Agricultural cooperatives differ in many attributes which affect members' farm and off-farm earnings in different ways. As a result, time allocation patterns between farm work and off-farm work will vary significantly across cooperatives. Participation equations in farm work and off-farm work of farmers who are members of Israeli moshavim are estimated jointly, including a cooperative-specific factor in each equation. The fixed effects are found to be significant and important, but can be only partly explained by observed cooperative attributes. This provides another support to the Monte Carlo results of Borjas and Sueyoshi, that controlling for group effects is superior to alternative models. The results also imply that unobserved factors have considerable effects on farmers' time allocation, and hence should be controlled for whenever possible.