The Demographic and Health Surveys (DHS) program routinely collects information about the local availability of family planning services in a separate survey in which the sample clusters ate the units of analysis. This paper proposes analytical and statistical models for using this information and illustrates the models with data from rural Guatemala. In most of the analysis, clusters are the units of analysis. The number of women in the cluster, the number of current users of modern methods, and the means and proportions of socioeconomic indicators such as literacy are attached to the cluster data. Then follows an assessment of the impact, upon contraceptive prevalence, of the presence of family planning facilities, the degree of access to these facilities, their quality, and the existence of a local distribution program. It is found that both distance and travel time to facilities are quite important and that a few indicators of quality are also significant, but specific coefficients are difficult to interpret because of missing data and the substitution of facilities. When the socioeconomic environment is taken into account, the effects of service availability are substantially reduced. Travel time and distance to facilities remain critically important, but the quality indicators lose their statistical significance.