Premium
Distinguishing Mediational Models and Analyses in Clinical Psychology: Atemporal Associations Do Not Imply Causation
Author(s) -
Winer E. Samuel,
Cervone Daniel,
Bryant Jessica,
McKinney Cliff,
Liu Richard T.,
Nadorff Michael R.
Publication year - 2016
Publication title -
journal of clinical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.124
H-Index - 119
eISSN - 1097-4679
pISSN - 0021-9762
DOI - 10.1002/jclp.22298
Subject(s) - mediation , psychology , causality (physics) , causation , heuristic , cognitive psychology , social psychology , epistemology , social science , philosophy , physics , quantum mechanics , sociology
Purpose A popular way to attempt to discern causality in clinical psychology is through mediation analysis. However, mediation analysis is sometimes applied to research questions in clinical psychology when inferring causality is impossible. This practice may soon increase with new, readily available, and easy‐to‐use statistical advances. Thus, we here provide a heuristic to remind clinical psychological scientists of the assumptions of mediation analyses. Approach We describe recent statistical advances and unpack assumptions of causality in mediation, underscoring the importance of time in understanding mediational hypotheses and analyses in clinical psychology. Example analyses demonstrate that statistical mediation can occur despite theoretical mediation being improbable. Conclusion We propose a delineation of mediational effects derived from cross‐sectional designs into the terms temporal and atemporal associations to emphasize time in conceptualizing process models in clinical psychology. The general implications for mediational hypotheses and the temporal frameworks from within which they may be drawn are discussed.