The authors of this paper tested this prediction directly, by comparing skin reflectance date of adult men and women from different populations with latitude, discovering that sexual dimorphism is not correlated with latitude. Hence, one of the predictions of the "sexual selection hypothesis" is not supported by the data.
From the conclusions of the paper:
We tested the hypothesis that human sexual dimorphism
of skin color should be positively correlated with
distance from the equator, a proposal generated by the
sexual selection hypothesis. We found no support for that
proposition. Before this paper was written, the sexual
selection hypothesis was based on stated male-preference
data in a number of human groups. Here, we focused on
the actual pattern of sexual dimorphism. We report that
the distribution of human sexual dimorphism in relation
to latitude is not that which is predicted by the sexual
selection hypothesis. According to that hypothesis, in
areas of low solar radiation, there should be greater sexual
dimorphism, because sexual selection for lighter
females is not counterbalanced by natural selection
for dark skin. Our data analysis does not support this prediction.
American Journal of Physical Anthropology (Early View)
Human skin-color sexual dimorphism: A test of the sexual selection hypothesis
Lorena Madrigal, William Kelly
Applied to skin color, the sexual selection hypothesis proposes that male preference for light-skinned females explains the presence of light skin in areas of low solar radiation. According to this proposal, in areas of high solar radiation, natural selection for dark skin overrides the universal preference of males for light females. But in areas in which natural selection ceases to act, sexual selection becomes more important, and causes human populations to become light-skinned, and females to be lighter than males. The sexual selection hypothesis proposes that human sexual dimorphism of skin color should be positively correlated with distance from the equator. We tested the prediction that sexual dimorphism should increase with increasing latitude, using adult-only data sets derived from measurements with standard reflectance spectrophotometric devices. Our analysis failed to support the prediction of a positive correlation between increasing distance from the equator and increased sexual dimorphism. We found no evidence in support of the sexual selection hypothesis.