Background There is some consensus on the value of cognitive‐behaviourally informed interventions in the criminal justice system, but uncertainty about which components are of critical value. Aims To test the hypothesis that change in prisoners ‐ criminal thinking and institutional misconduct ‐ will both follow completion of a brief cognitive behavioural intervention. Methods A one‐group pre‐test–post‐test quasi‐experimental design was used to assess change on the General Criminal Thinking (GCT) scale of the Psychological Inventory of Criminal Thinking Styles among 219 male prisoners completing a 10‐week cognitive behavioural intervention, referred to as ‘Lifestyle Issues’. Institutional misconduct was measured for 1 year prior to completion of the course and 2 years subsequently. Using variable‐oriented analysis, post‐test GCT scores were compared with change in prison conduct, controlling for the pre‐test thinking scores. Calculations were repeated by using person‐oriented analysis. Results Prisoners who displayed a drop in GCT scores between pre‐test and post‐test levels were significantly more likely to show a reduction in prison misconduct, whereas prison misconduct was likely to escalate among those who displayed a rise in criminal thinking scores from pre‐test to post‐test. Conclusions These findings must still be regarded as preliminary, but taken together with other work and with cognitive behavioural theory, they suggest that development of more prosocial thinking and abilities may have an early beneficial effect on institutional behaviour. Their measurement may offer a practical way in which men could be assessed for readiness to return to the community. Copyright © 2017 John Wiley & Sons, Ltd.