The Brief COPE is a self-completed questionnaire measuring coping strategies. It comprises 14 subscales for which psychometric properties are described. Three composite subscales measuring emotion-focused, problem-focused, and dysfunctional coping have proved useful in clinical research and have content validity. We report psychometric properties of these subscales for the first time. One hundred twenty-five family carers of people with Alzheimer's disease completed the Brief COPE at time 1, 92 (82.9% of eligible carers) a year later, and 74 (77.1%) 2 years later. Internal consistencies were good for emotion-focused, problem-focused, and dysfunctional subscales (alpha = 0.72, 0.84, 0.75). Test-retest reliability over a year was demonstrated for emotion-focused, problem-focused, and dysfunctional subscales among carers in whom burden scores did not change significantly (r = 0.58, r = 0.72, r = 0.68; p < 0.001). Change in burden score over 2 years correlated with change in problem-focused and dysfunctional (r = 0.33, r = 0.32; p < 0.01) subscales, indicating sensitivity to change, but not with change on the emotion-focused scale. Change in emotion-focused coping correlated with change in problem-focused and dysfunctional coping (r = 0.40, r = 0.26; p < 0.05). Regression analyses indicated convergent and concurrent validity: emotion-focused coping was predicted by secure attachment (beta = 0.23) and by problem-focused coping (beta = 0.68); dysfunctional coping by burden (beta = 0.36) and less secure attachment (beta = -0.25) and problem-focused coping (beta = 0.31;all p < 0.05). The model predicting problem-focused coping included avoidant attachment (beta = 0.22; p = 0.014), social support (beta = 0.10; p = 0.25), care recipient activities of daily living impairment (beta = 0.12; p = 0.14) and less secure attachment (beta = -0.25; p = 0.011) and emotion-focused (beta = 0.53; p < 0.001) and dysfunctional coping (beta = 0.25, p = 0.006). These subscales are potentially useful in clinical research as they reflect possible components of interventions to change coping, although more information about sensitivity to change of the emotion-focused subscale is needed.