Dear Drs. Baumeister and Vohs,
Perspectives on Psychological Science published the results of a “A Multi-Lab Pre-Registered Replication of the Ego-Depletion Paradigm Reported in Sripada, Kessler, and Jonides (2014).” The main finding of this replication project was a failure to demonstrate the ego-depletion effect across multiple labs with a large combined sample size.
You wrote a response to this finding (Baumeister & Vohs, in press). In your response, you highlight several problems with the replication studies and conclude that the results only show that the specific experimental procedure used for the replication studies failed to demonstrate ego-depletion.
At the same time, you maintain that ego-depletion is a robust phenomenon that has been demonstrated repeatedly for two decades; quote “for two decades we have conducted studies of ego depletion carefully and honestly, following the field’s best practices, and we find the effect over and over.”
It is regrettable that the recent RRR project failed to show any effect for ego-depletion because the researchers used a paradigm that you never approved and never used in your own two decades of successful ego-depletion research.
I would like to conduct my own replication studies using paradigms that have reliably produced ego depletion effects in your laboratories. As a single paradigm may fail for unknown reasons, I would like to ask you kindly to identify three paradigms that based on your own experience have reliably produced the ego-depletion effect in your own laboratory and that can produce the effect in other laboratories.
To plan sample sizes for my replication studies, it is also important that you provide an estimate of the effect size. A meta-analysis by Hagger et al. (2010) suggested that the average ego-depletion effect size is d = .6, but a bias-corrected estimate suggests that the effect size may be as small as d = .2 (Carter & McCollough, 2014). I would hate to end up with a non-significant result because my replication studies were underpowered and failed to detect the ego-depletion effect due to insufficient power. What effect size would you expect based on your two decades of successful studies?
Dr. Ulrich Schimmack
Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. D. (2010a). Ego depletion and the
strength model of self-control: A meta-analysis. Psychological Bulletin, 136, 495-525.
Carter, E. C., & McCullough, M. E. (2014). Publication bias and the limited strength model of self-control: Has the evidence for ego depletion been overestimated? Frontiers in
Psychology, 5, 823. doi: 10.3389/fpsyg.2014.00823
Sripada, C., Kessler, D., & Jonides, J. (2014). Methylphenidate blocks effort-induced depletion of regulatory control in healthy volunteers. Psychological Science, 25, 1227-1234. doi: 10.1177/0956797614526415
4 thoughts on “Open Ego-Depletion Replication Initiative”
I think it’s also great that Baumeister & Vohs are planning to do a pre-registered replication study themselves and i would like to add (if it makes sense) that i hope they will follow a Registered Report format so there is no chance of publication bias.
I am sure a journal like Comprehensive Results in Social Psychology would welcome their Registered Report:
It is great if they do their own replication study, but I want to see what paradigm they think will work and I want to see whether this paradigm will also work outside their labs.
It seems that Hagger et al’s failure to replication mainly results from the ineffectiveness of the manipulation, as revealed by a re-analysis, please see the link: http://journal.frontiersin.org/article/10.3389/fpsyg.2016.01155/full
Regarding which manipulation is more effective, an up to date meta-analysis showed the effect size for each frequently used manipulation, see this link: https://www.researchgate.net/publication/305493370_An_Updated_Meta-Analysis_of_the_Ego_Depletion_Effect
It’s a good idea to replicate ego depletion by taking into account the effectiveness of each manipulation task, which would shed light more on further studies in this field.