Alternate Test Forms Research Articles

INFERRING EXAMINEE ABILITY WHEN SOME ITEM RESPONSES ARE MISSING

ABSTRACTThe basic equations of item response theory (IRT) provide a foundation for inferring examinees' abilities and items' operating characteristics from observed responses. In practice, though, examinees will usually not have provided a response to every available item–for reasons that may or may not have been intended by the test administrator, and that may or may not be related to examinee ability. The mechanisms that produce missingness must be taken into account if correct inferences are to be drawn. Using concepts introduced by Rubin (1976), we discuss the implications for ability and item parameter estimation that are entailed by alternate test forms, targeted testing, adaptive testing, time limits, and omitted responses.

Response latencies and the computerized assessment of intelligence

A study was undertaken to examine the relationship between response latencies to verbal ability test items administered by computer and overall verbal intelligence test scores. Sixty-four undergraduate students responded to a test of verbal ability under four conditions of alternate test forms (A or B) and modes of administration (computerized vs paper-and-pencil). The response latencies recorded during computerized testing, averaged for each subject, showed a negative correlation with overall test scores as would be predicted from a speed-of-information-processing perspective of human intelligence (Jensen, 1982a, b; Vernon, 1983). This inverse relationship was evident in every condition of test form and mode of administration, thereby demonstrating the generalizability of these findings. Discussion considered the implications of test speededness for the results of this study and provided suggestions for future research employing response latency data as a means for studying the cognitive processes underlying intelligent behaviour.

Assessing the Test-Wiseness of Health Science Students

Two alternate test forms were developed to assess several test-wiseness (TW)principles for use with students in the health sciences. Each of the 24-item tests has six subtests consisting of four multiple-choice items that assess the examinee's skill in using the TWprinciples of similar options, umbrella term, item give-away, convergence principle, length of correct option, and stemoption cue. Analysis of responses from 39 entering medical students showed reasonable internal consistency (KR-20 = .79 for the combinedforms) and alternate form (r = .68) reliabilities. The means (64.1% and 65.7%, standard deviations (13.9% and 14.1%, and KR-20 reliabilities (.65 and .65) were similar for the two forms of the test. Some degree of validity for measuring TW is indicated by the significant gain made on the test by nursing and medical students specifically instructed in test-taking skills and by the correlations with course grades. The test is appropriatefor use with health science students because of the medical nature of the content of the test items and the level of performance on the test by both nursing and medical students, although some question remains concerning the appropriateness of retaining the current items that measure the use of the convergence principle.

The standardized mental stress test protocol: test-retest reliability and comparison with ambulatory blood pressure monitoring.

ABSTRACTThe mental stress test protocol is used extensively in research, but different laboratories often employ different stress tasks, utilize different dependent variables to index the stress response, and perform different transformations on the gathered data. The present study determined the test‐retest reliability of 11 cardiovascular dependent variables during a resting baseline and three common stress tasks: playing a video game, performing a choice reaction‐time test, and performing a cold‐pressor test. Sixty healthy, middle‐aged males underwent testing twice, approximately three months apart. Instructions were delivered via videotape and data were gathered on‐line by computer to ensure a standard laboratory environment. Each task elicited significant increases in blood pressure, vascular rigidity, LVET, heart rate, and stroke volume. In addition, the cold‐pressor test led to increases in total systemic resistance and mean systolic ejection rate. The absolute levels of the 11 dependent variables were correlated across tasks (partialr, baseline removed, = .06 to .69, 32 of 33 comparisons significant atp<.05), indicating that reactivity to stress generalizes across alternate test forms. The absolute levels also showed significant test‐retest reliability (r= .32 to .82; 40 of 44 comparisons significant atp<.05). In addition, for 19 of 33 comparisons, absolute levels showed greater test‐retest reliability than change scores derived by subtracting the initial resting baseline value from the stress‐task value. Finally, blood pressures taken during the stress tests were more highly correlated with the average blood pressures measured via ambulatory monitoring than casual office pressures, suggesting that such stress values may more accurately reflect average blood pressure.

