As part of the analysis, investigators should first describe their study sample or patient population. Descriptive analysis involves descriptive statistics, which are numbers that summarize data with the purpose of describing the study sample or patient population. Frequency distribution, mean, median, mode, range, and standard deviation are the most used statistics. Characteristics of the study sample or patient population that are important to describe are commonly demographic (e.g., age), clinical (e.g., health conditions), and geographic (e.g., rural/urban) patterns. Descriptive analysis summarizes data that would be useful information for a specific topic area and/or program planning, since it constitutes describing the study sample or patient population.

Inferential analysis involves inferential statistics, which are based on probability theory and the logic of hypothesis testing. Because researchers rarely have access to an entire population, inferential statistics allow researchers to use a sample to draw inferences about the larger population of interest. Therefore, inferential statistics rely on appropriate sampling methods, adequate sample size, and adherence to statistical assumptions.

There are three basic categories of inferential analysis that differ based on their intended purpose, including: evaluating differences (e.g., t-tests), examining relationships (e.g., correlations), and making predictions (e.g., simple or multiple regression analysis). Researchers should keep in mind that the type of data determines the type of test (e.g., parametric vs. non-parametric).

Inferential analysis used in primary care research commonly answers questions about the effect of interventions or changes to clinical practices on patient outcomes, as opposed to research questions on the causes of a health condition or disease.