Studies of primate social cognition aim at a) documentation of social behaviors that appear to require an understanding in the animals of their own relationships and the relationships among others; an ability to anticipate behavioral effects; intentionality, and so on, and b) determination of the psychological mechanisms and mental abilities underlying this complex sociality. Thus far, the first area of investigation, which is largely descriptive, has reveived more attention than the second. The most parsimonious assumption concerning nonhuman primates is that if their behavior resembles human behavior the psychological and mental processes involved are probably similar too. This is not the same as uncritically accepting the existence of these processes; alternative hypothesis need to be formulated and tested. The present paper advocates research over a broad spectrum, with reliance on a variety of methods, and emphasis on covergent evidence. Because each method has its own strengths and weaknesses, the most convincing support for a particular explanation will involve data from a number of sources, both observational and experimental. Four research methods are discussed: Qualitative description. Sometimes referred to as "anecdotalism", this method seeks to document the animals' spontaneous solutions to unusual social problems, and exploitation of rare opportunities. This method provides a starting point for research if it suggests remarkable cognitive capacities. It is unsuitable, however, for a conclusive comparison of alternative explanations. Quantitative description. The first goal of systematic research is to delineate the range of tenable hypotheses about the phenomenon of interest, i.e. to establish the cognitively least demanding explanation that cannot be rejected as well as the most demanding explanation that seems reasonable given the species' general intelligence level. The second goal is to compare alternative explanations. The advantage of quantitative description is that it can be applied in natural or naturalistic settings; the disadvantage is that is lacks control over variables. Controlled observation. A pseudo-experimental procedure that aims at controlling certain variables. An observation following a particular event is compared with an observation unpreceded by the same event but matched to the first observation on a number of other dimensions. A disadvantage, compared to experimentation, is the dependency on spontaneous events. Experimentation. Manipulation of the animal's experience and available information is a powerful tool to select among alternative cognitive explanations. The disadvantage is that this procedure requires an unusual environment and human involvement, both of which may alter behavior. Experimental results are particularly convincing, therefore, if they agree with knowledge gained by means of observational techniques.