There is growing concern about the detrimental effect of term-time employment on university students' academic success. We report results from an online survey of 1837 students of the University of Canterbury in New Zealand, whose responses were later matched to their academic records for a semester. The majority of employed students reported working out of financial necessity. There was no difference in grades between employed and non-employed students, but hours worked had a direct negative linear effect on the grades of employed students. Subsequent analysis suggested that employed students might have had significantly higher grades than the non-employed subsample if they had not worked. Mediating and moderating effects on the relationship between hours worked and academic performance were identified, and a model of the work--achievement relationship developed. Universities could do more to accommodate the reality of part-time work by their students.