Clemens S. Bernhardson

Dogmatism, defense mechanisms, and social desirability responding.

The influence of social desirability responding on the correlation between the Dogmatism and Repression-Sensitization Scales was examined. A positive relationship was found between these two scales (r = .44, p < .001). This relationship decreased only slightly when a partial correlation was used to remove the influence of social desirability responding (r = .40, p < .001). In addition, correcting for attenuation in the controlled variable still resulted in a significant relationship between these two variables (r = .39, p < .001). Thus social desirability responding does not have an undue influence on the relationship between the Dogmatism Scale and the Repression-Sensitization Scale.

Comparison of the Three-Decision and Conventional Multiple-Choice Tests

Two multiple-choice tests, one with five alternatives for each question and one with four alternatives for each question, were scored as a Three-decision Multiple-choice Test and as a conventional multiple-choice test. In addition, the five-alternative test was scored as a modified conventional multiple-choice test by giving half marks if the correct alternative was picked as the second choice. The different scoring systems were evaluated by correlating the scores with the average mark obtained by each student in all his courses during the year. The results indicated that the conventional multiple-choice test was not improved by scoring methods which gave credit for partial knowledge.

DETERMINATION OF THE CHANCE SCORE ON THE THREE-DECISION MULTIPLE-CHOICE TEST

The chance score on the Three-decision Multiple-choice Test was determined theoretically. The theoretical predictions were then tested empirically by using a table of random numbers to answer 100 sets of 20 multiple-choice questions like a three-decision multiple-choice test would be answered. The results indicated that the average chance score on the Three-decision Multiple-choice Test is over two times as large as the average chance score on the traditional five-alternative multiple-choice test.

Relationship between individual social desirability ratings and Hand Acquiescence and I-E scores.

The relationship between two measures of individual social desirability ratings and the test scores of college students was determined for Hands Acquiescence Scale and the I-E Scale. A measure of the imbalance of intensity of social desirability on the two subscales of each test was significantly related to the scores on Hands Acquiescence Scale (r = .38, p < .001) and to the scores on the I-E Scale (r = .82, p < .001). A measure of the imbalance of the individual social desirability ratings on pairs of items was significantly related to Hands Acquiescence Scale scores (r = 32, p < .001) and to the scores on the I-E Scale (r = .79, P < .001).

COMMENT ON "LOGIC OF ANALYSIS OF VARIANCE IN RELATION TO PREDICTIONS AND EVALUATIONS OF INTERACTION MEANS"

Cutter and Heilizer (1971) suggested the use of adjusted interaction means for examining interactions in analysis of variance designs. However, the adjusted interaction means were estimates of the parameter, αβIJ, which represents the interaction between the Ith level of factor A and the Jth level of factor B. Therefore, estimated α β I J   ( α ∘ β i j ) would have been more straightforward terminology than adjusted interaction means.

Social Desirability, Endorsement and Judged Frequency of Occurrence on the Items from Four Social Desirability Scales

The usual positive relationship between the social desirability of traits and the proportion of the group endorsing the traits was found on items from 3 of 4 social desirability scales. However, judged frequency of occurrence of the traits was positively related to endorsement on the items from all 4 scales. Partialling out judged frequency of occurrence resulted in a reduced relationship between social desirability and endorsement when the group of items had a positive relationship between social desirability and judged frequency of occurrence. Computer simulation estimated that a considerable component of an individuals social desirability score could be due to how many of the traits which characterized the individual were included in each scale.

PERCEPTIONS OF SOCIAL DESIRABILITY AND FREQUENCY OF OCCURRENCE OF TRAITS AND THE PROBABILITY THAT THE TRAITS WILL BE ENDORSED

One group estimated the percentage of their fellow students characterized by each one of 100 traits while another group responded True or False to each trait. The correlation between the social desirability values of the traits and probability of endorsement was .80 while the correlation between the perceived frequency of occurrence of these traits and the probability of endorsement was .86. The multiple correlation of frequency of occurrence and social desirability with probability of endorsement was .88 and .10 of the variance in the endorsement variable was directly associated with variance in the social desirability variable while .37 of the variance in the endorsement variable was directly associated with variance in the frequency of occurrence variable.

Intra-maze cues and alternation after forced trials

Rats were given four forced trials and one free trial each day for 15 days. The use of an opaque or clear door to block the nonforced arm did not influence alternation. When one arm was black and the other was white, more alternation occurred when the same maze arms were used for forced and free trials than when new arms were used for the free trials.

INDEPENDENCE OF HAND'S ACQUIESCENCE AND SOCIAL DESIRABILITY SCALES

Hands Acquiescence Scale (HAcq) and Hands Social Desirability Scale (HSD) have been reported to be independent measures (Hand, 1964). However, there are 30 items common to both scales. An abbreviated HAcq (AHAcq) was constructed by removing from HAcq the items which occurred on both HAcq and HSD. The correlation between HAcq and HSD was −0.10 (p > .10). However, the correlation between AHAcq and HSD was −0.19 (p < .01). These results indicate HAcq and HSD are not independent when the effect of the overlapping items is removed.

Influence of Guessing on the Three-Decision Multiple-Choice Test: A Reply to C. F. Willey

Willey (1966) has recently commented that my demonstration of the effect of chance on the scores of the Three-decision Multiple-choice Test (Bernhardson, 1966) was not relevant if the scores were changed to percentiles or standard scores. This comment seems to imply that guessing is a problem only if it tends to raise che absolute level of most scores and thus make ic difficult to apply the traditional cut-off points for letter grades. This considers only one aspect of the problem. W e must bear in mind chac a score s h o ~ ~ l d reflect how much the examinee knows about the content being tested. Guessing still obscures the measure of the amount the examinee knows about the content even if percentiles or standard scores are used because guessing can influence the relative position of the examinee in the group. Thus an individual will do better in relation to the other examinees than he would on the basis of his knowledge of the content alone if he is luckier at guessing than che others. This increment due to guessing will be reflected in the percentile or standard score the individual obtains. Thus, using a score which indicates how well each examinee did in relation to all the ocher examinees as Willey suggested does not prevent guessing from influencing the scores. Nor does the solution seem to be the use of a formula to correct for guessing because it is difficult to determine how much any given student guessed. The best solution is to use the amount test scores are influenced by guessing as one of the criteria for selecting different types of tests. According to this criterion, the Three-decision Multiple-choice Test would not be an improvement over the conventional multiple-choice test. However, the case might be different for other criteria such as the assessment of partial knowledge.