Samuel B. Green

Limitations of Coefficient Alpha as an Index of Test Unidimensionality1

Confusion in the literature between the concepts of internal consistency and homogeneity has led to a misuse of coefficient alpha as an index of item homogeneity. Coefficient alpha is actually a complexly determined test statistic, item homogeneity only being one influence on its magnitude. The related statistic, the average intercorrelation, has similar difficulties. Several indices of item homogeneity derived from the model of common factor analysis are offered as alternatives.

Factors Affecting Energy Consumption: Two Field Tests of the Fishbein-Ajzen Model

Summary One hundred and thirteen students and 506 consumers in a statewide sample participated in two studies which investigated the relationship among variables defined by the Fishbein-Ajzen model, situational factors, and energy consumption measures. Correlational analyses with the student sample indicated that one of the situational variables, knowledge, predicted energy usage relatively accurately. In addition, the standard Fishbein-Ajzen model appeared to be moderately useful in understanding energy consumption. With the statewide sample, similar results were found except that the situational variable of income rather than knowledge was the most potent predictor. A manipulation attempting to alter the information level of the students yielded a change in knowledge, but no change in the Fishbein-Ajzen measures. It was concluded that given the complex nature of energy consumption, multiple treatments are probably necessary to increase conservation.

Coffee drinking and cigarette smoking: I. coffee, caffeine and cigarette smoking behavior

Abstract Two experiments designed to assess the relationship between coffee intake and smoking are reported. In Experiment I, coffee drinking smokers were randomly assigned to four groups in which they received 0, 1, 2, or 3 cups of coffee during two one-hour sessions while they worked on crossword puzzles. Results showed that subjects receiving coffee in any amount smoked more than subjects who were not provided coffee. Moderate and low rate smokers were then randomly assigned to one of five groups in Experiment II, in which they were provided no drink, water, Postum® (a coffee substitute), caffeinated, or decaffeinated coffee. These groups were selected to assess the characteristics of coffee that may have influenced increased smoking. Results for number of cigarettes smoked and puff rate generally showed that subjects receiving caffeinated or decaffeinated coffee smoked more than subjects in the no drink or water control groups. The results of this study provide experimental evidence of the role of coffee in setting the occasion for smoking, as well as ruling out the presence of a liquid or caffeine as the important characteristics of coffee in influencing smoking.

Identifiability of spurious factors using linear factor analysis with binary items

The purpose of this study was to evaluate the ro bustness of some linear factor analytic techniques to violations of the linearity assumption by factoring product-moment correlations computed from data con forming to an extended, three-parameter logistic model of item responding. Three factors were crossed to yield 81 subcases: the number of underlying dimen sions (0, 1, or 2), the number of items (10, 15, 20, 25, 30, 35, 40, 45, or 50), and the number of subjects (100, 250, or 500). The mean eigenvalues for the sub cases were evaluated using parallel analysis and the scree technique. The mean eigenvectors were visually inspected. For almost all subcases with one or two un derlying dimensions, a single spurious factor was able to be identified using parallel analysis. However, in comparison with the nonspurious factors, it was small in magnitude and, in practice, factors of this relative size might be interpreted as trivial. It was concluded that researchers may have some confidence in inter preting linear factor analysis with binary items if they are using a test instrument that has been carefully de veloped.

Accuracy of observational data with whole interval, partial interval, and momentary time-sampling recording techniques

The purpose of the study was to compare the accuracies obtained from whole interval, partial interval, and momentary time-sample recording procedures. Two types of accuracies were defined. Between-methods accuracy is the degree to which an observer using a particular recording method agrees with a standard using a continuous recording method. Within-methods accuracy is the degree to which an observer and a standard agree when both use the same recording method. Fifty-four undergraduate students viewed a videotape of a woman twisting her hair and recorded the occurrence or nonoccurrence of the behavior using one of the three recording methods. Tapes were divided into segments with low, intermediate, and high rates of behavior. Partial and whole interval recording obtained poorer between-methods accuracies than momentary time sampling. Results showed that whole interval and momentary time-sampling procedures yielded higher within-method accuracies than partial interval recording. Thus, momentary time-sampling recording provided the advantages of both greater representativeness and as few or fewer observer errors than whole or partial interval recording procedures. Use of the interval recording methods by researchers must be justifiable within this framework.

