Blog Entry

Regression effect

[...] A much more subtle flaw is technically termed a regression effect. That is, processes appear to "regress" from less likely states to more likely ones simply because the more likely ones are likely to occur at later points in time. For example, people are not often extremely happy (or extremely unhappy). It follows that when they are, they are less likely to be as extremely unhappy (or happy) later--no matter what happens in the meantime. Because most people enter therapy when they are extremely unhappy, they are less likely to be as unhappy later, independent of the effects of therapy itself. Hence, this "regression effect" can create the illusion that the therapy has helped to alleviate their unhappiness, whether it has or not. [p.44]

[...]

The direct relevance of regression effects to evaluating psychotherapy is that people often enter therapy at times when they are particularly unhappy and distressed. But if their problem is one that varies over time rather than having a consistently downward course, regression effects alone could result in "improvement"--and an illusion that the improvement is due to psychotherapy: "If treated, a cold will go away in seven days, whereas if left alone, it will last a week." [p.45]

House of cards: psychology and psychotherapy build on myth
Robyn Dawes

Posted on:
2008.09.13 -0500

Tags:
texts