Is the Dodo bird endangered in the 21st century? A meta-analysis of treatment comparison studies.

The Dodo bird hypothesis asserts that when bona fide treatments are compared they yield similar outcomes and this hypothesis is consistent with a common factors or contextual model of psychotherapy. Wampold et al. (1997), the most recent comprehensive meta-analysis to test the Dodo bird hypothesis, yielded consistent evidence of treatment equivalence. However, some of Wampold et al.'s analytic strategies, such as using multiple effect sizes from the same study and prioritizing long-term follow-up, may have obscured treatment differences. The current meta-analysis updated Wampold et al. by analyzing studies published in the subsequent 16 years (k=51). Separate effect sizes were calculated for primary outcomes versus secondary outcomes, at termination and follow-up. Contrary to the Dodo bird hypothesis, there was evidence of treatment differences for primary outcomes at termination. Furthermore, cognitive-behavioral treatments may be incrementally more effective than alternative treatments for primary outcomes. Consistent with the Dodo bird hypothesis, there was little evidence of treatment differences for the secondary outcomes at termination and follow-up. There are small, statistically significant differences between bona-fide treatments when the specific targets of those treatments are assessed, but not when secondary outcomes are assessed, providing mixed support for both specific factors and contextual models of psychotherapy.

The latent structure of generalized anxiety disorder in midlife adults.

Generalized anxiety disorder (GAD) is identified as a discrete disorder in the DSM-5, but evidence suggests that GAD and the related construct of pathological worry possesses a dimensional latent structure. The objective of this study was to ascertain the latent structure of GAD using taxometric methods. A subsample of adults (N=2061) from the Midlife in the United States Study, a national sample of Americans, provided the data. Additional data from individuals who were re-interviewed 10 year later (n=1228) were also analyzed. Items corresponding to the DSM-IV-TR diagnostic criteria for GAD were used to generate indicators for the taxometric analyses. Multiple taxometric procedures provided no evidence that GAD has a categorical or taxonic latent structure. Instead, the results were more consistent with the proposition that GAD exists on a continuum. Evidence that GAD is dimensional suggests that dichotomizing individuals into GAD versus non-GAD groups will typically result in decreased statistical power. They also suggest that any diagnostic thresholds for identifying GAD are likely to be arbitrary. The findings are consistent with models that locate GAD within the framework of extant dimensional models of personality and with research that emphasizes a multifactorial etiology for GAD.