The case raises important questions about when retractions are appropriate, and whether they can have a chilling effect on scientific discourse. Although Hanna Kokko of the University of Zurich, Switzerland — who co-authored both papers — agreed that the academic literature needed to be corrected, she didn’t want to retract the earlier paper; the journal imposed that course of action, said Kokko.
Let’s take a look at the retraction note:
The above article , published online on 31 March 2007 in Wiley Online Library (wileyonlinelibrary.com), has been retracted by the journal Editor in Chief, Ruth Shaw, due to inadvertent technical errors in analysis of the model. These errors are explained and corrected in a forthcoming article  by Lutz Fromhage, Michael Jennions and Hanna Kokko, which is endorsed by both authors of the original paper.
Kokko gave us a bit more background:
It was me who initiated looking at the topic, but not because I suspected that earlier ideas were wrong, rather because I wanted to expand on some of the results. Lutz then identified the problems, so the project changed from the initial idea.
When asked if scientists should be willing to correct the literary record, she added:
yes in the sense of willing to correct previous results, and openly admit where things were done in a less than ideal way; but no, in the sense of wanting earlier work to be readily retracted as soon as issues are discovered.
The retracted paper, “What Determines Sex Roles in Mate Searching?,” has been cited 49 times, according to Thomson Reuters Web of Science. The new paper, “The Evolution of Sex Roles in Mate Searching,” has not yet been cited.
Kokko noted that constant fears of retraction could make academics reluctant to admit past mistakes. She explained her stance:
In our view science is best served if (a) papers identifying a scientific problem and making initial attempts at solving them are, generally, kept in the literature even if flaws are later found, as there is usually quite a bit of merit left in the sense of identifying areas of research; (b) if I was criticizing someone else’s work (than my own) at a similar level of severity, I would still not want that person to retract (or to be forced to retract) the original, for that would leave counterproductive gaps in literature and change research atmospheres towards a much more destructive game.
She went further to question the journal’s decision:
If the journal rather than the person initiating the criticism is very trigger-happy in rejections, the consequence for me as a scientist is to become somewhat restrained in when I “dare” to criticize, which is a counterproductive result for the scientific progress.
Lutz Fromhage from the University of Jyväskylä in Finland, who co-authored the new paper, added more context:
this story started when we first submitted our 2016 paper, which included an appendix explaining the errors of the 2007 paper (which I was the first to notice).
Initially, a reviewer and associate editor suggested correcting the 2007 paper, Fromhage said, but after some discussion, the editors ruled out this possibility on the grounds that an erratum should have the same author list as the original paper — which was not the case here.
According to Fromhage, the errors stemmed from the “mathematical argument” in the 2007 paper, which ended up contradicting its conclusions about why males usually seek out females, and not vice versa. Here is how he presented the conclusions of the 2007 paper:
(A1) the evolution of mate searching (i.e., should males search for females or vice versa?) is not influenced by females spending more time being unavailable for mating (e.g., because of pregnancy), even if this means that available females are outnumbered by available males.
To explain why males are nevertheless usually the searching sex, Kokko and Wong 2007 suggested that
(A2) this pattern is created by polyandry (i.e. females mating with multiple males)
(A3) this pattern arises if searching is costlier for females than for males.
In comparison, he also listed the conclusions of the 2016 study:
B1) time spent unavailable for mating weakens selection for mate search (contradicting A1 above).
(B2) male-biased searching is not created by polyandry per se (contra A2), but by the presence of a time interval (‘mating window’) during which the timing of mating does not matter for females.
(B3) finding (A3) above was upheld.
(B4) male-biased searching evolves if the costs of searching are not temporally limited to the act of searching (i.e., cost ‘carry over’ to other parts of the life-cycle)
Fromhage noted that the contradiction between A1 and B1 “directly caused” the paper to be pulled, adding:
A2 is really a clarification of why polyandry had such an effect in the 2007 paper; and the novel finding B4 follows from a previously unstudied assumption.
Kokko and Fromhage have previously also co-authored a 2013 corrigendum in Nature Communications, and Kokko has issued another corrigendum in Ecology Letters in 2008.
Regarding both corrigenda, Kokko said:
Both cases are about typos in printed versions of a paper where one or, in the first paper, two equations appeared in an incorrect form. In neither case have actual analyses (production of figures, etc) used the incorrect form, so the corrections were necessary to restore the match between a printed equation and the analysis that was actually done in the paper.
When asked for comment, Ruth Shaw, the editor-in-chief of Evolution, who is based at the University of Minnesota, and Wiley’s managing editor, Katie Simmons, referred us to the retraction note and the new paper.
Update, 1 p.m. Eastern, 4/29/16: Please see a guest post from one of the reviewers of the updated paper.
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.