ABSTRACT: Large wage differences exist between married and single men. This paper argues that this marriage wage premium can arise from statistical discrimination on the basis of marital status in a labor market with incomplete information and employer learning. Using NLSY79, this hypothesis is tested separately for high school graduates and college graduates. The results suggest that high school graduates face a labor market with information friction and experience statistical discrimination on marital status, while college graduates are on the contrary. This paper also provides a framework to quantify the treatment effect of marriage in an incomplete information setting.