Paul D. Allison

Discrete-Time Methods for the Analysis of Event Histories

The history of an individual or group can always be characterized as a sequence of events. People finish school, enter the labor force, marry, give birth, get promoted, change employers, retire, and ultimately die. Formal organizations merge, adopt innovations, and go bankrupt. Nations experience wars, revolutions, and peaceful changes of government. It is surely the business of sociology to explain and predict the occurrence of such events. Why is it, for example, that some individuals try marijuana while others do not? Why do some people marry early while others marry late? Do educational

Missing data techniques for structural equation modeling.

As with other statistical methods, missing data often create major problems for the estimation of structural equation models (SEMs). Conventional methods such as listwise or pairwise deletion generally do a poor job of using all the available information. However, structural equation modelers are fortunate that many programs for estimating SEMs now have maximum likelihood methods for handling missing data in an optimal fashion. In addition to maximum likelihood, this article also discusses multiple imputation. This method has statistical properties that are almost as good as those for maximum likelihood and can be applied to a much wider array of models and estimation methods.

CHANGE SCORES AS DEPENDENT VARIABLES IN REGRESSION ANALYSIS

Change scores have been widely criticized for their purported unreliability and for their sensitivity to regression toward the mean. These objections are shown to be unfounded under a plausible regression model for the nonequivalent control group design. This model leads to inferences that are intuitively correct, as judged by changes in means over time, while the conventional model leads to inferences that are intuitively false. Moreover, the conventional model implies that regression toward the mean within groups leads to regression toward the mean between groups, an implausible result for naturally occurring groups. Nevertheless, the conventional model may be more appropriate when there is a true causal effect of the pretest on the posttest, or when cases are assigned to groups on the basis of their pretest scores.

Comparing Logit and Probit Coefficients Across Groups

In logit and probit regression analysis, a common practice is to estimate separate models for two or more groups and then compare coefficients across groups. An equivalent method is to test for interactions between particular predictors and dummy (indicator) variables representing the groups. Both methods may lead to invalid conclusions if residual variation differs across groups. New tests are proposed that adjust for unequal residual variation.

Fixed–Effects Negative Binomial Regression Models

This paper demonstrates that the conditional negative binomial model for panel data, proposed by Hausman, Hall, and Griliches (1984), is not a true fixed-effects method. This method—which has been implemented in both Stata and LIMDEP—does not in fact control for all stable covariates. Three alternative methods are explored. A negative multinomial model yields the same estimator as the conditional Poisson estimator and hence does not provide any additional leverage for dealing with over-dispersion. On the other hand, a simulation study yields good results from applying an unconditional negative binomial regression estimator with dummy variables to represent the fixed effects. There is no evidence for any incidental parameters bias in the coefficients, and downward bias in the standard error estimates can be easily and effectively corrected using the deviance statistic. Finally, an approximate conditional method is found to perform at about the same level as the unconditional estimator.

Multiple imputation for missing data : A cautionary tale

Two algorithms for producing multiple imputations for missing data are evaluated with simulated data. Software using a propensity score classifier with the approximate Bayesian bootstrap produces badly biased estimates of regression coefficients when data on predictor variables are missing at random or missing completely at random. On the other hand, a regression-based method employing the data augmentation algorithm produces estimates with little or no bias.

Productivity Differences Among Scientists: Evidence for Accumulative Advantage

The highly skewed distributions of productivity among scientists can be partly explained by a process of accumulative advantage. Because offeedback through recognition and resources, highly productive scientists maintain or increase their productivity, while scientists who produce very little produce even less later on. A major implication of accumulative advantage is that the distribution of productivity becomes increasingly unequal as a cohort of scientists ages. Cross-sectional survey data support this hypothesis for chemists, physicists, and mathematicians, who show strong linear increases in inequality with increasing career age. This increase is highly associated with a changing distribution of time spent on research. Another implication of accumulative advantage is also corroborated: the association among productivity, resources and esteem increases as career age increases.

Testing for Interaction in Multiple Regression

Contrary to a recent claim, the inclusion of a product term in a multiple regression is a legitimate way to test for interaction. The unstandardized coefficient and the t-test for the product term are unaffected by the addition of arbitrary constants to the variables in the model. Certain other statistics are affected by this change, however, indicating that some hypotheses relating to interaction are not meaningfully testable unless variables are measured on ratio scales.

RANK ADVANCEMENT IN ACADEMIC CAREERS: SEX DIFFERENCES AND THE EFFECTS OF PRODUCTIVITY*

Advancement in rank is critically important to the career of an academic scientist, and the highly visible nature of the event makes it idealfor studying stratification in science. Concern with universalistic factors in promotion has prompted debates over two issues. First, why do female scientists advance more slowly than male scientists, and why do so few reach the rank offull professor? Second, is promotion driven by the sheer volume of published work as opposed to its quality? Event history analyses clearly indicate that quantity of publications is far more important than various measures of quality of publications in predicting rank advancement; and women are less likely to be promoted than men. About one-half of this sex difference is attributable to differences in levels of variables affecting promotion. Remaining differences are a result of differences in expected timing of promotion to associate professor and to the negative effects of department prestige on promotion to full professor for women.

DEPARTMENTAL EFFECTS ON SCIENTIFIC PRODUCTIVITY

Productive scientists tend to hold jobs at prestigious university departments, but it is unclear whether this is because good departments hire the best scientists or because good departments encourage and facilitate research productivity. To resolve this issue, we studied the antecedents and consequences of 179 job changes by chemists, biologists, physicists, and mathematicians. Those who were upwardly mobile showed substantial increases in their rate of publication and in the rate of citation to those publications, while those who were downwardly mobile showed substantial decreases in productivity. Earlier analyses of these job changes found only a small effect of prior productivity on destination prestige. These results suggest that the effect of department affiliation on productivity is more important than the effect of productivity on departmental affiliation.