Test-taking Behavior Research Articles

A Comparison of Computer-Monitored Group Tests with Paper-and-Pencil Tests

A test was administered to pharmacy students in two ways: using a computer-monitored, group-response system and using the conventional, paper-and-pencil method. Student performance on the two administrative modes was compared. Test-taking behavior for the two modes, as measured by the number and pattern of changed answers, was also compared. Three item formats (true-false, multiple-choice and matching) were analyzed separately.

Cross-Situational and Longitudinal Stability of Social-Emotional Functioning in Young Children

KOHN, MARTIN, and ROSMAN, BERNICE L. Cross-situational and Longitudinal Stability of Social-emotional Functioning in Young Children. CHILD DEVELOPMENT, 1973, 44, 721-727. Evidence is presented that 2 major factor dimensions that jointly account for major proportions of the variance of the social-emotional functioning of young children represent relatively enduring personality predispositions. Kindergarten children (N = 287) were rated by their teachers on classroom functioning and by testers on test-taking behavior; most of the children (N = 271) were rated again by their classroom teacher at the end of the first grade and by testers in the first half of the second grade. The factor dimensions were found to have moderate stability over time and across settings. Longitudinal stability was enhanced substantially when matching factor scores were pooled across settings.

Examination Performance as a Function of Group and Individual Administration

Bloom ( 1968), in considering student input to instruction, has suggested that group-study procedures effectively increase students' performance. As an extension of Bloom's suggestion, effects of methods of test-taking behavior on performance were compared. While the two heads are better than maxim has logical merit since more alternatives or information could be considered in a group, no data are available on this point. Consequently, it was hypothesized that students who take exams in small groups working for a common grade will perform better than when taking exams individually. During the quarter, 72 college students (25 males, 47 females) enrolled in a course in educational measurement took eight examinations of objective questions taken from the course instructor's manual (Noll, 1972). Four tests (75 questions) were administered in small groups of three students, each of whom completed the test for one score to be received by each member. Ss completed the other four examinations (also 75 questions) individually. The order of method of administration was randomly determined and rhe instrucrional conditions and format did not vary with administration. Following each examination, test papers were scored according to the key in the instructor's manual. Mean number of correct test responses in the small groups was 70.35 (SD of 4.45). The mean for the individual condition was 63.61 (SD of 14.03). A t test of 14.28 (9 < .001; df = 142) indicated a significant difference between the mean scores of the test-taking conditions, i.e., Ss working in small groups and allowed to reach consensus scored significantly higher on such exams than when administered tests individually. The srudy lacked control for such relevant variables as intelligence, locus of control, and persuasiveness of group members. Findings do suggest that the groups of three improved test scores in this course. Further investigation of participation by learners seems in order. REFERENCES

Deriving a personality characteristic from test-taking behavior.

Students were asked to indicate how sure they were of items on a test. Whether a student was under- or over-confident in the accuracy of his knowledge was determined by comparing the number of correct items with the degree of confidence expressed. A moderate consistency in this trait was found from one test to another.

Anxiety and test-taking behavior.

Anxiety and test-taking behavior.

The Application of a Self-Instructional Technique to Develop a Test-Taking Strategy

Test-wiseness is recognized as an important source of variation in educational and psychological measurement. A test-wise examinee can be expected to obtain a higher test score than an equally competent examinee who lacks test sophistication. Test constructors have attempted to reduce such variation by writing clear and comprehensive directions concerning timing and scoring. However, no matter how good the directions, cursory observation of any group of examinees will demonstrate that a large number do not behave as directed. They fail to answer every question when told to do so, guess wildly when told not to, etc. The cumulative effect of previous testing experiences is apparently a more important determinant of test-taking strategy than any single set of directions. Several publications (Morgan and Deese, 1957; Wechsler, Blum, and Friedman, 1963) have been written to assist examinees in taking tests. However, this information appears to have the same limitations as test directions. A number of investigations (Frankel, 1960; French and Dear, 1959; Peel, 1952; Vernon, 1954; Wiseman, 1954; Yates, 1953) have been concerned with the effects of special coaching or practice, or both, on test performance. However, these studies have focused on substantive test content rather than on test-taking strategy. This study was concerned with standardizing the test-taking behavior of examinees in an optimal fashion. The examinees were taught to identify and to respond differentially to the relevant cues in test directions.

The Effects of the Purported Validity of A Test Upon Level of Motivation and Performance

DESPITE THE concern in educational psychol ogy today over lack of student motivation and the problems of under-achievement, w e know little about the level of student motivation in the actual testing situation. A survey of recent texts i n edu cational measurement and evaluation (for example, 3, 5, and 8) and in educational psychology (2 and 9) indicates that the topic is usually ignored, and the general feeling seems to be that on the basis of ob servations of students' test-taking behavior, we may assume that they are well-motivated. It is of inter est, however, to consider what effect purported characteristics of the testing instrument have upon students' motivation and performance. In one of the more fruitful approaches to this problem, McClelland and his associates (7, Chapter V) have found that in regard to s i m p 1 e tasks the strength of achievement motivation in some subjects may be experimentally aroused and manipulated through the use of instructions which focus on the importance or unimportance of the task. Other stud ies (1 and 6) indicate that for some simple labora tory tasks the level of performance is related to the strength of the S's typical (i. e., not experimentally aroused) level of achievement motivation. The purpose of this study is to apply this ap proach to a more complex task ? the taking of an intelligence test ? and to study the effect of a spe cific test characteristic ? the validity of the test ? upon the motivation and performance of the stu dents. On the basis of the previous research using simple tasks, it is hypothesized that students who believe the intelligence test tobe a valid one will perform better on it than will students who believe it to be invalid.

Development of Statistical Methods Especially Useful in Test Construction and Evaluation

concerning the development of statistical models for the analysis of testtaking behavior has been added. As in the corresponding chapter of three years ago, several empirical studies are mentioned that furnish evidence regarding the effectiveness of various statistical procedures when they are applicable to the analysis and evaluation of item and test data. It should be pointed out that developments in factor analysis were reviewed in the