Showing papers in "Sociological Methods & Research in 1991"

The Reliability of Survey Attitude Measurement: The Influence of Question and Respondent Attributes

Abstract: Several theoretical hypotheses are developed concerning the relation of question and respondent characteristics to the reliability of survey attitude measurement. To test these hypotheses, reliabil...

Adding Social Structure to Diffusion Models An Event History Framework

Abstract: Most diffusion models make the implausible assumption that diffusion is unstructured by relationships within the population of interest. This article proposes methods for incorporating a priori notions about social structure into analysis. Diffusion is modeled within an event history framework where the individual's rate of adoption is a function of prior adoptions by related actors. Two diffusion models are suggested: an epidemic model where adoption rates vary with the number of prior adoptions, and a salience model where adoption rates vary with time since the last event. This approach is illustrated in an examination of the decolonization of British and French colonies. Diffusion is shown to occur within regions rather than within empires or the world system as a whole.

The Analysis of Factorial Surveys

Abstract: Factorial surveys constitute a specific technique for introducing experimental designs in sample surveys. Respondents are presented with descriptions (vignettes) of a constructed world in which important factors are built in experimentally. Using balanced designs well known from the multivariate experimental tradition, it is possible to build in a relatively large number of factors and levels. Within this context, the normal hypothesis is that responses are consistent on the individual level, but not totally idiosyncratic. In the analysis, it is important to determine the influence of both the vignette and the respondent variables. Analysis models for this type of data should reflect the fact that factorial surveys produce data pertaining to two distinct levels: the individual level and the vignette level. Such models are available and are generally known as multilevel analysis models. This article discusses the properties of factorial survey designs and some analysis models that address the multilevel as...

The Statistical Analysis of Event Histories

Abstract: Event histories are generated so-called failure-time processes and take this form. The dependent variable—for example, some social state—is discrete or continuous. Over time it evolves as follows: For finite periods of time (from one calendar date to another) it stays constant at a given value. At a later date, which is a random variable, the dependent variable jumps to a new value. The process evolves in this manner from the calendar date when one change occurs to a later date when another change occurs. Between the dates of the changes, the dependent variable stays constant. Data on such processes typically contain information about (a) the date a sample member entered a social state; (b) the date the state later was left, if left; (c) the value of the next state entered; and so on. In analyzing such data, the foci are on what determines the amount of time spent in each state and on what determines the value of the next state entered. This article describes how one can use continuous-time hazard rate mo...

Acquiescence and Recency Response-Order Effects in Interview Surveys

Abstract: Split-ballot experiments for three sets of items measuring attitudes towards lawyers, anomia, and self-esteem were included in a telephone survey to test for both acquiescence and response-order effects. The experimental design also investigated whether these response effects would be reduced by giving respondents an explicit opportunity to say “don't know” (a filtered question form). Extensive evidence for both acquiescence and recency response-order effects was found. These response effects also often occurred for the same item. Thus the use of a forced-choice form to avoid acquiescence to agree-disagree items may often substitute one type of response effect (recency order effects) for another (agreeing-response bias). Furthermore, there was very little evidence that the use of a filtered-question form would reduce these response effects. In addition to these practical conclusions, the patterns of these response effects across the three different types of attitudes, as well as their relationships to edu...

A Critical Evaluation of the Randomized Response Method Applications, Validation, and Research Agenda

Abstract: Warner (1965) introduced the Randomized Response Method (RRM) 20 years ago. This method has been advocated as a useful tool in eliciting sensitive information. Much of the early research focused on various parameter estimation methods. This article provides a comprehensive review of applications of the randomized response method with emphasis on recent publications, identifies issues now being studied, and suggests future research directions. As such, the article addresses method validation, respondent jeopardy, and new applications. In so doing, the method is considered in the broad context of obtaining responses to sensitive questions. Judging by the number of articles published, randomized response method continues to be of interest in numerous and diverse disciplines.

The Microdemographic Community-Study Approach

Abstract: Survey methods have been criticized for producing unreliable, invalid data and for failing to provide contextual information to test complex causal hypotheses. We discuss a technique that combines survey and ethnographic methods at every stage of the data collection process to overcome these shortcomings. We use ethnographic and survey evidence to show how the combined approach reduces coverage errors, nonresponse errors and measurement errors arising from the interviewer, the questionnaire, and the respondent. Complete integration of the two methods during data collection can uncover information that a survey alone would have missed. Ethnographic data can also be used to understand the meaning behind relationships among survey variables that would have otherwise been unclear. Finally, although the combined approach is intensive, it is flexible enough to be used in a variety of settings to study many different research questions.

