Paper

Period

Nb of articles

Study features

(Budd et al., 1998)

1966-1997

235

Studied citations, post-retraction citations and citation types, type of citing publication, retraction reasons, time to retraction, who initiated the retraction. Main conclusion: the majority of citing articles cite the retracted papers as valid research.

(Nath et al., 2006)

1982-2002

395

The number of authors, type of study, date of publication, if retraction requested by authors. Analyzed misconduct vs mistakes: time to retraction 2 years for mistakes, 3.3 years for misconduct.

(Cokol et al., 2007)

1950-2004

596

Presents a model estimating the real number of papers that should be retracted, based on the impact factor of journals. Estimates the number of potentially retractable articles between 10,000 (best case) and 100,000 (worst case) out of 9,398,715 articles.

(Redman et al., 2008)

1995-2004

315

Analyzed pre and post-retraction citations, time to retraction, the primary reason for retraction, reproducibility as reason, and who requested the retraction. Main conclusion: pre-retraction highly cited studies continue to be highly cited after retraction. Significant differences between citations depending on retraction reason. Journal impact factor influence the number of citations.

(Steen, 2011)

2000-2010

788

Studied English language retractions: retraction reason, countries, repeat offenders, JIF, comparing fraud (data fabrication or falsification, data plagiarism) with errors (text plagiarism included). Conclusions: JIF higher for fraudulent articles, most of the retractions from USA.

(Wager & Williams, 2011)

1980-2009

312

Analyzed 312 out of 870 retractions: retraction reasons, who initiated the retraction, article types, number of authors. Conclusions: retracted articles have increased since 1980. Main retraction reasons are mistakes, overlap (redundant publication), and plagiarism, almost 90% of retractions are research articles, inconsistencies between journal retraction practices.

(Foo, 2011)

up to 2009

303

Analyzed only authors with at least two retractions (303 from 1239 PubMed records): time to retraction, citations, the number of authors. Conclusions: time to retraction decrease from 52 months (articles published before 2000) to 33 months (after 2000), almost 25% of citations received after article retraction.

(Samp et al., 2012)

2000-2011

102

Studied drug therapy retracted articles (102 out of 742 retractions): retraction reasons, time to retraction, funding sources, repeated retractions. Conclusions: mean time to retraction is 39 months, main reason is scientific misconduct (51%), 5.6 articles in average written by repeat offenders.

(Fang et al., 2012)

Up to 2012

2047

Studied retraction reasons, temporal trends, time to retraction, geographic origin, citations. Conclusions: fraud/suspected fraud/misconduct, plagiarism and redundant publication are the main retraction reasons. Mean time to retraction for all causes is 32.9 months. “USA, Germany, Japan and China account for 75% of papers retracted for fraud or suspected fraud”.

(Decullier et al., 2013)

2008

250

Studied retraction reasons, author countries, time to retraction, visibility of retracted articles, retraction notes characteristics, journals impact factor. Identified mistakes, plagiarism and fraud as the main causes for retraction and articles with mistakes being published in high impact factor journals.

(Singh et al., 2014)

2004-2013

2343

PubMed retracted articles data: retraction reasons, article types, time to retraction. Conclusions: main reasons are mistakes, plagiarism, and duplicate publication. Forty-five percent of the retracted articles represent original research, 33% clinical cases and 18% reviews. A decrease over time of the period between publication and retraction was noticed.

(Madlock-Brown & Eichmann, 2015)

2003-2010

1113

This article studied reasons for retraction, citations (for 740 papers matched in the WoS database), and those who initiated the retraction. Conclusions/findings: mistakes, overlap (duplication) and plagiarism are the top 3 retraction reasons. Self cites represent 5% of the retracted article citations. Authors requested retraction in almost 40% of cases, editorial retractions account for 60%.

(Azoulay et al., 2015)

1973-2007

1104

Studies a set of PubMed retractions for retraction reasons, retraction speed, citations and the impact of these papers on related articles. Findings: retraction speed is 3.2 years; main retraction reasons are: fake data (32.7%), mistakes/errors (24.55%), duplicated publication (8.33%) and plagiarism (8.15%).

(Rosenkrantz, 2016)

1983-2015

48

The article analyzed retracted papers published in radiology journals extracted from 3685 retracted PubMed articles: time to retraction, retraction reasons, citations Findings: mistakes, duplicate publication and plagiarism were the main retraction reasons, retraction notes inconsistency between journals.

(Pantziarka & Meheus, 2019)

2000-2018

1512

Retracted oncology articles research: lifetime, citations, authors. Findings: mean lifetime is 1318 days, the median is 870 days. Mean citation number is 35.1/article (median = 14). Multiple retractions/author in 14% of articles.

(Rapani et al., 2020)

up to 2018

180

Retracted dental articles: retraction reason, type of study, citations (PubMed and Retraction Watch). Findings: mean time to retraction = 2.1 years (median 1 year), 55% of retracted articles originate from Asia, 28% from Europe and 10% from America. Mean number of authors is 4.6. Author misconduct, scientific errors and publisher related issues are the main retraction reasons.

(Bhatt, 2021)

1959-2019

6936

PubMed retractions combined with WoS information: retraction trends over time, time until retraction, trends among countries. Findings: 3.8 years average time until retraction, retraction rate per 10,000 publications increased from 0.38 in 1985 to 5.95 in 2014, China authors publish in low impact journals while authors from US are published in high impact journals.