Many studies used more than one hypnosis scale in the research, (e.g., Balthazard & Woody, 1992; Crawford, 1982; Crawford, Brown & Moon, 1993; Moretti et al., 1980; Spanos, et al., 1983). Multiple correlations from a single sample can bias the results of a meta-analysis, this is commonly known as the independence problem. Moreover, given the significant differences with the different types of susceptibility measures, and the problem with the independence of the observations, the remaining analyses were conducted separately by hypnotizability scales for the objective response estimates. To provide comparability with the objective response information the data from the subjective response estimate measures were analyzed separately by hypnosis scale as well.

Tables 6a and 6b shows the differences in mean sample weighted correlations between those studies that assessed the hypnotic susceptibility and TAS relationship in or out of a hypnotic context. For both the HGSHS and SHSS: C, measuring the TAS in a hypnotic setting was not significantly different from assessing the TAS out of the hypnotic context, for the HGSHS, t(57)=0.97, ns; for the SHSSC, t(24) = 0.83, ns.

Insert Table 6a Here
Insert Table 6b Here

Unlike the HGSHS and SHSS: C, both the CURSS: Objective, t (20) = 2.98, p < .01, and CURSS: Subjective, t (20) = 3.23, p < .01, showed statistically significant context effects. When the TAS was measured in a hypnotic context there was a significantly higher mean sample weighted correlation than when the hypnotic context was kept separate from the TAS testing.
The results with the "Other" category were contrary to what Council et al. (1986) would predict. Those studies that measured the TAS in a hypnotic context had a significantly lower mean sample weighted correlation than those that kept context separate from the Absorption testing, t (27) = 2.37, p < .05.