An Evaluation of the Response Format and Scale Structure of the Job Diagnostic Survey

This research deals with an investigation of the factor structure of the Job Diagnostic Survey. Textile operatives (n = 114) in a southeastern textile plant completed the JDS. Factor analyses produced substantially different results than have appeared in the literature. Further correlational analyses revealed that the format used in collecting the data may be partially responsible for the different factor structures which have been found for the various types of workers. The researchers conclude that a simplification of the format may be needed.

Measurement of assertive behavior: Construct and predictive validity of self-report, role-playing, and in-vivo measures

Examined the predictive validity and construct equivalence of the three major procedures used to measure assertive behavior: Self-report, behavioral role-playing, and in-vivo assessment. Seventy-five Ss, who spanned the range of assertiveness, completed two self-report measures of assertiveness, the Rathus Assertiveness Scale (RAS) and the College Self-Expression Scale (CSES); two scales from the Endler S-R Inventory of General Trait Anxiousness, the interpersonal and general anxiety scales; eight role-playing situations that involved the expression of positive and negative assertiveness; and a telephone in-vivo task. In general, the study revealed the following: (1) assertiveness measures are task-dependent in that there was more overlap within task than between tasks; (2) there is a moderate degree of correspondence between self-report and role-playing measures, although this was true only for negative assertion; (3) positive and negative assertion do not appear to have the same topography of responding; and (4) there appears to be no consistent relationship between the in-vivo measure and any other type of assertiveness measure.

Coffee drinking and cigarette smoking: II. Coffee, urinary pH and cigarette smoking behavior.

Abstract This study was designed to evaluate the relationship between coffee consumption, urinary pH, and cigarette smoking. Urine acidity levels were manipulated and the resulting effect on cigarette smoking was monitored. Each of eight subjects participated in each of four conditions in which they received: water, coffee, coffee plus sodium bicarbonate, or coffee plus ascorbic acid. Subjects provided both pre- and post-session urine specimens, which were analyzed to determine pH, and cigarette butts were collected. The results showed that subjects smoke more cigarettes in a one hour session when they receive coffee. However, coffee did not have the effect of increasing urine acidity in one hour so that increased urine acidity cannot account for observed smoking increases. The results are discussed in terms of a stimulus control explanation for the relationship between coffee consumption and cigarette smoking.

Another look at gender-related differences in leadership behavior

This investigation examined self-related behavioral differences in terms of broad leadership styles using a methodology that minimized biases inherent in research designs previously employed to investigate this area. Subjects consisted of 40 males and 41 females who served in the role of leader, and who instructed both a male and a female subordinate in a card-sorting task. Experimental sessions were tape-recorded and transcribed, after which judges classified word phrases into various language categories. Factor analysis of the language categories yielded two factors of verbal behavior within a leadership situation: a socially oriented style and a task-oriented style. In contrast to generally accepted stereotypes, results indicated that within the present experimental context male leaders preferred to use phrases that indicated a more social style than female leaders. However, both male and female leaders used language that indicated a task-oriented approach to a significantly greater degree when instructing a female rather than a male subordinate. Implications and interpretations of the results are discussed.

Use of the Coombs Elimination Procedure in Classroom Tests.

AbstractThe effectiveness of the Coombs elimination procedure was evaluated with 29 college students enrolled in a statistics course. Five multiple-choice tests were employed and scored using the Coombs procedure. Specifically, students were instructed to mark incorrect rather than correct options and were informed about the scoring method, which discourages guessing. Four indices of performance were calculated to obtain specific indicants of guessing and partial knowledge. Simulated scores were calculated for each test as if the questions were administered using standard procedures. The results suggest that the Coombs procedure decreased guessing, and this effect increased over the grading period. Decreased guessing may account for modest gains in reliability in comparison to reliability of a standard test; however, the decrease from the first to final exam did not improve reliability.