Michael Gillespie

Log-Linear Techniques and the Regression Analysis of Dummy Dependent Variables

Although previous comparisons of log-linear techniques with the regression analysis of dummy dependent variables have focused on the statistical superiority of the log-linear techniques, this paper presents three advantages of dummy dependent-variable regressions. First, dummy dependent-variable regression is able to accommodate both categorical and continuous independent variables. Second, the calculation of indirect effects and reciprocal effects is possible only with dummy dependent-variable regression. Third, the slopes yielded by dummy dependent-variable regression possess the properties of fundamental parameters; the effect parameters of log-linear models do not. In particular, controlling for a variable that is irrelevant to the causal system under investigation can yield a partial relationship whose expected value differs from that of the zero-order relationship when using log-linear techniques.

Attachment in African American and European American older adults: The roles of early life socialization and religiosity

The significance of attachment relationships in later life has recently emerged as an important topic of study. Yet little attention has focused on attachment dynamics among older minority adults. This paper extends the literatures on ethnicity, attachment, and later life by examining attachment patterns in two large community-dwelling samples of older (65 + years) African American (n=671) and European American (n=447) adults. Data gathered during face-to-face interviews included demographic information, adult attachment, early rearing experiences, and current religiosity. Hierarchical regression analyses revealed that adult attachment dimensions were differentially predicted by childhood socialization patterns and current religiosity. Significant ethnic differences in relations between adult attachment and childhood socialization practices also were found. The results highlight the importance of examining contextual differences in attachment in later life.

Suicide Following Homicide in Canada

Sucide following homicide is fairly rare. Even more rare are studies of this event that use national data. In Canada, where national data are available, homicide offenders commit suicide in about 10% of the cases. Following Henry and Short, we hypothesize that the probability of suicide following homicide increases when the offender has close social ties to the victim and/or to society. We use data on all homicides committed in Canada by male offenders between 1961 and 1983 to explore this issue. Our results show that the closer the tie between the offender and the victim, the higher the probability that the offender will commit suicide. Furthermore, the probability of suicide increases with the offenders age and education, is higher when the offender uses a gun, and is higher when the victim is female.

Using Mokken scale analysis to develop unidimensional scales

This paper describes Mokken scale analysis as a method for assessing the unidimensionality of a set of items. As a nonparametric stochastic version of Guttman scale analysis, the Mokken model provides a useful starting point in scale construction since it does not impose severe restrictions on the functional form of the item trace lines. It requires only that the item trace lines are monotonically increasing and that they do not cross. After describing the Mokken method, we illustrate it by analyzing six abortion items from the 1975–1984 NORC General Social Surveys. In contrast to earlier parametric analyses of these items (regular and probit factor analyses), we find that these items form a single dimension. We argue that the two-dimension solution of these earlier analyses is an artifact of the differences in the difficulty of the items.

The differential roles of early emotion socialization and adult attachment in adult emotional experience: testing a mediator hypothesis.

The goal of the present study was to model the relations among self-reported early emotion socialization, adult attachment styles, and positive and negative adult emotion experiences in younger (mean age — 28) and older (mean age — 74) adults. Using structural equation modeling, we found that reports of early emotion socialization had both direct and indirect effects (the latter mediated by attachment style) on adult emotional experiences. There was also some support for the expectation that age would interact with emotion socialization in the effect on adult emotions. Results are discussed in terms of their implications for understanding how attachment and emotion relate across the lifespan.

Secular Trends in Abortion Attitudes: 1975-1980-1985

In this study, we compared the attitudes toward abortion of respondents from the National Opinion Research Centers (NORC) General Social Survey (GSS) for the years 1975, 1980, and 1985. The major theme of the GSS is that an adequate description of changes in abortion attitudes is predicted on resolving the question of whether the six abortion items used in the surveys measure one or two dimensions. We used the Mokken method of scale analysis, a stochastic extension of Guttman scaling, to resolve this question. We concluded that the six items are unidimensional and, therefore, create a single scale to measure the changes in abortion attitudes across the three periods. The data revealed virtually no change from 1975 to 1980, but a significant, though small, drop in approval from 1980 to 1985.

Breast self-examination practices among women from six ethnic groups and the influence of cancer worry

SummaryDiverse samples of women (N = 1364) from Brooklyn, New York, were interviewed regarding their breast cancer screening practices. Of interest here is the relation between cancer worry and adherence to breast self-examination (BSE) guidelines among the six ethnic groups identified – European American, African American, Haitian, Dominican and English-speaking Caribbean women. There was a significant difference in cancer worry by ethnicity. Logistic regression analysis indicated that education, cancer worry, and perceived efficacy of BSE significantly predicted adherence. Furthermore, there were significant differences among women of African descent in BSE adherence and cancer worry. These differences reflect the emerging need for researchers to empirically and methodically investigate ethnic and cultural factors, as well as emotions and affect in preventive health behaviors.

Using Mokken methods to develop robust cross-national scales: American and West German attitudes toward abortion

A fundamental methodological issue in cross-national research on attitudes is the comparability of the attitude measures across populations. We address this issue by presenting the Mokken method and accompanying Mokken test as a means for developing equivalent attitude scales. We apply these methods to an analysis of the responses to seven abortion items in the 1982 NORC GSS and West German ALLBUS combined files. We find that the seven items form a unidimensional scale in both countries and that four of these items constitute a scale that is robust across the two populations. We conclude by describing how such results can be used to guide the development of Rasch and LISREL models.

Testing the assumptions and interpreting the results of the Rasch model using log-linear procedures in SPSS

This paper shows how to use the log-linear subroutine of SPSS to fit the Rasch model. It also shows how to fit less restrictive models obtained by relaxing specific assumptions of the Rasch model. Conditional maximum likelihood estimation was achieved by including dummy variables for the total scores as covariates in the models. This approach greatly simplifies the specification of the Rasch models. We illustrate these procedures in an analysis of four items selected from the Reiss Premarital Sexual Permissiveness Scale. We found that a modified version of the Rasch model with item dependencies fits the data significantly better than the simple Rasch model. We also found that the item difficulties are the same for men and women, but that the item dependencies are significantly greater for men. Apart from any substantive issues these results raise, the value of this exercise lies in its demonstration of how researchers can use the procedures of popular, accessible software packages to study an increasingly important set of measurement models.

Anger Experience and Anger Inhibition in Sub-Populations of African American and European American Older Adults and Relation to Circulatory Disease

This study examined ethnic differences in the link between anger experience and anger inhibition and that of circulatory disease (CD). To ascertain the effects of anger inhibition in older persons, health data from groups of African American, African Caribbean, Eastern European and European American adults were collected. Experienced anger and anger inhibition were significant predictors of CD only for the African American group and the relation between experienced anger and CD was mediated by anger inhibition. The data suggest that cultural factors play a role in the development of an angerinhibitory style and that this trait may pose a serious risk factor for circulatory disease.