Search

Data editing, a crucial task in the data production process, has received little scientific attention. Consequently, there is no consensus among social scientists about how data should be edited or by whom. While some argue that it should be left to data managers and data users, others claim that it is primarily a task for fieldworkers. The authors review these divergent approaches to editing and evaluate the underlying theoretical arguments. Results are reported from a methodological experiment in which different types of actors who are party to the data production and research process were asked to solve artificially generated inconsistencies in real survey data. Results are informative on two counts. First, the least accurate editors were the researchers with no field experience in the survey sites. Second, when provided with only partial information on which in make editing decisions, fieldworkers edited more accurately than both data managers and data users. [ABSTRACT FROM AUTHOR]

This article is a follow-up to Jones, Nagin, and Roeder (2001), which described an SAS procedure for estimating group-based trajectory models. Group-based trajectory is a specialized application of finite mixture modeling and is designed to identify clusters of individuals following similar progressions of some behavior or outcome over age or time. This article has two purposes. One is to summarize extensions of the methodology and of the SAS procedure that have been developed since Jones et al. The other is to illustrate how group-based trajectory modeling lends itself to presentation of findings in the form of easily understood graphical and tabular data summaries. [ABSTRACT FROM AUTHOR]

Surveys often sample adults across a broad range of ages, measuring the same outcomes in several interviews spaced during a period of years and comparing the changes observed across segments of the adult life course. Put in sequence, those change vectors provide a composite image of the outcome's life course trajectory. To illustrate, the authors estimate depression vectors in a sample of U.S. adults ages 18 and older at baseline in 1995, with follow-up interviews in 1998 and 2001. They show the vector equations and their graphs and also their synthetic-cohort projection. The authors introduce the trend-function and virtual-cohort projection, showing how they provide tests of "convergence" and other hypotheses about trajectories and trends. Results show depression dropping and then rising across adulthood more steeply than suggested by cross-sectional differences among age groups. They also indicate a rise and fall in age-specific levels of depression across cohorts. [ABSTRACT FROM AUTHOR]

McDonald and Moffitt's (1980) decomposition method for Tobit coefficients disaggregates the total changes in a censored outcome variable into the changes in the probability of attaining noncensored values and the changes in the noncensored values themselves. McDonald and Moffitt, however, did not highlight two limitations of this decomposition, and sociological applications of the Tobit decomposition that do not acknowledge these limitations have been misleading. First, the decomposition ratio of the two effects depends on the choice of an observation point in the sample. Some sociological studies invalidly generalized a decomposition at the mean point of the sample to the whole sample. Second, once the observation point is fixed, the decomposition ratio is the same across all coefficients or across all predictors. That is, the two effects remain proportionally constant, but studies often erroneously assume they differ across predictors. Further implications are discussed in relation to other selection bias models. [ABSTRACT FROM AUTHOR]

As the counterfactual model of causality has increased in popularity, sociologists have returned to matching as a research methodology. In this article, advances over the past two decades in matching estimators are explained, and the practical limitations of matching techniques are emphasized. The authors introduce matching methods by focusing first on ideal scenarios in which stratification and weighting procedures warrant causal inference. Then, they discuss how matching is often undertaken in practice, offering an overview of the most prominent data analysis routines. With four hypothetical examples, they demonstrate how the assumptions behind matching estimators often break down in practice. Even so, the authors argue that matching techniques can be used effectively to strengthen the prosecution of causal questions in sociology. [ABSTRACT FROM AUTHOR]

In this article, the authors examine the most common type of improper solutions: zero or negative error variances. They address the causes of, consequences of, and strategies to handle these issues. Several hypotheses are evaluated using Monte Carlo simulation models, including two structural equation models with several misspecifications of each model. Results suggested several unique findings. First, increasing numbers of omitted paths in the measurement model were associated with decreasing numbers of improper solutions. Second, bias in the parameter estimates was higher in samples with improper solutions than in samples including only proper solutions. Third, investigation of the consequences of using constrained estimates in the presence of improper solutions indicated that inequality constraints helped some samples achieve convergence. Finally, the use of confidence intervals as well as four other proposed tests yielded similar results when testing whether the error variance was greater than or equal to zero. [ABSTRACT FROM AUTHOR]

In this article, the author discusses on contradictions between generality and accuracy issues in sociological methodologies with considering the critical article by David L. Weakliem about the Bayesian information critenon (BIC). In sampling, efficient samples are often expensive to collect. In estimation, robust estimators tend to be inefficient. In data analysis, parsimonious models may not fit observed data. These kind of so-called trade-offs create tensions because researchers ideally prefer efficiency and low cost in sampling, robustness and efficiency in estimation, and parsimony and good fit in data analysis. Weakliem has pointed out two major limitations of the BIC. At first, Weakliem shows that Bayes factors depend on priors pertaining to the distribution of unknown parameters, whereas the BIC does not. Second, the BIC takes into consideration only the total sample size but not its distributions across tables, categories, or cells. Weakliem makes a good case that the BIC has certain limitations. The key question is whether the limitations are so severe as to render the BIC useless in applied research.

