Abstract
All experimenters know that human and animal subjects do not respond uniformly to experimental treatments. Yet theories and findings in experimental psychology either ignore this causal effect heterogeneity or treat it as uninteresting error. This is the case even when data are available to examine effect heterogeneity directly, in within-subjects designs where experimental effects can be examined subject by subject. Using data from four repeated-measures experiments, we show that effect heterogeneity can be modeled readily, that its discovery presents exciting opportunities for theory and methods, and that allowing for it in study designs is good research practice. This evidence suggests that experimenters should work from the assumption that causal effects are heterogeneous. Such a working assumption will be of particular benefit, given the increasing diversity of subject populations in psychology.
Original language | English |
---|---|
Pages (from-to) | 601-618 |
Number of pages | 18 |
Journal | Journal of Experimental Psychology: General |
Volume | 148 |
Issue number | 4 |
DOIs | |
State | Published - Apr 2019 |
Bibliographical note
Publisher Copyright:© 2019 American Psychological Association.
Keywords
- Causal effect heterogeneity
- Causal processes
- Mixed models
- Repeated measures
- Theory development