If hours can be freely varied within jobs, the effect on hours of
changes in preferences for those who do change jobs should be similar to
the effect on hours for those who do not change jobs. Conversely, if
employers restrict hours choices, then changes in preferences should affect
hours more strongly when the job changes than when it does not change. For a
sample of married women we find that changes in many of the labor supply
preference variables produce much larger effects on hours when the job
changes.