METHODS
Statistical methods
Choose a method by starting with the question you need to answer. Is the outcome a number or a category? Are you comparing groups, measuring an association, or making a prediction? The data and study design help narrow the choice.
A useful way to choose
First describe the outcome and how each observation was collected. Then decide what comparison or estimate you need. For example, a paired measurement calls for a different analysis from two unrelated groups. Check assumptions, missing data and sample design before interpreting a test result.
These guides use small examples, state their assumptions and explain what the result does not prove. For exploratory checks and plots, see NIST’s exploratory data analysis handbook.
WORKED EXAMPLES
Statistical method guides
New guides are added as they reach their scheduled release time.