Here's a snip of the opening of the article justifying why they chose to go back and question the findings of the original meta-analysis:
(Rockmaster, PM me your email, I'll send you the PDF of the whole article)
Any evidence that there is an efficacious treatment for obesity
that results in well-maintained weight loss is cause for enthusiasm,
given that the prevalence of obesity is increasing
(Kuczmarski, Flegal, Campbell, & Johnson, 1994). However,
such evidence merits careful examination, given that long-term
weight loss is notoriously difficult to achieve (Brownell & Rodin,
1994) with all but surgical interventions.
Several factors led us to be skeptical of Kirsch et al's (1995)
conclusion. First, it is inconsistent with the conclusion of several
qualitative reviews of this topic (Cochrane, 1992; Mott & Roberts,
1979; Vanderlinden & Vandereycken, 1994; Wadden & Anderton,
1982).
Second, the reported weighted mean effect size (standardized mean difference; d) of 1.96 was surprisingly large, as were two individual effect sizes (d = 3.65 and d = 5.57). By way of contrast, Abelson reviewed published metaanalyses and found that "it is unusual for this measure [d] to be as big as 1.0, quite rare for it to be as big as 1.4, and extraordinary for it to be as big as 2.0" ( 1995, p. 89).
Third, the distribution of reported effect sizes appeared to be markedly heterogeneous both within and among studies.
Fourth, Kirsch et al. (1995, Figure 1 ) suggested absolutely
no weight regain after hypnosis even at 2-year follow-up. Such
perfectly horizontal lines are generally not even found in studies
of bariatric surgery. Furthermore, these 2-year follow-up data
were based on a single study ( Bolocofsky, Spinier, & Coulthard-
Morris, 1985 ). Among the remaining studies, follow-up lengths
were limited to 3 and 6 months.
Fifth, not all studies appeared to sample clinically obese clients. Specifically, the mean weight of participants in one study was 70 kg (Bolocofsky et al., 1985 ), which barely qualifies as obesity even among 25-year-old women. It is also noteworthy that 24% of clients in that study dropped out of treatment in the first 9 weeks, an especially high attrition rate for that brief time period. Equally problematic is the fact that participants in Goldstein's ( 1981 ) study were allowed to come for additional weekly treatment if desired, which might have confounded this study's effect size estimates.
Finally, several unpublished dissertations found that hypnosis was no more effective than cognitive-behavioral interventions in the treatment of obesity. These studies were not included in Kirsch et al.'s meta-analysis, but were reviewed elsewhere (Wadden & Anderton, 1982). For these reasons, we chose to re-meta-analyze the obesity studies meta-analyzed by Kirsch et al. ( 1995 ).
RSS
Twitter
Facebook
YouTube
Launched in August 2010, TheMetropreneur.com is a local online resource devoted to small business development and entrepreneurship. Its aim is to tell the stories of Central Ohio's business community, foster regional economic development and assist entrepreneurs with its resource-heavy focus.