BACKGROUND The study was designed to investigate whether seasonal mood and behavioural changes are detectable prospectively in a non-clinical population in the way they have been reported in retrospective studies. The specificity of any seasonal fluctuation in affective state was also investigated by measuring anxiety as well as depression.
METHOD To measure seasonal fluctuations in affect and behaviour prospectively, 25 women were interviewed every month for one year using four scales (depression, anxiety, stress, and behavioural change). Retrospective accounts of mood and behaviour at the end of the year were collected with the Seasonal Pattern and Assessment Questionnaire (SPAQ).
RESULTS Seasonal depression peaked in winter as did atypical behaviour when measured either prospectively or retrospectively, but the difference between winter and summer was much more pronounced in the retrospective data. No seasonal effect was found for anxiety or psychosocial stressors.
CONCLUSIONS The results obtained by retrospective techniques have limited reliability. In the future, more prospective studies with unbiased, standardised instruments are recommended to measure seasonal variations in affect and behaviour.