Beneficial and detrimental modeling effects on creative response production.

Enhancement of creativity through various educational "training" methods or external factors fostering creative production has long been demonstrated. The assumption has been that creativity exists as a given quantity in trait form, or that progressive learning takes place. Social observational modeling is investigated here as a technique that can both increase and decrease level of creative production in the individual despite every opportunity presented to respond freely and without (a) resort to any direct education or reinforcement methods or (b) the presence of environmental constraints. One hundred and twenty female university students were assessed for creativity on one figural and one verbal task of the Torrance Tests of Creative Thinking and the components of fluence, flexibility, originality, and elaboration. Ss then observed either high, low, or no figural creative model performance based on normative responses of past Ss and were reassessed on alternate test forms. Reassessment showed both predicted increases and decreases in component verbal creativity scores.

Amount and loci of classroom answer copying, spaced seating, and alternate test forms.

Amount and loci of classroom answer copying, spaced seating, and alternate test forms.

Amount and loci of classroom answer copying, spaced seating, and alternate test forms.

Amount and loci of classroom answer copying, spaced seating, and alternate test forms.

A STUDY OF EMOTIONAL STATES AROUSED DURING EXAMINATIONS*

ABSTRACTAre there some groups of candidates who cannot perform at their best levels when working under the conditions of pressure that prevail during a College Board examination? Does excessive anxiety of the candidates reduce the validity of the test? An attempt has been made to answer these questions by administering a “relaxed” test to some candidates a few days before and to others a few days after the January 1960 SAT. These tests were followed by a questionnaire asking about the students' background, academic ambitions, feeling of anxiety, and reasons for feeling anxious. Two scales measuring generalized anxiety were included. Teachers' grades in four different high‐school subjects were collected.The results show the effects of anxiety to be small. A comparison of scores on the relaxed tests with alternate test forms taken with the SAT indicates no effect of anxiety for boys, while anxiety seems to be associated with a slight but significant improvement of girls' mathematical scores relative to their verbal scores. Since the relaxed test, the “anxious” test taken with the SAT, and the SAT itself were all found to have substantially the same concurrent validities, it is also unlikely that anxiety has had any effect on predictive validity. Responses to the questionnaire brought out expected statements about anxious feelings, and a small, possibly significant, tendency for “academically minded” boys to do relatively poorly on an anxious test as compared to a relaxed test, while academically minded girls show the opposite tendency. These effects are all well below the standard errors of measurement of the test.

Thurstone Scale for Measuring Attitudes of College Students toward Physical Fitness and Exercise

Abstract Employing modified Thurstone techniques, two equivalent forms of an equal-appearing intervals attitude scale were constructed around the topic “physical fitness and exercise.” Experimental test items were subjected to an extensive refinement process preceding the formulation of the final alternate test forms. Validity of the test was based upon authoritative opinion and expert judgment. A reliability coefficient derived from test, retest sampling was .83 ± .06. A correlation coefficient of .87 ± .03 was obtained on the parallel lest forms.

A Health Knowledge Test for the Seventh Grade

Abstract A valid and reliable instrument was developed for the purpose of measuring the extent of health knowledge possessed by seventh grade pupils. Appropriate vocabulary burden, level of difficulty, and curricular validity were primary factors considered in item formulation. Experimental test forms were subjected to an extensive refinement process prior to the construction of two final, alternate test forms. The alternate forms were administered in a random sample of public schools in Indiana. Analysis of test results yielded statistical evidence of the equivalence of forms.

An appraisal of the validity of the factor loadings employed in the construction of the primary social attitude scales

In this article the authors examine the effect of including alternate test forms in a factor matrix upon the validity of the resultant factor loadings, finding that in this particular instance the effect is negligible. Comparisons of the factor loadings derived from matrices in which only one of the alternate test forms is included with those in which both forms are included reveal practically no difference in the magnitude of either the original or rotated factor loadings, or in that of the computed communalities.