On structural equation modeling with data that are not missing completely at random

A general latent variable model is given which includes the specification of a missing data mechanism. This framework allows for an elucidating discussion of existing general multivariate theory bearing on maximum likelihood estimation with missing data. Here, missing completely at random is not a prerequisite for unbiased estimation in large samples, as when using the traditional listwise or pairwise present data approaches. The theory is connected with old and new results in the area of selection and factorial invariance. It is pointed out that in many applications, maximum likelihood estimation with missing data may be carried out by existing structural equation modeling software, such as LISREL and LISCOMP. Several sets of artifical data are generated within the general model framework. The proposed estimator is compared to the two traditional ones and found superior.The research of the first author was supported by grant No. SES-8312583 from the National Science Foundation and by a Spencer Foundation grant. We wish to thank Chuen-Rong Chan for drawing the path diagram.     

Dissociation of semantic priming, word and non-word repetition effects by event-related potentials

Two experiments investigated the modulation of event-related potentials (ERPs) by semantic priming and item repetition. In Experiment 1, subjects silently counted occasional non-words against a background of words, a proportion of which were either semantic associates or repetitions of a preceding word. Compared to control items, ERPs to repeated words were distinguished by an early (ca. 200 msec) transient negative-going deflection and a later, topographically widespread and temporally sustained positive-going shift. In contrast, semantically primed words showed a relatively small, topographically and temporally limited positive-going modulation peaking around 500 msec. These data were interpreted as evidence against models of priming and repetition which postulate similar loci for these effects. In Experiment 2, subjects counted occasional words against a background of non-words, some of which were repeated. ERPs to repetitions showed a similar early ERP modulation to that in Experiment 1, and also displayed a later slow positive shift. This latter effect was smaller in magnitude and had a delayed onset in comparison to Experiment 1. It was concluded that the effects of repetition differ as a consequence of whether, prior to their first presentation, items possess a representation in lexical memory.     

Smoking,personality, and imagery preference in Relation to spontaneous bilateral electrodermal activity

Effects of smoking on spontaneous bilateral skin conductance were studied as a consequence of interactions with set and setting. Bilateral skin conductance was recorded in six men and six women who smoked and an equal number of nonsmokers on two different days. On the first day, spontaneous electrodermal activity was recorded before and after a 10-minute rest period during which half the smokers smoked a cigarette. On the second day, one week later, spontaneous activity was again recorded before and after a 10-minute rest period during which the remaining smokers smoked a cigarette. Nonsmokers did not smoke at any time. Following smoking there was a significant negative correlation between a preference for visual over nonvisual imagery and spontaneous skin conductance responsivity in the left as compared to the right hand. Correlations were significantly different in smokers and nonsmokers. Smokers showed significantly greater preferences for either visual or nonvisual imagery than nonsmokers. Nonresponding nonsmokers were higher on the psychoticism (P) scale of the Eysenck Personality Questionnaire (EPQ) than responding nonsmokers. Smokers scored higher on P than nonsmokers. Interactions between personality, cognitive style, and the experimental situation, set, and setting were discussed in relation to the arousing effects of nicotine.     

Developing probability limits for longitudinal physiological data

The objective of this study was to determine what amount of change in heart electrophysiology is necessary to yield a significant difference between a baseline condition and an experimental treatment condition. In a single-subject longitudinal design, a modification of at test was applied to baseline data, and the performance levels for both .05 and .01 probability levels were obtained. It was concluded that it is both ethically sound and physiologically safe to permit a subject in this type of experiment to physiologically respond at levels that would reach both the .05 and the .01 levels of significance.     

Adolescent alcoholism in Benin City, Nigeria

The consumption of alcohol by Nigerian adolescents is investigated. Adolescence and alcoholism as applied in this study are defined, and the financial costs and physical and mental effects of alcohol use are described. The author used unobtrusive methods to gather data, visiting three nightclubs and an amusement park where alcohol is served in order to observe and interview adolescents. The findings show that adolescents begin drinking at an early age and consume large quantities of beer. The findings are consistent with those of other epidemiological studies. Finally, recommendations for controlling adolescent alcohol use are made.     

Belief systems and attitudes toward the death penalty and other punishments

This study focused on the relationship of belief systems as a configural construct and conservatism-liberalism to attitudes toward the death penalty and other punishments for offenses of varying severity Extrapersonalists, the most concretely functioning of the four conceptual or belief systems posited by Harvey, Hunt, and Schroder (1961), were most in favor of the death penalty They also endorsed the most severe punishment for a variety of crimes, to a particularly greater extent than did representatives of either of the two more abstractly functioning systems The belief dimensions of Openness, Evaluativeness, and Complexity, as well as the frequency of church attendance, correlated more highly and consistently with attitudes toward punishment than did either Conservatism-Liberalism or gender Both of the latter variables failed to correlate with a number of the outcome variables and related at only low levels to the others The greater predictive power of a configural conception of personality or belief systems over a unidimensional conception seems to have been demonstrated Configural concepts may additionally be generally superior to multidimensional concepts treated linearly