Knowledge of the major methods that have been used in sex research and of the problems and merits associated with each is necessary for understanding and evaluating sex research.

Ideally, sex research should employ probability sampling techniques.

Large-scale surveys of sexual behavior generally rely on people's self-reports, which may be inaccurate because of purposeful distortion, problems of memory, or an inability to estimate some of the information requested. Direct observations of sexual behavior avoid these problems, but they lead to an even more restricted sample. They also answer questions that are somewhat different from those answered by surveys. Web-based surveys offer new opportunities for sex research.

In all behavioral research, the ethical principles of informed consent, protection from harm, and justice must be observed, although historically some sex researchers did not do this. One major sex survey was Kinsey's large-scale interview study of the sexual behavior of Americans, done during the 1940s. The NHSLS is a recent, large-scale survey of the sexual behavior of Americans; it was based on probability sampling. Other large surveys have been done, including some through magazines; the samples in magazine surveys, however, are so restricted that we cannot draw general conclusions from them.

Studies of special populations include Coxon's daily diary study of the sexual behavior of gay men.

In media content analysis, researchers use systematic coding categories to analyze what is represented in the media, such as on television, in romance novels, or in magazine ads. In participant-observer studies, the scientist becomes a part of the community to be studied and uses a combination of direct observations and interviewing. Examples are studies of sexual behavior in other cultures and Humphreys's study of the tearoom trade.

In experimental sex research, the goal is to discover what factors cause or influence various aspects of sexual behavior. The researcher manipulates an independent variable and measures a dependent variable.

The following statistical terms were introduced: mean, median, variability, incidence, frequency, and correlation.