In job ads, employers express demand for personality traits when seeking workers to perform tasks that can be completed with different behaviors (e.g., communication, problem-solving) but not when seeking workers to perform tasks involving narrowly prescribed sets of behaviors such as routine and mathematics tasks. For many tasks, employers appear to demand narrower personality traits than those measured at the Big Five factor level. The job ads also exhibit substantial heterogeneity within occupations in the tasks mentioned. Workers may thus sort based on personality-derived comparative advantages in tasks into jobs rather than occupations. In the National Longitudinal Survey of Youth 1997, we confirm that personality sorting based on tasks occurs at both the occupation and job levels. In this sample, however, there is little evidence of task-specific wage returns to personality traits, which would influence the supply of traits to jobs with particular tasks. This may explain why personality sorting based on tasks in the sample is very limited in spite of the correlations between tasks and employers' demands for traits.