This article is concerned with a method for making inferences about various measures of vaccine efficacy. These measures describe reductions in susceptibility and in the potential to transmit infection. The method uses data on household outbreaks; it is based on a model that allows for transmission of infection both from within a household and from the outside. The use of household data is motivated by the hope that these are informative about vaccine-induced reduction of the potential to transmit infection, as household outbreaks contain some information about the possible source of infection. For illustration, the method is applied to observed data on household outbreaks of smallpox. These data are of the form needed and the number of households is of a size that can be managed in a vaccine trial. It is found that vaccine effects, such as the mean reduction in susceptibility and the mean reduction in the potential to infect others, per infectious contact, can be estimated with precision. However, a more specific parameter reflecting the reduction in infectivity for individuals partially responding to vaccination is not estimated well in the application. An evaluation of the method using artificial data shows that this parameter can be estimated with greater precision when we have outbreak data on a large number of small households.