Lasso Regularization for Selection of Log-linear Models: An Application to Educational Assortative Mating

Bucca, Mauricio; Urbina, Daniela

doi:10.1177/0049124119826154

Cited by 7 publications

(5 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this context, I turn to Least Absolute Shrinkage and Selection Operator (lasso) logistic regressions. Perhaps the most well-established SML method among social scientists (Bucca and Urbina 2021), lasso regressions offer two key advantages over more complex SML approaches (e.g., neural nets or classification trees). First, lasso regressions still provide regression coefficients for all selected predictors, providing a face-validity test that is not possible with more complex, black-box approaches.…”

Section: Explaining Lgb Outcomes: a “Gender Predictive” Approachmentioning

confidence: 99%

“…Second, lasso regressions are easily implemented in all statistical software programs and provide a natural extension of the regression methods already used in sociological research. Part of a class of functions called penalized regressions, lasso regressions balance overfitting and underfitting by iteratively drawing on a large number of variables while shrinking their coefficients toward zero (for a more complete overview, see Bucca and Urbina 2021).…”

Section: Explaining Lgb Outcomes: a “Gender Predictive” Approachmentioning

confidence: 99%

See 1 more Smart Citation

Intersecting the Academic Gender Gap: The Education of Lesbian, Gay, and Bisexual America

Mittleman

2022

Am Sociol Rev

View full text Add to dashboard Cite

Although gender is central to contemporary accounts of educational stratification, sexuality has been largely invisible as a population-level axis of academic inequality. Taking advantage of major recent data expansions, the current study establishes sexuality as a core dimension of educational stratification in the United States. First, I analyze lesbian, gay, and bisexual (LGB) adults’ college completion rates: overall, by race/ethnicity, and by birth cohort. Then, using new data from the High School Longitudinal Survey of 2009, I analyze LGB students’ performance on a full range of achievement and attainment measures. Across analyses, I reveal two demographic facts. First, women’s rising academic advantages are largely confined to straight women: although lesbian women historically outpaced straight women, in contemporary cohorts, lesbian and bisexual women face significant academic disadvantages. Second, boys’ well-documented underperformance obscures one group with remarkably high levels of school success: gay boys. Given these facts, I propose that marginalization from hegemonic gender norms has important—but asymmetric—impacts on men’s and women’s academic success. To illustrate this point, I apply what I call a “gender predictive” approach, using supervised machine learning methods to uncover patterns of inequality otherwise obscured by the binary sex/gender measures typically available in population research.

show abstract

Section: Explaining Lgb Outcomes: a “Gender Predictive” Approachmentioning

confidence: 99%

Section: Explaining Lgb Outcomes: a “Gender Predictive” Approachmentioning

confidence: 99%

Intersecting the Academic Gender Gap: The Education of Lesbian, Gay, and Bisexual America

Mittleman

2022

Am Sociol Rev

View full text Add to dashboard Cite

show abstract

“…For example, methodological adaptations of a wide variety of ML methods have appeared in life course research (Billari, Fürnkranz, and Prskawetz 2006), criminology (Baćak and Kennedy 2019), survey design (Fu, Guo, and Land 2018), text and image analysis (Rona-Tas et al 2019;Zhang and Pan 2019), causal inference (Brand et al 2021;Liu 2021;Torrats-Espinosa 2021), log-linear modelling (Bucca and Urbina 2019), and general evaluation of the predictability of life and physical outcomes (Daoud, Kim, and Subramanian 2019;Salganik et al 2020).…”

Section: Introductionmentioning

confidence: 99%

Leveraging Machine Learning to Estimate Effect Modification

Yu¹,

Park²,

Kang³

et al. 2021

Preprint

View full text Add to dashboard Cite

Sociologists are often interested in estimating and testing whether some causal effect varies by a modifier of interest. The conventional regression estimator for effect modification is inflexible in functional form and prone to misspecification bias. Machine Learning (ML) algorithms can aid the estimation of effect modification in observational studies by controlling for confounders in a highly flexible, automated, yet principled way. Therefore, leveraging ML for effect modification helps reduce misspecification bias and enhance the credibility of causal identification. We introduce a novel estimator that estimates effect modification in a familiar regression framework after using ML algorithms to fit nuisance components of the model. We show that this estimator is more flexible than the conventional regression model while more efficient and suitable for theory-driven sociological research than other ML-based methods. We use the new estimator to study the modification in the effect of a college degree on adult family income by gender and family income in adolescence in the United States. Along these two dimensions, the benefits of a college degree are rather equally distributed.

show abstract

“… 5. By following a set inferential logic (namely that of the original model), complexity in the proposed framework is introduced solely through the functional relationship among variables and distinguishes itself from “model complexity” as is often understood as the problem of selecting among a (large) set of potential covariates (Bucca and Urbina 2019). …”

mentioning

confidence: 99%

“…Given these variables, though, we are interested in a model that fits the data well. Note that a different approach to ML methods is to identify relevant variables among a (large) set of possible covariates (Bucca and Urbina 2019; Grimmer et al 2021). This is another fruitful application of ML into sociological model-building, but outside the scope of this article.…”

mentioning

confidence: 99%

Incorporating Machine Learning into Sociological Model-Building

Verhagen

2024

Sociological Methodology

View full text Add to dashboard Cite

Quantitative sociologists frequently use simple linear functional forms to estimate associations among variables. However, there is little guidance on whether such simple functional forms correctly reflect the underlying data-generating process. Incorrect model specification can lead to misspecification bias, and a lack of scrutiny of functional forms fosters interference of researcher degrees of freedom in sociological work. In this article, I propose a framework that uses flexible machine learning (ML) methods to provide an indication of the fit potential in a dataset containing the exact same covariates as a researcher’s hypothesized model. When this ML-based fit potential strongly outperforms the researcher’s self-hypothesized functional form, it implies a lack of complexity in the latter. Advances in the field of explainable AI, like the increasingly popular Shapley values, can be used to generate understanding into the ML model such that the researcher’s original functional form can be improved accordingly. The proposed framework aims to use ML beyond solely predictive questions, helping sociologists exploit the potential of ML to identify intricate patterns in data to specify better-fitting, interpretable models. I illustrate the proposed framework using a simulation and real-world examples.

show abstract

Lasso Regularization for Selection of Log-linear Models: An Application to Educational Assortative Mating

Cited by 7 publications

References 66 publications

Intersecting the Academic Gender Gap: The Education of Lesbian, Gay, and Bisexual America

Intersecting the Academic Gender Gap: The Education of Lesbian, Gay, and Bisexual America

Leveraging Machine Learning to Estimate Effect Modification

Incorporating Machine Learning into Sociological Model-Building

Contact Info

Product

Resources

About