A prevailing assumption in survey research is that the higher the response rate, the more representative the sample. However, given real-world budget constraints, it is necessary to consider whether the benefits of marginal increases in response rates outweigh the costs. It is possible that, at a certain level of effort, other means of increasing the sample or reducing survey error may be more cost-effective than continued attempts to increase the response rate (see Groves 1989). In this article, we examine the costs and benefits of increased response rates on the composition of fathers in the Fragile Families and Child Wellbeing Study, a national longitudinal survey of new parents. Because we have mother-reported data on fathers who were not interviewed, we are able to compare the sample compositions of fathers at various levels of response rate with that of all fathers who were eligible for the study. Most of the fathers in the sample are unwed, poor, and live in large cities. They therefore represent a population that is notorious for being difficult to locate and interview. A mixed-mode approach was used to maximize father response rates, and this resulted in a wide range in costs per case.