Talk:Chamberlain's approach to unobserved effects models