The Microdemographic Community-Study Approach: Improving Survey Data by Integrating the Ethnographic Method

Abstract: Survey methods have been criticized for producing unreliable, invalid data and for failing to provide contextual information to test complex causal hypotheses We discuss a technique that combines

Estimating Nonresponse and Response Bias: Resident and Nonresident Parents' Reports about Child Support

Abstract: Although researchers are aware that nonresponse and response bias may compromise the accuracy of estimates from survey data, it is difficult to obtain estimates of these biases. In this article, we estimate nonresponse and response bias for a particular case—child support awards and payments. The sample of divorced resident and nonresident parents was drawn from Wisconsin court records and subsequently interviewed by telephone in 1987. The court records provide the criterion to use in estimating nonresponse and response bias. The analysis shows that those not interviewed are less likely to have awards or to pay support than participants, and the average award and average amount paid are lower for them than for survey participants. Both resident mothers and nonresident fathers overreport the amount of support owed and paid, but the bias is larger for fathers. Nonparticipation bias is greater for measures of paying support than for measures of owing support. Response bias contributes more to total nonsampli...

Research on Survey Quality

Abstract: This article provides a review of recent research on the quality of survey data and serves as an introduction to this special issue of SMR. The article reviews standard statistical treatments of su...

Reducing Bias in Estimates of Linear Models by Remeasurement of a Random Subsample

Abstract: 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 tried 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.

Assessing bias and fit of global and local hazard models.

Abstract: This paper assesses the bias and fit of a hazard rate model by comparing predicted and nonparametric estimates of survivor probabilities. These comparisons also generate several diagnostic displays for event history data. We illustrate these diagnostic methods for several global and local hazard models using data on age at first marriage for women in the June 1980 [U.S.] Current Population Survey. Our results suggest that a nonproportional local Gompertz model performs better than other models we examine. (EXCERPT)

On the Treatment of Interrupted Spells and Initial Conditions in Event History Analysis

Abstract: This article deals with some problems of event history analysis that arise in special types of sampling plans where spells in progress at time t0, 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 t0 and the duration of the spell after t0 have to account for the condition that the individual is in the state y0 in question at time t0. 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 y0 at time t0), in particular the joint distribution of pre-interview and post-interview durations. In general, this distribution depends on the entry rate into the state y0. The distribution of the sampled spells for the special sampling plans differs from the cor...

The Measurement of Parental Influence: Assessing the Relative Effects of Father and Mother

Abstract: The influence that parents have on the social, behavioral, and psychological status of their children is the focus of a large and important body of literature in the social sciences. The data collected to study parental influence often describe individual mothers and fathers, leaving the researcher with the problem of representing the joint influence of these individuals as parents. A diagonal ANOVA model is proposed to represent parental influence as the weighted effects of one variable that describes father and one that describes mother. Extensions of the diagonal model that facilitate tests of various hypotheses pertaining to variation in the relative influence of mother and father are proposed. Models that represent the effects of asymmetry of parental characteristics are discussed as well. This application of the diagonal ANOVA model is illustrated using data on the effects of social bonds to mother and to father as a deterrent to the delinquent behavior of black and white male adolescents.

Autocorrelation and Density Dependence in Organizational Founding Rates

Abstract: This article assesses the robustness of recent estimates of the effect of density on the founding rates in organizational populations. It reports reanalyses of data on founding rates of six populations of organizations using a generalization of quasi-likelihood estimation that allows specification of autocorrelation processes. Autocorrelation is indeed present in five of the six data sets. However, the main substantive finding of earlier research proves to be robust-a nonmonotonic relationship between density and founding rate-continues to hold in most cases even when autocorrelation is taken into account. In other words, the predicted pattern of nonmonotonic density dependence is robust with respect to the form of autocorrelation investigated.

Approximations to the Mean and Variance of the Index of Dissimilarity in 2 x C Tables Under a Random Allocation Model

Abstract: Approximations to the mean and variance of the index of dissimilarity in 2 x C tables under a random allocation model are derived, using a multivariate normal approximation to the multivariate hypergeometric distribution for the cell frequencies and the appropriate multivariate folded-normal distribution for the components of the index. To illustrate one application of these approximations, we standardize values of the dissimilarity index computed from tract-level census data to compare the levels of racial segregation in Birmingham and Richmond in 1970 and 1980, following Cortese, Falk, and Cohen (1976).

A Hierarchy for Some Latent Structure Models

Abstract: In 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 models and propose strategies for model selection, demonstrating those strategies in several examples. We also discuss the usefulness of the various models.

The Analysis of Mobility Regimes: Implementation Using SAS Procedures and an Australian Case Study

Abstract: A topological model of a mobility regime is a mapping of the cells in a two-way cross-classification of data into a number of coherent sets of cells or density levels. Such density levels identify cells in the table that are characterized by similar tendencies to association or dissociation between pairs of origin and destination categories. This article demonstrates how to automate categorical model fitting of such data within the SAS statistical package. We discuss two substantive examples: U.S. male intergenerational occupational mobility: and a new account of intergenerational class mobility in Australia, that draws on contemporary neo-Marxist theorizing about class structures in capitalist societies.