25.11.2014 Views

Developmental psychology.pdf

Developmental psychology.pdf

Developmental psychology.pdf

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

Tests and Measurement 363<br />

Evaluation of the Approach Criterion-referenced testing was developed primarily<br />

for individualized and self-paced instruction, in which each student is expected to<br />

achieve mastery in his or her own way and at his or her own rate. Mastery in these<br />

instances is indicated as a percentage of the material learned. But there are difficulties<br />

here too. What is adequate mastery? What, for example, should be the standard for<br />

airline pilots? What is the lowest acceptable score on the simulator test? Setting a<br />

criterion involves an arbitrary decision, just as it does in norm-referenced testing.<br />

The concept behind criterion-referenced testing has a long history in early<br />

education, especially for promotion from one grade to the next. Each student is judged<br />

on his or her ability to perform at the next level, regardless of how many others pass<br />

as well. Licensing and certification in most professions function in this way (Shimberg,<br />

1981). But particularly for promoting growth, testing programs should be flexible,<br />

adapted to individual students with specific assets and limitations, rather than set at<br />

the same absolute standard for everyone (Millman, 1980).<br />

Purposes and Practices<br />

These two types of tests have different purposes. Norm-referenced tests identify differences<br />

among individuals; they are used primarily for selection. Criterion-referenced<br />

tests reveal changes within an individual; they serve chiefly to measure progress. Expressed<br />

differently, we can say that norm-referenced tests indicate the individual's<br />

standing in comparison with others. Criterion-referenced tests indicate the percentage<br />

of the material that an individual has learned. The examiner must keep these differences<br />

in mind (Figure 13.17).<br />

Characteristic Norm-referenced Criterion-referenced<br />

The chief concern is<br />

with. . .<br />

The individual's test score is<br />

expressed as a. . .<br />

selection, based on<br />

differences among<br />

individuals. How does the<br />

performance of pilot X<br />

compare with that of<br />

others taking the test?<br />

percentile. Pilot X ranks<br />

second among all 20<br />

students who took the<br />

test, placing at the 95th<br />

percentile.<br />

growth, based on differences<br />

within individuals. Has the<br />

performance of pilot X<br />

improved over the course<br />

of instruction?<br />

percentage. Having<br />

answered 16 out of 20<br />

questions correctly, the<br />

percentage score for pilot<br />

X is 80.<br />

Figure 13.17<br />

Norm-Referenced and Criterion-<br />

Referenced Tests.<br />

These approaches are not mutually exclusive, however. A norm-referenced<br />

test can be used for evaluating learning, especially by including a critical score that<br />

serves as a criterion for passing. A criterion-referenced test can be constructed from<br />

normative data, and it can be used for selection. But skillful examiners will find that<br />

most tests are more suitable for one purpose than for the other (Carver, 1974; Johnson<br />

& Martin, 1980).<br />

Tests and Users Within this context, two dimensions of any testing situation should<br />

be emphasized. First, one sometimes hears psychological tests described as "good" or<br />

"bad," perhaps dependent upon the speaker's general attitude toward <strong>psychology</strong> or<br />

experience in a particular testing situation. Instead, it is more appropriate to speak of<br />

the potential of a test. A well-constructed test has a higher potential for measuring<br />

intelligence, or whatever, than one developed by inexperienced people, who may not<br />

evaluate the test carefully. A test with high validity and an adequate set of norms has<br />

a high potential for success in its purported use (Figure 13.18).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!