Instructor: Guanglei Hong
Causal mediation analysis is suitable for investigating WHY an intervention generates or fails to generate an impact and thereby explicitly tests the intervention theory. The evidence may suggest whether there is a need to modify the intervention theory or to simply enhance its implementation. This issue is critical for evaluating new STEM curricula such as the one described in the stylized case. A randomized experiment guarantees unbiased estimates of the treatment effects on the outcome and on the mediators but not of the effect of a mediator on the outcome. This is because the latter is often confounded by selection bias. Further complications arise when the mediator-outcome relationship differs across the treatment conditions. For example, a certain instructional strategy may show great benefit only under the experimental condition that trains teachers to take advantage of its benefit; the same strategy may not be as effective under the control condition. Conventional statistical methods most familiar to STEM education researchers tend to be inadequate for studying causal mechanisms because they typically overlook the threats to causal claims. The goal of this course is to equip students with basic knowledge and analytic skills for conducting causal mediation analysis. The course will introduce cutting-edge methodological concepts (Pearl, 2001) and analytic approaches and will contrast them with conventional strategies including multiple regression, path analysis, and structural equation modeling. We will examine several innovative strategies for causal mediation analysis (Imai, Keele, & Tingley, 2010; Valeri & VanderWeele, 2013) and will highlight a new weighting-based strategy with application to STEM education research (Hong, 2010; Hong & Nomi, 2012; Qin & Hong, 2017; Qin et al, 2019, in press). The latter has been endorsed as “a simple unified approach” for causal mediation analysis (Lange, Vansteelandt, & Bekaert, 2012). Last but not least, the course will conclude with a discussion of sensitivity analysis for assessing the potential impact of hidden bias (Hong, Qin, & Yang, 2018). The textbook “Causality in a Social World: Moderation, Mediation, and Spill-Over” (Hong, 2015) will be supplemented with other readings reflecting latest developments and controversies. The course content is organized around application examples, many of which are drawn from STEM education research. Instructional format will include lectures, small group and whole class discussions, and exercises for assessing conceptual understanding and analytic skills. Fellows will practice applying the causal concepts and methods to their own data.
REFERENCES:
Hong, G. (2010). Ratio of mediator probability weighting for estimating natural direct and indirect effects. In JSM Proceedings, Biometrics Section. Alexandria, VA: American Statistical Association, pp.2401-2415.
Hong, G. (2015). Causality in a social world: Moderation, mediation, and spill-over. West Sussex, UK: John Wiley & Sons, Inc.
Hong, G., & Nomi, T. (2012). Weighting methods for assessing policy effects mediated by peer change. Journal of Research on Educational Effectiveness special issue on the statistical approaches to studying mediator effects in education research, 5(3), 261-289.
Hong, G., Qin, X., & Yang, F. (2018). Weighting-based sensitivity analysis in causal mediation studies. Journal of Educational and Behavioral Statistics, 43(1), 32-56.
Imai, K., Keele, L., & Tingley, D. (2010). A general approach to causal mediation analysis. Psychological methods, 15(4), 309.
Lange, T., Vansteelandt, S., & Bekaert, M. (2012). A simple unified approach for estimating natural direct and indirect effects. American Journal of Epidemiology, 176(3), 190-195.
Pearl, J. (2001). Direct and indirect effects. In Proceedings of the Seventeenth Conference on Uncertainty in Artificial Intelligence. Morgan Kaufmann, San Francisco, CA, pp.411–420.
Qin, X., & Hong, G. (2017). A weighting method for assessing between-site heterogeneity in causal mediation mechanism. Journal of Educational and Behavioral Statistics, 42(3), 308-340.
Qin, X., Hong, G., Deutsch, J., & Bein, E. (2019). Multisite causal mediation analysis in the presence of complex sample and survey designs and non-random nonresponse. Journal of the Royal Statistical Society, Series A, 182, Part 4, 1343-1370.
Qin, X., Deutsch, J., Hong, G. (in press). Revealing heterogeneity in complex mediation mechanisms: Two concurrent mediators. Journal of Policy Analysis and Management.
Valeri, L., & VanderWeele, T. J. (2013). Mediation analysis allowing for exposure-mediator interactions and causal interpretations: Theoretical assumptions and implementation with SAS and SPSS macros. Psychological Methods, 18, 137-150.