This article presents a commentary evaluation of the critical article, by David L. Weakliem, on the Bayesian Information Criterion (BIC) for Model Selection, contained in this issue of the journal "Sociological Methods & Research." The author describes that Weakliem makes two main points about the BIC approximation. The first is that BIC yields an approximation to Bayes factors that corresponds closely to a particular prior on the model parameters. This may or may not be similar to the researcher's actual prior information or beliefs, and the Bayes factor it yields may or may not be close to the Bayes factor resulting from the researcher's prior. In most cases, BIC is more conservative than the alternative priors proposed by Weakliem and other critics of BIC, in the sense that the alternative priors are more likely to find evidence for the effect being studied. In practice, therefore, there seems to be some agreement that BIC is sufficiently conservative, perhaps too conservative. It follows that, in most cases, if BIC finds evidence for an effect, we should agree that the data support the effect.

In this article authors discusses over the critical article, by David L. Weakliem, on the Bayesian Information Criterion (BIC) for Model Selection, contained in this issue of the journal "Sociological Methods & Research" itself. Authors report that Weakliem's article makes three main points. First, that the BIC corresponds to a specific probability model that, in many examples, does not make good scientific sense. Second, that the mathematical form of the BIC can lead to pathologies with respect to sample size, data grouping, and model formulation, as illustrated with several real examples in Weakliem's article. Third, that using the BIC to select a model leads to scientifically misleading conclusions. This third point is a problem with any automatic model selection method. One can generally use both analytical and empirical approaches to evaluate applied statistical methods, but empirical evaluations and narratives of applications are suggestively more important determinants of how useful researchers view a statistical method to be in a specific application context.

This article presents a commentary note on the critical article, by David L. Weakliem, on the Bayesian Information Criterion (BIC) for Model Selection, contained in this issue of the journal "Sociological Methods & Research" itself. According to the authors of this article, Weakline has exposed some fundamental problems with the use of the BIC and with "automatic" model selection criteria generally. As is pointed out by Weakliem, a Bayes factor or the BIC, suggestively, cannot tell which statistically significant effects are important in relation to substantive questions of interest. The model or models most appropriate to the purposes of an investigation will typically depend on what those purposes are. In particular, the existence and quantification of small effects, which for some purposes might be best neglected, are often a focus of substantive interest, perhaps especially in studies involving large amounts of data. For judging the substantive importance of effects found to be "real" by a test of significance, responsibility rests with the investigator. Bayesian hypothesis-testing paradigm revolves around a probability, the probability that the effect under study is nonzero; subject-matter importance of the effect plays no part in the motivation or calculation of such probability. Thus, when a Bayes factor or the BIC succeeds in identifying an effect of substantive importance, such success are supposed to be regarded as just a positive luck.

This article presents a critical note on the Bayesian information criterion (BIC) which has recently been proposed as an effective method of model selection in sociological research. The BIC is defined as L2-rlog(N), where "N" is the total sample size and "r" is the number of degrees of freedom remaining after fitting the model, with smaller values representing better fits. Because it includes an adjustment for sample size, the BIC often favors a simpler model even when that model would be rejected by standard hypothesis tests. The BIC was initially proposed for log linear models, and has been employed in most articles using log linear or log multiplicative models published in the leading American sociology journals in the 1990s. The BIC is based on a Bayesian approach to hypothesis testing. In this approach, the observed data and prior beliefs about the probable values of the parameters are combined to produce a Bayes factor; that is, a measure of the weight of evidence that the data provide for or against the null hypothesis. In general, different prior beliefs produce different Bayes factors, and in principle anyone with access to the data could compute the Bayes factor that follows from his or her own beliefs.

