Examination of interaction effects within a meta-analysis: An example with the Balanced Placebo Design
Meta-analysis offers a set of quantitative techniques that allow synthesizing the results of many types of research including surveys, correlation studies, experimental studies and regression analyses. With emphasis on broad generalizations, meta-analysts tend to overlook or to down play the importance of contingency-specifying interactions that in most situations have an inferential precedence over statements about main effects. ^ The Balanced Placebo Design is a general research paradigm that employs a 2 × 2 factorial design and controls for both expectancy set and pharmacologic action. Subjects are told that their drinks either contain an alcoholic or a nonalcoholic beverage and the beverage itself either contains alcohol or it does not. This results in four conditions and at low to moderate doses of alcohol, subjects cannot discriminate correctly the placebo and antiplacebo conditions. ^ The expectancy belief caused by the deception of the Balanced Placebo Design can be accentuated by the introduction of moderator variables. Moderator variables are external design variables that can affect the outcome of the experiment by creating interactions. In this dissertation, it was hypothesized that two major moderator variables would affect the deception value of the expectancy effect of the balanced placebo design–the setting of the experiment and the method of assessment of the dependent variable. A meta-analysis using sixty-two studies was conducted and effect sizes were calculated. Homogeneity testing was done on the effect sizes. Significant heterogeneity was found, indicating the presence of moderator variables. ^
"Examination of interaction effects within a meta-analysis: An example with the Balanced Placebo Design"
(January 1, 2008).
ETD Collection for Fordham University.