Reading focus
Syllabus: How Does Psychotherapy Produce Change?
Separate a mediator, a mechanism, a moderator and a treatment procedure.
Why you’re reading it. Specify the evidence still missing after a credible treatment effect or process association.
A question to consider. For the hypothesis that increased activity reduces depressed mood, design a sequence of measurements and one comparison that would challenge reverse causation.
Why this reading
Showing that a treatment helps leaves its mechanism open. Kazdin lays out the kinds of converging evidence needed to explain a change process and distinguishes that task from finding a statistical mediator.
Before you read
A mediator is an intervening variable that may statistically account for a treatment–outcome relation. A mechanism explains the process responsible for change. A moderator identifies circumstances under which an association or treatment effect differs.
Read for
- Examine temporal ordering, experimental manipulation and alternative explanations in the proposed evidential criteria.
- Distinguish removing a treatment procedure from intervening directly on a hypothesized mechanism.
Consider afterwards
- Design a measurement schedule that could detect whether a proposed mediator changes before symptoms.
- How could the same outcome arise through different mechanisms for different people?
Editorial context
This 2009 methodological argument favors accumulated evidence across designs. Its opening assessment of the field and examples from genetics, neuroscience and treatment research are historical, not a current consensus review. Mechanistic explanation need not be biological, and neither a significant mediation test nor temporal precedence alone supplies a complete causal account.