3-238 - New (or Under Appreciated) Statistical Methods for Probing Causal Hypotheses in Developmental Research

Session Type: Paper Symposium

Integrative Statement

Causal inference is a matter of central significance to the field of child development. There are, however, many obstacles to conducting studies with internal validity for developmental scientists whose research extends beyond tightly controlled laboratory studies into real-world settings. Even randomized field experiments often involve complex interpretative challenges beyond the average treatment effect. In this symposium, four papers will be presented that offer new (or under appreciated) statistical methods that developmental scientists can easily apply when the causes of individual differences in development are of primary concern. The first paper provides a primer on sensitivity analysis and its value for examining the robustness of results from correlational studies of child development in which potential omitted variable bias is a ubiquitous concern. The second paper addresses the value of Marginal Structural Models for controlling time-varying confounds in longitudinal studies. The third paper compares two methods that improve upon instrumental variable techniques. And, the fourth paper …heterogeneous treatment effects. For each of the papers, presenters will approach the topics from an applied perspective, focusing on conceptual issues, empirical examples, and resources for implementing these methods (e.g., software and syntax).

Sub Unit

• Methods, History, Theory

Chair

• Henrik Daae Zachrisson, University of Oslo

Individual Presentations

• Taking Selection and Sensitivity (to Confounders) Seriously in Correlational Studies of Child Development - Presenting Author: Eric Dearing, Boston College; Non-Presenting Author: Henrik Daae Zachrisson, University of Oslo

• Weight, Balance, Repeat: Marginal Structural Models for Improving Causal Inference With Time-Varying Treatments - Presenting Author: Dan Berry, University of Minnesota; Non-Presenting Author: Clancy Blair, New York University; Non-Presenting Author: W. Roger Mills-Koonce, University of North Carolina at Chapel Hill

• Addressing Endogeneity Selection Bias in a Non-Experimental Study of Academic Achievement - Presenting Author: Jordan Lawson, Boston College; Non-Presenting Author: Laura O'Dwyer, Boston College; Non-Presenting Author: Mary Walsh, Boston College

• Estimating Treatment Heterogeneity in Early Childhood Contexts: Lessons Learned and Implications for Study Design - Presenting Author: Kathryn Gonzalez, Harvard University; Non-Presenting Author: Terri Sabol, Northwestern University; Non-Presenting Author: Dana McCoy, Harvard Graduate School of Education; Non-Presenting Author: Luke Miratrix, Harvard University; Non-Presenting Author: Jessaca Spybrook, Western Michigan University