Collaboration and its influence on retraction based on retracted publications during 1978-2017

Scientific collaboration has considerably grown in recent years, and has been recognized as a positive factor for research productivity and quality. This study is aimed at exploring the collaboration in retracted publications and the influence on different kinds of scientific misbehaviors based on 4842 retracted publications in Web of Science Core Collection from 1978 to 2017. Search results showed that the proportion of international collaborations in retracted publications has fluctuated at around 20% in the past 17 years, and different collaboration patterns of major countries are revealed. The USA and China contributed to more than half of all retracted publications, while retraction rate of Chinese papers was three times higher compared to papers from the USA. A negative correlation between the proportion of internationally collaborative publications and the retraction rate was found. Countries involved in more international collaborations tend to have lower retraction rates. Also, the occurrences of plagiarism was found to be lower when a paper is contributed by more authors, while falsification or manipulation, errors or unreliable results appeared more frequently under the same condition. Conclusions are made with a focus on the need for more actions to strengthen the scientific norms and integrity within collaborative groups.