A linear structural equation model (SEM) without free parameters has two parts: a probability distribution and an associated path diagram corresponding to the causal relations among variables specified by the structural equations and the correlations among the error terms. This article shows how path diagrams can be used to solve a number of important problems in structural equation modeling; for example, How much do sample data underdetermine the correct model specification? Given that there are equivalent models, is it possible to extract the features common to those models? When a modeler draws conclusions about coefficients in an unknown underlying SEM from a multivariate regression, precisely what assumptions an being made about the SEM? The authors explain how the path diagram provides much more than heuristics for special cases; the theory of path diagrams helps to clarify several of the issues just noted. [ABSTRACT FROM AUTHOR]

Social research often involves estimating the effects of a categorical treatment on a dependent outcome variable. Endogenous switching regression models are natural extensions of classical experimental designs, which allow tests of assumptions about the exogeneity of treatment effects from survey data. Endogenous switching regression applications are pervasive in the econometric literature and are becoming widely used by sociologists. However, sociologists have devoted somewhat less attention to endogenous switching models involving limited dependent variables. Switching regression models for continuous variables can be generalized to account for binary and censored dependent variables. These models are applied to estimate the effect of family structure on early family formation. [ABSTRACT FROM AUTHOR]

Multinomial logit discrete-choice models are now being used more frequently in sociological research, but often without appropriate attention to the central role played by the independence of irrelevant alternatives (IIA) property. In this article, three tests of the IIA property are discussed and then how to apply them is illustrated, using a data set previously analyzed by Hoffman and Duncan. The IIA tests decisively reject the IIA property in the single/welfare/remarriage choice model. [ABSTRACT FROM AUTHOR]

Data are called ipsative if they are subject to a constant-sum constraint for each observation. Usually, ipsative data are the consequence of transformation of their corresponding preipsative data. In this article, two kinds of ipsative data are defined. They are the additive ipsative data (AID) and the multiplicative ipsative data (MID). Jackson and Alwin proposed a method to analyze AID in exploratory factor analysis. However, they failed to provide the estimates of the original factor loadings. In this study, first, their method is modified in the context of covariance structure analysis. It is discovered that the original parameter estimates can usually be recovered provided the model of the preipsative data is well defined. An artificial example is used to demonstrate the suggested method Second, the method is extended to the case of MID. A real example is considered also. Finally, some related issues, problems, and generalizations are addressed and discussed. [ABSTRACT FROM AUTHOR]

Model modification in covariance structure analysis through reducing constraints can have an impact on the estimates of the maintained free parameters if the model is reevaluated. Three new statistics that focus on the estimated changes of parameter estimates and estimated sampling variability of the maintained free parameters are developed in this study. Another new statistic focuses on the significance of the estimated change in restricted, especially, fixed, parameters. An empirical investigation of the performance of these four new statistics showed that they can provide valuable supplementary information to the Lagrange multiplier statistic. [ABSTRACT FROM AUTHOR]

Although methods for latent variable modeling that allow a joint analysis of measurement and theory have become popular, they are not without difficulties. As these difficulties have become more apparent, several researchers have recently called for a “two-step approach” to latent variable modeling in which measurement is evaluated separately from theory. This implies that programs for covariance structure analysis are not needed because factor analysis and regressions would suffice for analysis. Before a return to earlier practice using seemingly simpler analysis tools can be recommended, it seems prudent to consider the assumptions underlying a two-step approach. At least four implicit assumptions can be identified: (a) theory and measurement are independent, (b) results of factor analysis specifications can be generalized to other specifications, (c) the estimators have desirable statistical properties, and (d) the statistical test in one step is independent of the test in the other. The authors show that these assumptions cannot be met in general and that some of them are logically inconsistent. Thus any wholesale adoption of a two-step approach could have serious consequences. [ABSTRACT FROM AUTHOR]

Fornell and Yi (1992 [this issue]) have discussed four assumptions that they contend underlie two-step approaches to structural equation modeling. Each of these assumptions is demonstrated to not be an assumption of the two-step approach recommended by Anderson and Gerbing (1988). In doing so, an attempt is made to provide some clarification and guidance to researchers interested in employing structural equation modeling to test and develop theory. Given the comparative strengths of the two-step approach over a one-step approach in practice, it is concluded that the preferred approach to the modeling task is the two-step approach of Anderson and Gerbing (1988). [ABSTRACT FROM AUTHOR]

Measurement error in independent variables produces biased estimates of the coefficients in linear models. These biases can be reduced by obtaining repeated measurements of the variables and then estimating structural equation models with multiple indicators of latent variables. Remeasurement is usually costly, however, raising the question of whether the same benefits can be obtained by remeasuring only a fraction of the sample. Although this strategy has been fried previously, there were no appropriate statistical methods for combining the data in the remeasured subsample and the single measurement subsample. We demonstrate here how recently developed methods for incomplete data provide an attractive solution to this estimation problem. The methodology is illustrated by a reanalysis of Bielby, Hauser, and Featherman's (1977a) study of the OCG-II data. [ABSTRACT FROM AUTHOR]

hi this article we consider some latent variable models for dichotomous manifest variables that have recently appeared in sociology. The latent trait models considered are a logit/probit one-parameter, a logit/probit two-parameter, and a log-linear model. Some related latent class models for response consistency are also considered. We discuss both statistical and logical relations among these model and propose strategies for model selection, demonstrating those strategies in several examples. We also discuss the usefulness of the various models. [ABSTRACT FROM AUTHOR]

