The purpose of this thesis is to examine the effect of working at home on the hourly wage for men and women in the United States. Data are taken from the May 1997 Current Population Survey. A human-capital model of hourly earnings is specified which allows for the endogenous determination of the choice of home-based or office-located work. The model is estimated by a two-step, instrumental-variables procedure. First, a reduced-form equation determining the probability of working at home is estimated by the maximum-likelihood Probit method. The predicted values from this equation are then substituted for the observed, dichotomous homeworker-status variable in the (log) hourly wage equation. The empirical results reveal a statistically significant wage premium for persons who work at home, relative to their observationally equivalent office-located counterparts. Several explanations for this finding are discussed, and suggestions for future research are offered.