Sociological Methodology

This paper advocates an actor-centered, relational view of agency and proposes Quantitative Narrative Analysis (QNA) as a promising method for operationalizing and measuring agency. QNA organizes the information contained in narrative texts by exploiting the invariant linguistic structural properties of narrative—namely, sets of SVOs (Subject, Verb, Object) organized in predictable sequences and where in narrative S are actors and V are actions. The relational data made available by QNA are ideally suited for analysis with geographic information systems (GIS) tools, sequence analysis, or network analysis. These tools preserve the centrality of agency (actors and their actions) in social scientific explanation of social reality. An application of QNA to newspaper stories of lynchings in Georgia (1875–1930) will illustrate the power of this approach. The paper complements the illustration of this quantitative way of measuring agency with discourse analysis—another popular social science approach to texts. We will rely on this approach to illustrate how linguistic and rhetorical strategies can be used to hide agency in texts and the challenges (and solutions) this poses for measurement: How can we measure something that is not there?

Although there has been much optimistic discussion of integrating quantitative and qualitative findings into sociological analysis, there remains a gap regarding the application of mixed approaches. We examine the potential gains and pitfalls of such integration in the context of the growing analytic power of contemporary qualitative data analysis software (QDAS) programs. We illustrate the issues with our own research in a mixed-methods project examining low fertility in Italy, a project that combines analysis of large nationally representative survey data with qualitative in-depth interviews with women across four cities in Italy. Despite the enthusiasm for mixed-methods research, the available software appears to be underutilized. In addition, we suggest that the sociological research community will want to address several conceptual and inferential issues with these approaches.

This paper reviews methods for analyzing both individual preferences and choices about where to live, and the implications of these choices for residential patterns. Although these methods are discussed in the context of residential choice, they can be applied more broadly to individual choices in a range of social contexts where behavior is interdependent. We also discuss specific problems with residential mobility data, including the treatment of one’s current location as a potential choice, the aggregation of units and the need to take into account variations in neighborhood size, the problem of very large choice sets, and the link between residential mobility and patterns of neighborhood change.

This article presents a new simulation method to make global network inference from sampled data. The proposed method takes sampled ego network data and uses exponential random graph models (ERGM) to reconstruct the features of the true, unknown network. After describing the method, the author presents two validity checks of the approach: the first uses the 20 largest Add Health networks while the second uses the Sociology Coauthorship network in the 1990s. For each test, I take random ego network samples from the known networks and use my method to make global network inference. The method successfully reproduces the properties of the networks, such as distance and main component size. The results also suggest that simpler, baseline models provide considerably worse estimates for most network properties. The paper concludes with a discussion of the bounds/limitations of ego network sampling as well as possible extensions to the proposed approach.

Techniques for sampling from networks have grown into an important area of research across several fields. For sociologists, the possibility of sampling from a network is appealing for two reasons: (1) A network sample can yield substantively interesting data about network structures and social interactions, and (2) it is useful in situations in which study populations are difficult or impossible to survey with traditional sampling approaches because of the lack of a sampling frame. Despite its appeal, methodological concerns about the precision and accuracy of network-based sampling methods remain. In particular, recent research has shown that sampling from a network using a random walk–based approach such as respondent-driven sampling (RDS) can result in a high design effect (DE): the ratio of the sampling variance to the sampling variance of simple random sampling (SRS). A high DE means that more cases must be collected to achieve the same level of precision as SRS. In this article, we propose an alternative strategy, network sampling with memory (NSM), which collects network data from respondents to reduce DEs and, correspondingly, the number of interviews needed to achieve a given level of statistical power. NSM combines a “list” mode, in which all individuals on the revealed network list are sampled with the same cumulative probability, with a “search” mode, which gives priority to bridge nodes connecting the current sample to unexplored parts of the network. We test the relative efficiency of NSM compared with RDS and SRS on 162 school and university networks from the National Longitudinal Study of Adolescent Health and Facebook that range in size from 110 to 16,278 nodes. The results show that the average DE for NSM on these 162 networks is 1.16, which is very close to the efficiency of a simple random sample (DE = 1) and 98.5 percent lower than the average DE we observed for RDS.