This article deals with some problems of event history analysis that arise in special types of sampling plans where spells in progress at time t??? the start date of the observation period, are sampled. Then, the available data are random samples of interrupted spells, or else may be spells that begin after the start date of the sample. For the interrupted spells, all relevant probability distributions of the elapsed duration until to and the duration of the spell after t[sub 0] have to account for the condition that the individual is in the state Y[sub 0] in question at time t[sub 0] Relying on results of Ridder (1984) and Heckman and Singer (1986), we discuss the distribution of the “sampled” spells (i.e., conditional on the presence of the individual in state y[sub 0] at time t[sub 0]), in particular the joint distribution of pre-interview and post-interview durations. In general, this distribution depends on the entry rate into the state Y[sub 0]. The distribution of the sampled spells for the special sampling plans differs from the corresponding duration distributions in the population. We then show that, sampling spells in progress, the distribution of the explanatory variables in the sample becomes informative for the parameters of interest. Next, we investigate the distribution of the sampled spells of individuals who enter or leave the state in question. This referred to as “sampling the flow.” We then discuss the special case of a time invariant entry rate. Appropriate likelihood contributions are given depending on which pre-interview, post-interview, and completed durations are observable. We also discuss a conditional likelihood approach due to Lancaster (1979) that does not require any assumptions on the entry rate. Next we derive the appropriate likelihood contributions of right censored observations and both left and right censored observations for the specific sampling plans considered here. Finally, we present the... [ABSTRACT FROM AUTHOR]

This issue of Sociological Methods & Research contains four excellent illustrations of current methods for analyzing dichotomous data: dichotomous factor analysis (Muthén, 1989), latent trait analysis (Eaton et al., 1989), and grade of membership analysis (Woodbury and Manton, 1989). These methods are now sufficiently developed for use by researchers, though a lack of available software for the grade of membership model renders that method mostly theoretical. Of these methods the factor analysis model succeeds in carrying out the tasks set by the editors with the fewest problems. Researchers would do welt, however, to look over all the methods to find the best approach for their particular problem. [ABSTRACT FROM AUTHOR]

Theories of cognitive consistency are formally integrated and compared by applying a multiple regression modeling strategy that has developed in the analysts and formal integration of theories of social consistency. Several important studies of cognitive consistency theories are reviewed to locate conceptual omissions in tests of the theory. In addition, several issues that are problematic in testing consistency theories are reviewed, alternative solutions are discussed, and areas in which additional work is needed are noted. [ABSTRACT FROM AUTHOR]

This paper is concerned with the interpretation of fitted log-linear models in contingency table analysis. The two most commonly used alternatives are the direct interpretation of the estimated coefficients of the fitted model and the interpretation of statistics, such as odds ratios, based upon the expected cell frequencies under the fitted model. There are strong reasons for preferring the latter since the statistics will be the same whatever computer program is used to fit the model whereas the estimated parameter coefficients will depend upon the set of constraints used to solve the normal equations Thus, different computer programs can (correctly) provide different parameter estimates from the same data whereas statistics such as odds ratios based upon the expected cell frequencies will be unaffected. Interpretation based upon the estimated parameters of the fitted model is therefore potentially misleading. These problems are not unique to log-linear analysis but occur in analyses involving the generalized linear model. [ABSTRACT FROM AUTHOR]

Social scientists have available a wide array of techniques for the analysis of complex contingency tables. This paper discusses the factors which should be considered in choosing which of the various techniques to use. The theme emphasized here is that the analyst should evaluate the techniques within the context of the substantive questions he seeks to answer. The specific factors discussed are categorized into three groups: representation of the dependent variable, independent variable models, and estimation procedures. [ABSTRACT FROM AUTHOR]

It is noted that a ‘forwards incremental’decomposition of variance meshes particularly well with a recursive path analysis formulation of a social process, and it can readily be calculated from the model, knowing only the paths. From this its suggested that it may be substantively illuminating to trace residual effects forwards through a standardized model, and not to confine interpretation of residual effects to the ‘parent’ variables. A simple interpretation of the residual path is also presented. [ABSTRACT FROM AUTHOR]