David Allison writes:

I hope you will consider covering this in your blog. I admire the authors for simply admitting they made an error and stating clearly and without equivocation that their original conclusions were not substantiated.

More attention to the confusing effects of regression to the mean are warranted as is more praise for people who just come out and say “I made a mistake and my original conclusions were incorrect” without spin, varnish, or denial.

The article in question is Correction to: Effectiveness of Nutrition Intervention in a Selected Group of Overweight and Obese African-American Preschoolers, by Yulyu Yeh, Kathryn Brogan Hartlieb, Cynthia Danford, and K.-L. Catherine Jenn, in the Journal of Racial and Ethnic Health Disparities, where they write:

Part of our original analyses was performed with overweight and obese preschoolers only. This procedure may have created a subgroup of individuals with extreme values at baseline and this may likely be inappropriate. It is conceivable that these subjects were in an extreme state when they were measured at baseline and that over time, one would expect them to regress back to a more normal state, independent of any intervention. We did not take into consideration the phenomenon called regression to the mean [1, 2]. As a result, we cannot make any affirmative statements about the effectiveness of our interventions since a) there were no differences in changes in BMI percentile among the three groups when all preschoolers were analyzed together; and b) in the subgroup of overweight and obese preschoolers, all three groups, including the controls, regressed back towards a more normal state, even though the changes observed in intervention Group A were not statistically significant.

It is indeed good when researchers admit error. We all make mistakes; what is important is to learn from these mistakes.