At least one study uses household income [1]. The effect isn't adjusted for family size in any way I can see. Do note that the linked study differentiates between 'enjoyment of life' which they estimate starts to plateau around $75k / household and 'life satisfaction' which keeps going up with earnings [2]. That difference may explain much of the supposed controversy as outlined in the original post.
[1] http://www.pnas.org/content/107/38/16489.full#T2
[2] http://blogs.wsj.com/wealth/2010/07/02/money-can-buy-satisfa...