Moderated regression analysis and Likert scales: too coarse for comfort.

One of the most commonly accepted models of relationships among three variables in applied industrial and organizational psychology is the simple moderator effect. However, many authors have expressed concern over the general lack of empirical support for interaction effects reported in the literature. We demonstrate in the current sample that use of a continuous, dependent-response scale instead of a discrete, Likert-type scale, causes moderated regression analysis effect sizes to increase an average of 93%. We suggest that use of relatively coarse Likert scales to measure fine dependent responses causes information loss that, although varying widely across subjects, greatly reduces the probability of detecting true interaction effects. Specific recommendations for alternate research strategies are made.     

Prejudice Against a Progay Man in an Everyday Situation: A Scenario Study1

A scenario study was performed to investigate whether people in the street would be less likely to respond positively to a stranger's request for a small favor if the stranger wore a T-shirt bearing a progay slogan. One hundred and eighteen female and 112 male participants were each presented with six scenarios, all of which asked them to imagine being approached by a person requesting change for a banknote. The dependent variable in the study was the participant's judgment, on a 7-point scale, of how likely he or she would be to help. Five of the scenarios were included as distractors. In the critical scenario, the requester was described as wearing a plain T-shirt or one with either a progay or a control slogan printed on it. The male participants in the progay condition showed a significantly lower level of help than those in either of the control groups. The antigay pattern, however, was not obtained in the responses of the female participants. The results are discussed with reference to other findings in the literature, and their implications for people's responses to progay persons in everyday life are considered.     

The Discrediting Effect in Eyewitness Testimony1

Loftus (1974) had subjects read summaries of criminal trials that contained the testimony of either credible or discredited prosecution eyewitnesses, and found no effect of discrediting an eyewitness. Instead, almost as many subjects voted guilty with a discredited eyewitness as with a credible eyewitness; this led Loftus to the conclusion that jurors tend to overbelieve eyewitness testimony. Loftus's conclusion was subsequently challenged by others who reported a strong discrediting effect. A series of three experiments using college students was conducted to explore the characteristics of trial summaries that might account for the discrepancy in results, such as inclusion of judicial instructions concerning proof beyond a reasonable doubt, or an eyewitness's reaffirmation of his testimony following discrediting. In all cases, a strong discrediting effect was found. Apparently the discrediting effect appears regardless of wide variation in content of trial summaries. The present data do not support the overbelief claim.     

An "Epidemic" Model of Adolescent Cigarette Smoking

The prevalence of regular smoking increases rapidly during adolescence. We applied modeling methods from the study of epidemic infectious disease to smoking behavior. Our "epidemic" models dealt with the process of transition among four states: nonsmoker (susceptible), trier (exposed), regular smoker (infected), and ex-smoker (recovered). The best fitting model was one in which the transition from nonsmoker to trier was a social contagion process where nonsmokers contacted either triers or regular smokers. All other transitions were treated as constant rate processes in which a constant proportion of eligible individuals made the transition in a given year. The recursive equation model embodying these assumptions was able to predict accurately the proportion of adolescents in each state in Grades 6–11. Our results support a "stage" model of smoking, suggesting that different psychological and physiological processes are involved in the initiation of smoking than in developing a habit of regular smoking. A possible sex difference was found in the non-contact transition between trier and smoker states, with the transition being more probable in females than in males.     

Social support, depressed mood, and adjustment to stress: a genetic epidemiologic investigation.

A survey of 821 same-sex female twin pairs from a population-based registry assessed 8 dimensions of social support and social integration. Twin analyses documented significant common environmental influences on 5 of these 8 measures and significant genetic influences on 5 of the 8. A decomposition of the multiplicative association between support and a measure of stressful life experiences in predicting depressed mood--an association typically interpreted as providing evidence for a stress-buffering effect of social support--shows clearly that it is the environmental and genetic factors that cause support, rather than support itself, that buffer the effects of stress on mood in most cases. We discuss the implications of this result for future research on the relationship between social support and psychopathology.     

Evidence for an error deadzone in compensatory tracking

Humans and monkeys show intermittent arm movements while tracking moving targets. This intermittency has been explained by postulating either a psychological refractory period after each movement and/or an error deadzone, an area surrounding the target within which movements are not initiated. We present a technique to detect and quantify the size of this deadzone, using a compensatory tracking paradigm that distinguishes it from a psychological refractory period. An artificial deadzone of variable size was added around a visual target displayed on a computer screen. While the subject was within this area, he received visual feedback that showed him to be directly on target. The presence of this artificial deadzone could affect tracking performance only if it exceeded the size of his intrinsic deadzone. Therefore, the size of artificial deadzone at which performance began to be affected revealed the size of the intrinsic deadzone. Measured at the subjects' eye, the deadzone was found to vary between 0.06 and 0.38 degrees, depending on the tracking task and viewing conditions; on the screen, this range was 1.3 mm to 3.3 mm. It increased with increasing speed of the target, with increasing viewing distance, and when the amplitude of the movement required was reduced. However, the deadzone size was not significantly correlated with the subjects' level of performance. We conclude that an intrinsic deadzone exists during compensatory tracking, and we suggest that its size is set by a cognitive process not simply related to the difficulty of the tracking task.     

Affect, personality, and social activity.

These studies examined relations between social activity and state and trait measures of Positive and Negative Affect. In Study 1 Ss completed scales relevant to 3-factor models of personality and a weekly mood and social activity questionnaire for 13 weeks. In Study 2 Ss completed measures of the 5-factor model of personality and a daily mood and social activity survey for 6-7 weeks. In within- and between-Ss analyses, socializing correlated significantly with state measures of Positive Affect and with trait measures of Extraversion/Positive Emotionality. These relations were relatively general across various types of positive affect and social events; however, specific types of social events also were differentially related to affect. In contrast, social activity had no consistent association with measures of Negative Affect or the other personality dimensions. The results support a temperamental view of Extraversion.     

Repression at encoding: discrete appraisals of emotional stimuli.

In previous research, the emotions associated with repressors' memorial representations were found to be more discrete than those associated with nonrepressors'. In each of the 3 experiments reported here, repressive discreteness was apparent in repressors' appraisals of emotional stimuli at the time they were encoded. In 1 experiment, Ss appraised individual facial expressions of emotion. Repressors judged the dominant emotions in these faces as no less intense than did nonrepressors, but they appraised the blend of nondominant emotions as less intense than did nonrepressors. In the remaining 2 experiments, Ss appraised crowds of emotional faces as well as crowds of geometric shapes. In both crowd experiments, the repressive discreteness was evident in appraisals of crowds of emotional faces but not in appraisals of crowds of geometric shapes. The repressive discreteness effect did not appear to reflect a general repressor-nonrepressor difference in the appraisal of stimulus features. Rather, the results suggested that repressive discreteness may be constrained to appraisals of emotions.     

Rorschach object relations of adolescents who committed homicide.

We investigated the object relations of adolescents who committed homicide. A clinical sample of 55 adolescents who committed homicide did not differ from a comparison group of nonviolent delinquents on Rorschach measures of object differentiation, mutuality of autonomy, and aggressive content. However, the subgroup of adolescents who committed homicides in the context of another crime (e.g., robbery or burglary) did manifest significantly lower object relations (poorer object differentiation and more victim responses) than the subgroup of adolescents whose homicides were committed in the context of an interpersonal conflict or dispute with the victim. These findings support the need for differentiated classification of violent individuals as urged by Megargee (1970) over 20 years ago.