Although log-linear and log-multiplicative analysis of association between variables is useful, its major limitation has been a lack of association with causal analysis. The objective of this article is to introduce a causal analysis on the basis of an adjusted cross-classified frequency table having such counterfactual odds ratios between a treatment variable X and its dependent variable Y that would be realized if the effects of confounding covariates on the association between X and Y were eliminated. The author first clarifies that the attainment of statistical independence between X and the covariates is not sufficient, because the issue of collapsibility remains even under completely randomized treatment allocation, and then proposes a specific standardization method for covariate states to solve this issue. By assuming semiparametric regression models, the author then introduces a new method of retaining the causal effects of X on Y, and the possible dependence of those effects on the third variable, in adjusted two-way or three-way contingency tables. As a result, the author also introduces a method for applying semiparametric logit and multinomial logit regression models that do not specify the effects of the confounding covariates V on Y. The estimation of standard errors for applying log-linear and log-multiplicative association models for the data of adjusted contingency tables is also described.

Logit and probit models are widely used in empirical sociological research. However, the common practice of comparing the coefficients of a given variable across differently specified models fitted to the same sample does not warrant the same interpretation in logits and probits as in linear regression. Unlike linear models, the change in the coefficient of the variable of interest cannot be straightforwardly attributed to the inclusion of confounding variables. The reason for this is that the variance of the underlying latent variable is not identified and will differ between models. We refer to this as the problem of rescaling. We propose a solution that allows researchers to assess the influence of confounding relative to the influence of rescaling, and we develop a test to assess the statistical significance of confounding. A further problem in making comparisons is that, in most cases, the error distribution, and not just its variance, will differ across models. Monte Carlo analyses indicate that other methods that have been proposed for dealing with the rescaling problem can lead to mistaken inferences if the error distributions are very different. In contrast, in all scenarios studied, our approach performs as least as well as, and in some cases better than, others when faced with differences in the error distributions. We present an example of our method using data from the National Education Longitudinal Study.

Individuals differ not only in their background characteristics but also in how they respond to a particular treatment, intervention, or stimulation. In particular, treatment effects may vary systematically by the propensity for treatment. In this paper, we discuss a practical approach to studying heterogeneous treatment effects as a function of the treatment propensity, under the same assumption commonly underlying regression analysis: ignorability. We describe one parametric method and two nonparametric methods for estimating interactions between treatment and the propensity for treatment. For the first method, we begin by estimating propensity scores for the probability of treatment given a set of observed covariates for each unit and construct balanced propensity score strata; we then estimate propensity score stratum-specific average treatment effects and evaluate a trend across them. For the second method, we match control units to treated units based on the propensity score and transform the data into treatment-control comparisons at the most elementary level at which such comparisons can be constructed; we then estimate treatment effects as a function of the propensity score by fitting a nonparametric model as a smoothing device. For the third method, we first estimate nonparametric regressions of the outcome variable as a function of the propensity score separately for treated units and for control units and then take the difference between the two nonparametric regressions. We illustrate the application of these methods with an empirical example of the effects of college attendance on women’s fertility.

Although group interaction is a rich source of data, it offers many challenges for analysis. We discuss how pretesting can solve many coding and reliability issues before the actual data collection and analysis. We also consider how coding schemes can be developed, how coders can be trained, and how coder reliability can be ascertained. We examine the properties of different estimates of intercoder reliability and detail how these relate to decisions about the data.

The author presents a nonparametric approach to measuring stratification that highlights the distinction between stratification and inequality. Using pairwise comparison of ranks, the author develops an index of stratification that gauges the overall degree to which population subgroups occupy distinct strata with respect to a hierarchical outcome. This new index possesses a number of desirable properties that are not satisfied by existing measures of stratification. The overall index can be decomposed as a weighted average of pair-specific indices of stratification, which capture the extent of separation between any two particular groups. Besides, this index can be easily extended to measure conditional stratification through control of a third variable. In addition, the author builds a parallel between stratification and inequality in their measurement by developing a general formula of which the index of stratification and the Gini index of inequality can be considered as two special cases. Finally, this new approach is applied to depict the temporal trends of wage stratification by gender, race, and educational attainment over the past three decades in the United States.