Retraction of research articles is increasing but the reasons and characteristics of retractions involving authors from Africa have not been studied. Using records from the Retraction Watch database, we analyzed information on articles retracted between 2014 and 2018 with at least one author or co-author affiliated with an institution in the African region to determine the most prevalent types of misconduct, subject fields, and the characteristics of researchers or research teams associated with retraction. Plagiarism was the most frequent form of misconduct, followed by duplication. International collaboration was associated with fewer retractions for plagiarism and errors in data, but increased retractions due to authorship issues. Teams with at least one senior member were associated with fewer retractions due to plagiarism but more due to duplication of articles. We conclude by making recommendations for best practice, further research, and highlighting implications for education.