Table 2.
Selected extracts of retraction notices by type of error
| Problems with Getting or Acquiring Data |
|---|
| “…there may have been some fluorescence impurity/contamination with cAMP when they conducted their original experiments.” [79] |
| “…the authors discovered after publication that one of the cell lines described in the article had been unintentionally misidentified” [37] |
| “… we discovered a technical error in the measurement of CF sputum phenazines” [36] |
| “… incorrect cohort identification (ie, we missed many patients who were eligible for cardiac rehabilitation). We did not query all relevant codes used to identify the cohort of patients with ischemic heart disease.” [39] |
| “…more than 500 cases of the total 1882 cases of hernia patients presented in the paper were actually hydrocele of tunica vaginalis, not hernia” [80] |
| “transgenic mice reported in Supplementary Figs. 3 and 6 and in Figs. 4 and 5 were misidentified” [38] |
| “… the data collected from self-report of the HBV vaccination remains unverified, and potentially subject to errors.” [81] |
| “…a number of subject data points had been mistakenly duplicated …” [82] |
| “… we identified incorrectly entered data for six subjects on two variables.” [40] |
| “…some data points that should have been entered as a positive result were instead entered as having a negative result…” [41] |
| “…data that was input in SPSS is from another questionnaire” [83] |
| “…underlying data for the reported experiments are unavailable due to issues including the amount of time that has passed (seven years)…” [84] |
| “Following inquiries, it turns out that the raw data are no longer available having been lost as a result of computer failure.” [85] |
| “… the original image data for experiments shown in Figs 2, 3, and 7 are no longer available.” [86] |
| Problems with Preparing or Analyzing Data |
| “Through the automated process of the analysis the authors mistakenly failed to identify that these values were inverse values, and thus, the direction of changes in the individual gene analysis is opposite to those reported in the article.” [87] |
| “While there are no concerns about the data themselves, the experimental and control groups were inadvertently switched during the original analysis. This error unfortunately lead to the opposite results being reported.” [43] |
| “…the responses for ‘attitude’ and ‘intention’ measures were switched and may have influenced the findings from the developed regression model and its results” [42] |
| “…The purpose of the recoding was to change the randomization assignment variable format of “1, 2” to a binary format of “0, 1.” However, the assignment was made incorrectly and resulted in a reversed coding of the study groups.” [44] |
| “There was a major error in the coding in their dependent variable of marital status” [88] |
| “identified a mistake in the way the original data were merged” [89] |
| “…we had miscoded the National Health and Nutrition Examination Survey variable of trying to lose weight by missing a skip pattern that started in the 1999–2000 survey.” [90] |
| “…the code created to manually anonymize the data was accidentally run twice. During the first run, anonymized subject identifiers were successfully assigned to both biosamples and clinical data. However, after this first run had passed quality control checks, the anonymization code was re-run inadvertently, replacing the first correct set of identifiers with a random and incorrect set.” [91] |
| “The models did not include random slopes for the term perceived makeup attractiveness, and we have now learned that the Type 1 error rate can be inflated when by-subject random slopes are not included” [45] |
| “For analysis of repeatedly-assessed time-related data, the authors used comparison of groups at identical time points. This form of cross-sectional comparison at individual time points is not appropriate as it fails to account for patient variability.” [92] |
| “the statistical analysis to handle risk factors with more than two categories is incorrect” [93] |
| “The reason for this decision is that the statistical methodology we used did not adequately limit the impact of outlier data points on our findings. This was evident after reanalysis of the data using a different method.” [94] |
| “A methodological error has led to immortality bias within the findings of this article; therefore, the survival intervals for participants used in this survey were unsound.” [46] |
| “…review has confirmed firstly that within-group changes were highlighted rather than between-group differences as appropriate for a randomized trial…” [95] |
| “the authors discovered statistical errors which need further validation” [96] |