Background Several studies have found a connection between attentional deficit hyperactivity disorder (ADHD) and criminal behaviour in clinical and prison samples of adults, but there is a lack of representative general population data on this. Aim To test relationships between histories of ADHD and arrest. Our main research question was whether any such relationship is direct or best explained by co‐occurring variables, especially indicators of social bonds. Method Data were from a sample of 5,376 adults (18+) representative of the general population of Ontario, Canada. Logistic regression analysis was used to explore the relationship between self‐reported arrest on criminal charges and ADHD as measured by the Adult Self Report Scale (ASRS‐v1.1). Indicators of strong social bonds (post secondary education, household size) and weak bonds (drug use, antisocial behaviours, alcohol dependence) were also obtained at interview and included in the statistical models. Results In a main effects model, screening positive for ADHD was twice as likely (OR 2.05 CI 1.30, 3.14) and past use of medications for ADHD three times as likely (OR 3.94 CI 2.46, 6.22) to be associated with ever having been arrested. These associations were no longer significant after controls for weak and strong social bonds were added to the models. In the best fitting statistical model, ever having been arrested was not associated with ADHD, but it was significantly associated with indicators of strong and weak social bonds. Conclusions The observed connection between ADHD and criminality may be better understood through their shared relationships with indicators of poor social bonds. These include antisocial behaviour more generally, but also drug use and failure to progress to any form of tertiary education, including vocational training. Copyright © 2017 John Wiley & Sons, Ltd.