Re: differential mechanism

Part 4.

With regard to the Davidson and Parker (2001) meta-analysis, these authors separately compared EMDR vs. Eyes Fixed Desensitization and Reprocessing (EFixDR) and EMDR vs. other alternating stimuli. In comparing EMDR vs. Eyes Fixed DR, Davidson and Parker offered the following caveat:

The present data show that the EMDR-EFixDR effect
size is marginally significant if one examines only clinical populations satisfying Diagnostic and Statistical Manual of Mental Disorders (DSM)
diagnostic (selecting PTSD and other Anxiety
Disorders in Table 6). (p. 311)

However, Davidson and Parker go further than simply noting a marginally significant effect. They conduct analyses to determine "[i]f we limit ourselves to clinical populations, what sample size would give sufficient power to permit the effect of eye movements to be assessed?" (p. 311) They provide two answers to this question.

The first answer is in terms of how large a single study would have to be in order to achieve a power of .80. They conclude that the study would have to have between 144-780 subjects for a two-group study, depending on assumptions about the population effect size. The second answer is in terms of how many additional smaller n studies it would take to change the current meta-anlytic conclusions. Depending on assumptions about the magnitude of standard deviations, it would take between 38-190 additional studies.

The author's apparently do not think this "caveat" is a very serious one, as they state their conclusions without reservation:

Is eye movement -- or indeed, any alternating
movement -- a necessary component of EMDR? The
near-zero effect size indicates that eye movement
is unnecessary. Dismantling studies that examined
the use of alternating movements other than were
were not shown to be different from EDMR by the
data available. The effect sizes for the outcome measures are small, and we have no evidence that
eye movements or other alternating stimuli are necessary. (p. 313)

While readers of the Davidson & Parker analysis (and this post) can draw their own conclusions about how meaningful the "marginally significant" effect size among clinical populations is, my own conclusion that using their caveat as support for the efficacy of eye movements constitutes grasping at straws.

Citation:
Davidson, P.R., & Parker, K.C.H. (2001). Eye movement desensitiation and reprocessing (EMDR): A meta-analysis. Journal of Consulting and Clinical Psychology, 69, 305-316.

Replies:

Re: differential mechanism, by Cahill, 11/03/01

| Behavior OnLine Home Page | Disclaimer |

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