Background Anger is commonly measured as if it were a single, simple construct. This may be particularly unhelpful if the main purpose of a measure is to determine change and responsiveness to interventions. Aims Our primary aim was to assess five anger parameters in parolees – frequency, duration, intensity, latency, and threshold – and to test for effects of psychoeducation and social desirability bias on parolees' scores. Average anger scores for the offender sample were compared with those in a non‐offender sample. Methods The offender sample was drawn from male parolees in San Antonio. Age‐matched volunteers recruited at public libraries were engaged for baseline comparisons. The Anger Parameters Scale and the Marlow–Crowne Scale were used to assess anger and social desirability, respectively. Parole officers delivered a psychoeducation course to parolees over 12 weeks, and anger and social desirability measures were taken before and afterwards. Results At baseline, parolees were angry more often, stayed angry longer, and reached higher levels of anger than the non‐offenders, confirming their eligibility for the programme. Mean anger scores were not significantly different after psychoeducation than before it. Parolees' reported anger was significantly and negatively correlated with social desirability scores. Conclusions Only three of the five anger parameters were prominent among these offenders: frequency, duration, and intensity of their outbursts. Psychoeducation did not produce improvement, possibly because it was instructional rather than therapeutic, but also because group means may mask useful individual differences. Concurrent assessment of social desirability is likely to assist in interpretation of anger measures. Copyright © 2017 John Wiley & Sons, Ltd.