Behavioural syndromes describe consistent and correlated individual differences in behavioural traits. Quantifying individual differences often requires researchers to capture and hold animals in captivity while short-term behavioural assays are recorded. We compared behavioural responses of adult, territorial Steller's jays in short- and long-term field assessments of behavioural traits in two ecological contexts, risk taking and exploration. Individuals' risk taking was similar in short-term and long-term contexts (i.e. alarm calling in the presence of a predator mount and while re-entering a trap, respectively). However, a measure of short-term exploration of a novel object in a feeding context was not related to a long-term index of annual habitat exploration (i.e. travel distance outside home territory). Risk-taking and exploration indices were correlated across ecological contexts, indicating that these traits contributed to a behavioural syndrome in jays. Annual assessments of risk-taking and exploration behaviours were repeatable. Individuals with high scores in risk taking and exploration were more likely to be recaptured in a familiar trap. We conclude that short-term experiments are adequate measures of specific behavioural strategies, but because short-term responses did not necessarily predict long-term annual behaviours in related contexts, expression of behavioural types and associated ecological strategies should be regarded as species and context specific. Long-lived residents are useful study species to overcome sampling biases for traits measured in captivity, because they provide opportunity to evenly sample a population over all personality types, including trap-shy individuals.