<strong>in</strong>g: An aid to mathematical<br />

problem solv<strong>in</strong>g. Journal of Developmental Education, 26(2), 22–33.<br />

Sw<strong>in</strong>g, S. R., & Peterson, P. L. (1982). The relationship of student ability<br />

and small group <strong>in</strong>teraction to student achievement. American Educational<br />

Research Journal, 19, 259–274.<br />

Swanson, H. L. (1999). Instructional components that predict treatment<br />

outcomes for students with learn<strong>in</strong>g disabilities: Support for a comb<strong>in</strong>ed<br />

strategy and direct <strong>in</strong>struction model. Learn<strong>in</strong>g Disabilities Research and<br />

Practice, 14, 129–140.<br />

Toml<strong>in</strong>son, C. (1995). How to differentiate <strong>in</strong>struction <strong>in</strong> mixed-ability classrooms.<br />

Alexandria, VA: Association for Supervision and Curriculum<br />

Development.<br />

U.S. Department of Education. (1997). Amendments to the IDEA, P.L.<br />

105-17. Retrieved February 27, 2007, from http://www.ed.gov/offices/<br />

OSERS/Policy/IDEA/the_law.html<br />

Van de Walle, J. A. (2004). Elementary and middle school mathematics:<br />

Teach<strong>in</strong>g developmentally (5th ed.). Boston: Allyn & Bacon.<br />

Verschaffel, L., De Corte, E., Lassure, S., Van Vaerenbergh, G., Bogaerts,<br />

H., & Rat<strong>in</strong>ckx, E. (1999). Learn<strong>in</strong>g to solve mathematical application<br />

problems: A design experiment with fifth graders. <strong>Mathematical</strong> Learn<strong>in</strong>g<br />

and Th<strong>in</strong>k<strong>in</strong>g, 1(3).<br />

Webb, N. M. (1991). Task-related verbal <strong>in</strong>teraction and mathematics<br />

learn<strong>in</strong>g <strong>in</strong> small groups. Journal for Research <strong>in</strong> Mathematics Education,<br />

22, 366–389.<br />

Webb, N. M., Troper, J., & Fall, J. R. (1995). Constructive activity and<br />

learn<strong>in</strong>g <strong>in</strong> cooperative small groups. Journal of Educational Psychology,<br />

87, 406–423.<br />

Wood, D. K., Frank, A. R., & Wacker, D. P. (1998). Teach<strong>in</strong>g multiplication<br />

facts to students with learn<strong>in</strong>g disabilities. Journal of Applied Behavior<br />

Analysis, 31, 323–338.<br />

Zawaiza, T. B. W., & Gerber, M. M. (1993). Effects of explicit <strong>in</strong>struction<br />

on community college students with learn<strong>in</strong>g disabilities. Learn<strong>in</strong>g Disabilities<br />

Quarterly, 16, 64–79.


324-328 Bracey M_J 07 5/8/07 11:01 AM Page 328<br />

Among Those Present<br />

STEVEN B. SHELDON is an associate research scientist,<br />

Center on School, Family, and Community<br />

Partnerships, Johns Hopk<strong>in</strong>s University. He studies the<br />

development of family and community <strong>in</strong>volvement<br />

programs <strong>in</strong> schools, along with the impact of these<br />

programs on student outcomes. He also conducts<br />

research <strong>in</strong>to the <strong>in</strong>fluences of parent <strong>in</strong>volvement,<br />

<strong>in</strong>clud<strong>in</strong>g parent beliefs and social relations, and school<br />

outreach.<br />

KERRY HOLMES is an assistant professor of elementary<br />

education, School of Education, the University<br />

of Mississippi. Her research <strong>in</strong>terests focus on<br />

clos<strong>in</strong>g the read<strong>in</strong>g achievement gap through sound<br />

early childhood read<strong>in</strong>g practices and the motivation<br />

of reluctant readers. SARAH POWELL is an honor<br />

student <strong>in</strong> the School of Education at the same university.<br />

She graduated with honors from the Sally<br />

McDonnell Barksdale Honors College. STACY<br />

HOLMES is an associate professor of eng<strong>in</strong>eer<strong>in</strong>g at<br />

the University of Mississippi. He is <strong>in</strong>terested <strong>in</strong><br />

br<strong>in</strong>g<strong>in</strong>g quantitative techniques to the study of read<strong>in</strong>g<br />

acquisition. EMILY WITT is a fourth-grade<br />

teacher at Oxford Elementary School. She graduated<br />

with a master’s degree <strong>in</strong> elementary education from<br />

the University of Mississippi <strong>in</strong> 2005.<br />

ASHA K. JITENDRA is professor, College of Education,<br />

Lehigh University. Her areas of research <strong>in</strong>terests<br />

<strong>in</strong>clude mathematics <strong>in</strong>struction and assessment.<br />

Other <strong>in</strong>terests <strong>in</strong>clude read<strong>in</strong>g <strong>in</strong>terventions, textbook<br />

analysis, and curriculum-based assessment.<br />

CYNTHIA C. GRIFFIN is an associate professor,<br />

Department of Special Education, University of Florida.<br />

She is <strong>in</strong>terested <strong>in</strong> beg<strong>in</strong>n<strong>in</strong>g special educators<br />

and <strong>in</strong>clusive practice <strong>in</strong> mathematics and read<strong>in</strong>g.<br />

ANDRIA DEATLINE-BUCHMAN is a special<br />

education consultant, Easton Area School District,<br />

Pennsylvania. She provides academic support to elementary,<br />

middle, and high school special education<br />

students. Her areas of research <strong>in</strong>terests <strong>in</strong>clude writ<strong>in</strong>g<br />

<strong>in</strong>struction and curriculum-based assessment.<br />

EDWARD SCZESNIAK is also a special education<br />

consultant with the Easton Area School district. He<br />

provides behavioral support to elementary, middle,<br />

and high school special education students by conduct<strong>in</strong>g<br />

functional behavioral assessments and develop<strong>in</strong>g<br />

behavioral <strong>in</strong>tervention plans. He is <strong>in</strong>terested<br />

<strong>in</strong> mathematics <strong>in</strong>struction and assessment.<br />

HERBERT WARE is associated with George Mason<br />

University, Fairfax, Virg<strong>in</strong>ia. As a former public school<br />

teacher, research coord<strong>in</strong>ator, and adm<strong>in</strong>istrator, his<br />

<strong>in</strong>terests lie with the application of quantitative methods<br />

to the study of school climate. He is particularly<br />

<strong>in</strong>terested <strong>in</strong> the use of national data sets as a way of<br />

study<strong>in</strong>g the relationship of school climate to student<br />

achievement. ANASTASIA KITSANTAS is an associate<br />

professor, College of Education and Human<br />

Development, also at George Mason University. Her<br />

research <strong>in</strong>terests focus on social cognitive processes,<br />

self-regulated learn<strong>in</strong>g, and motivational beliefs.<br />

KELLIE J. ANDERSON is a school psychologist, Student<br />

Services Center, Anne Arundel County Public<br />

Schools, Maryland. Her areas of <strong>in</strong>terest <strong>in</strong>clude family<br />

and school collaboration, crisis <strong>in</strong>tervention, and adolescent<br />

development. KATHLEEN M. MINKE is a<br />

professor, School Psychology Program, University of<br />

Delaware. Her research <strong>in</strong>terests <strong>in</strong>clude family–school<br />

collaboration, self-concept, and professional issues <strong>in</strong><br />

school psychology.


Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!