Mediation

This week's class focuses on using regression to test mediation. More generally, we discuss the issue of controlling for alternative effects and spurious relationships. In this context you might find the hemline index and the Superbowl hypothesis interesting. 

Primary Readings

• Hayes, A. F. (2013).  Introduction to mediation, moderation, and conditional process analysis. New York: The Guilford Press. Chapter 1 [pdf]

  • Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of personality and social psychology, 51(6), 1173. [pdf]

    • Kenny handout on testing mediation. [html]

• Zhao, Lynch, and Chen (JCR 2010) -- Reconsidering Baron and Kenny: Myths and Truths About Mediation Analysis [^pdf]

• Spurious correlation

Secondary Readings

• Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879-891

• Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408-420

• Simpson's paradox (aka ecological fallacy)

Slides

• mediation.pdf

Stata Do Files

• Mediation files

Datasets

• • Neogen -- create perf, knowjob and iadvice variables from longer-named originals

• GSS 2008

  • hsbdemo --  use https://stats.idre.ucla.edu/stat/data/hsbdemo, clear

Software tools for SPSS

• PROCESS macro for SPSS and SAS

  • • Download latest from http://www.processmacro.org or use this (possibly old) zip file (zip)

    • Documentation:

      • Brief handout [html]

    • Hayes, A. F. (2013).  Introduction to mediation, moderation, and conditional process analysis. New York: The Guilford Press.  [web page]

• After installing PROCESS, we will be working with the gss2008 dataset. A few commands we will execute:

• • compute revhappy = 4 - happy

    • filter the cases to retain only sibs < 20 and income > 0 and income < 13

  • We will also be working with the protest dataset. A few commands we will execute:

  • process vars = protest liking respappr /y=liking/x=protest/m=respappr/total=1/model=4/boot=5000/effsize=1/normal=1.

  • process vars=protest liking sexism/y=liking/x=protest/m=sexism/model=1/jn=1/quantile=1/plot=1.

PROCESS commands

• • Open GSS dataset

    • process vars= prestg80 rtheism rself/y=prestg80/x=rtheism/m=rself/model=4/total=1/effsize=1/boot=10000/normal=1.

    • process vars= sex childs polviews prestg80/y=polviews/x=sex/m=prestg80 childs/model=4/total=1/effsize=1/boot=10000/normal=1

  • Open Neogen dataset

    • create perf and iadvice variables from longer-named originals

    • process vars= perf iadvice knowjob/y=perf/x=iadvice/m=knowjob/model=4/total=1/effsize=1/boot=10000/normal=1