This one seems like an honest mistake: a paper on dietary supplements during pregnancy has been retracted based on an error in data recording.
In the BMC Pregnancy & Childbirth paper, “Folic acid supplementation, dietary folate intake during pregnancy and risk for spontaneous preterm delivery: a prospective observational cohort study,” women for whom the researchers had no data on folic acid supplementation were classified as taking no supplements. Despite the error, the authors claim the overall conclusion remains the same: taking folic acid supplements didn’t protect women from preterm deliveries.
Here’s the retraction notice:
This article [1] has been retracted by the authors, as some of the reported data was incorrect. The error concerned approximately 8% of the cohort (women who answered the first version of the food frequency questionnaire, before 2002) and was caused by misclassification of the amount of the women’s reported folic acid supplementation (no supplement use instead of missing information on supplement use). The error led to minor changes in the results.
However, the estimate of the association between folate intake and spontaneous preterm delivery did not change the conclusions after the error was corrected.
In consultation with the Journal’s editors we decided to retract the article from BMC Pregnancy and Childbirth.
A corrected manuscript will be resubmitted to the Journal at the time of retraction.
On behalf of all authors,
Verena Sengpiel and Bo Jacobsson.
The paper has yet to be cited, according to Thomson Scientific’s Web of Knowledge.
Author Verena Sengpiel offered us a more detailed explanation for the mistake:
As stated in the retraction text, we have submitted a new corrected manuscript, and this manuscript is currently under review at BMC Pregnancy & Childbirth. As BMC makes the submission and review process public, all readers will be able to follow the changes made to the manuscript. If possible, you could wait a couple of weeks until republication as we at present do not know whether the new manuscript will be accepted in its current form.
When discovering the honest error, we first recalculated everything in the exact same dataset and contacted the journal. In consultation with the BMC Pregnancy and Childbirth editors we then decided to retract and resubmit, and also to recalculate the results in an updated version of the data files released from the Norwegian Mother and Child Cohort Study (MoBa). After publication of the original manuscript, we discussed our paper with other researchers and agree that it would be of interest to not exclude women due to complications that manifest first during pregnancy. The new manuscript that is currently under review differs therefore from the published paper by:
- All calculations are based on version 6 of the quality-assured data files from MoBa.
- Exclusion of those women that answered the old version of the MoBa food frequency questionnaire
- No longer exclusion based on pregnancy complications
In some versions of MS Excel, empty data fields are treated as containing zeros. Much worse, however, if you paste some data into them, containing numbers along with some empty cells, the empty cells may suddenly become cells containing zeros. Maybe that’s what happened here!