I report a study, based on a representative sample of white Canadian m
en, of self-employment/paid-work choice. The main findings were as fol
lows: (1) Relative potential earnings was the main choice determinant.
(2) Potential earnings differences between groups were primarily due
to unobserved factors: market values of observed characteristics were
similar between groups, and relative potential earnings, based only on
observed characteristics, were not an explainer of choice. (3) Paid w
orkers had higher potential earnings in both sectors, with a greater a
dvantage in paid work. But the apparent comparative advantage was not
significant.