Re: differential mechanism

Two objections to my discussion of dismantling studies can be anticipated and dealt with proactively.

Objection 1. The dismantling studies are not convincing because most were poorly done: Therapy was often too short for the subject population; treatment fidelity is not reported in most studies and may be questioned in some/many; some studies used "analogue" or student populations which are not severe enough to test the effects of removing a single element of a powerful multi-component treatment; therapists were not adequately trained; studies had inadequate power to detect differences between variations of EMDR; or what have you.

Fine. Let's prune out all of these studies. What are we left with? In my opinion, only one study would survive such a selection process: Feske & Goldstein's study of panic disorder finding modest superiority for the eye movement condition at posttreatment (statistical significance was obtained on 2 of five measures) that disappeared at 3-month follow-up. Not impressively large effect, not a durable effect, open to an alternative explanation in terms of differential credibility of treatent, and yet to be replicated. Awfully thin ice to be skating on after almost 10 years of research into this issue. Weeding out the bad studies does not result in convincing evidence for the efficacy of eye movements or any other form of bilateral stimulation.

What are the implications of this? Proponents may wish to argue that because the dismantling literature is so poor, we can't be confident in the conlcusion that eye movements make little or no difference. However, we will NEVER be justified in concluding that eye movements make little of no difference. This is because you can't prove the null hypothesis. The burden of proof lies with showing that eye movements (or other bilateral stimuli, or dual attention) make a difference, not in showing that they don't make a difference.

Objection 2. EMDR is a powerful, multicomponent therapy. Therefore, simply removing one element won't significantly impair outome.

This may be true, but some EMDR proponents have attempted to swing this sword both ways. For example, Ron Acierno's study of eye movments for treating phobias which found no effects for eye movements was dismissed because he told subjects to "relax" between sets of eye movements, rather than "blank it out." Such a small differenc in language is proposed to be responsible for completely elminating EMDR's effectiveness, but changes in things such as removing eye movements or installation trials have no effect? This seems a bit far fetched to me, and it appears that EMDR is either extremely sensitive to alterations in procedure or extremely robust (i.e., effective despite changes in procedure) depending on the data being interpreted. When results are potentially in favor of EMDR, then EMDR is a robust treatment. When null results are obtained, EMDR is very sensitive to alterations in the procedure. It certainly seems someone is trying to both have and eat their cake.

Replies:

Re: differential mechanism, by Susan Rogers, 11/05/01 Re: differential mechanism, by Susan Rogers, 11/06/01 Re: differential mechanism, by Cahill, 11/10/01

| Behavior OnLine Home Page | Disclaimer |

Copyright © 1996-2004 Behavior OnLine, Inc. All rights reserved.