Most article retractions are due to misconduct, but about 20% are retracted because of an unintentional error or methodological flaw.
In a free-access article in The Lancet, Tea Marasović, Ana Utrobiĉić and Ana Maruŝić highlight a method which several journals have begun using to credit the correction of an honest error, by republishing a corrected article instead of retracting. As well as highlighting this positive step, the paper also comments on the inconsistencies between stakeholders in communicating research in handling corrections for such honest errors.
Marasović T, Utrobićić A, Marusić A. Transparency of retracting and replacing articles. The Lancet 2018;391(10127):1244-1245 (doi: 10.1016/S0140-6736(18)30487-2)