A meta-analysis of criterion effects for the Implicit Relational Assessment Procedure (IRAP) in the clinical domain. (September 2015)

• Record Type: Journal Article
• Authors: Vahey, Nigel A.; Nicholson, Emma; Barnes-Holmes, Dermot
• Is Part Of: Journal of behavior therapy and experimental psychiatry. Volume 48(2015)
• Page Start: 59
• Page End: 65
• Publication Date: 2015-09
• Subjects: IRAP -- Meta-analysis -- Criterion effects -- Clinically-focused -- Statistical power; Behavior therapy -- Periodicals; 616.89142
• DOI: 10.1016/j.jbtep.2015.01.004
• Languages: English
• ISSNs: 0005-7916

Abstract

Abstract: Background and objectives: The Implicit Relational Assessment Procedure (IRAP) is a technique that is attracting a substantial body of research literature, particularly within the clinical domain. Method: In response, the present paper outlines a meta-analysis of clinically-focused IRAP effects ( N = 494) to provide the first estimate of how well such effects validate against their respective criterion variables in general. Results: The meta-analysis incorporated clinically-focused IRAP effects from 15 studies yielding a large effect size, r ¯ = .45, with a desirably narrow 95% credibility interval (.23, .67). The funnel plot and subsequent sensitivity analyses indicated that this meta-effect was not subject to publication bias. Limitations: The present meta-effect is an estimate based upon an IRAP literature that is still evolving rapidly in the clinical domain, and so as per its accompanying credibility interval, all conclusions that follow are necessarily provisional even if bounded. Apart from the fact that the current meta-effect might be subject to inadvertent under- and/or over-estimations of the current literature, the present meta-effect might strengthen with further refinements of the IRAP. Conclusions: The current meta-effect provides the means to calculate what sample size would be required to achieve a statistical power of .80 when testing the criterion validity of clinically-focused IRAP effects using a given parametric statistic. For example, first-order Pearson correlations would hypothetically require an N of 29–37 for such purposes depending upon how conservatively over-estimation of the present meta-effect is controlled for. Overall, the IRAP compares favourably with alternative implicit measures in clinical psychology. Highlights: A meta-analysis of clinically-focused IRAP effects from 15 published studies. Meta-effect of r ¯ = .45 was observed with desirably narrow 95% credibility interval. Suggests that a sample size of 29 needed to achieve a statistical power of .80. The IRAP compares favourably with alternative implicit measures.