A meta-analysis of the effects of meditation and hypnosis on measures of anxiety

Abstract

This meta-analysis was conducted to determine the effects of meditation and hypnosis techniques on psychometric measures of anxiety. The chief measure employed in the evaluated research was the State-Trait Anxiety Inventory (Spielberger, 1970; 1983). An extensive literature search identified several hundred related studies. A combination of the desired search terms narrowed this to 109 highly related studies. The final analyzed group included 21 hypnosis studies and 54 meditation studies. Effect sizes were then calculated for each included study, employing a pooled estimate of the population standard deviation (Hedges and Olkin, 1985). Where appropriate, statistical comparisons of effect sizes were undertaken. Both techniques were effective in reducing measures of state anxiety (Hypnosis = -.858; Meditation = -.745). However, for measures of trait anxiety, meditation was more effective (Meditation = -.686; Hypnosis = -.347). This difference was statistically significant. Data concerning the impact of practice time indicated that practice length influenced hypnosis effect sizes. A moderate amount of practice (three to six weeks) was associated with larger effect sizes. No significant differences were found for meditation studies. This study also indicated that journal source articles produced significantly larger hypnosis effect sizes. Meditation effect sizes did not significantly differ on the basis publication source. Data from this project found non-significant differences in effect sizes for the different hypnosis techniques. For meditation studies, Transcendental Meditation (TM) was somewhat more effective than other techniques. However, the only statistically significant difference was between TM and Relaxation Response trait anxiety effect sizes. Finally, results also suggested that the level of subject-instructor contact did not significantly influence hypnosis state anxiety effect sizes. Hypnosis trait anxiety effect sizes did, however, significantly differ in favor of moderate levels of contact. Effect sizes did not significantly differ on the basis of contact level for meditation studies.