The only way for a research method to determine causality is through a properly controlled experiment. Although other methods may clarify or complement an experiment, they cannot replace an experiment in determining causality. To determine causality, the scientist designs an experimental study to test a hypothesis and uses two or more conditions (variables) to compare. High-quality experiments control for other, sometimes hidden, variables that could account for the experimental results.

For example, suppose the scientist hypothesizes that exercise can reduce blood pressure. They set up two conditions, one in which participants exercise and a second condition in which they do not. After a period of time, the scientist then measures their blood pressure to see if the exercise group had lower blood pressure. To improve the quality of the experiment and increase the confidence that exercise causes reduced blood pressure, the scientist could measure blood pressure before the experiment begins. Imagine if the exercise group already had lower blood pressure! Additionally, the scientists should control for other variables that could account for the results, which in this case would be diet and other medications. Both groups should eat the same diet throughout the study to confirm that that is not the reason for any difference in blood pressure. They should also make sure none of the participants are on medication for their blood pressure and perhaps control (as best they can) for other lifestyle factors such as stress. This would yield the most accurate results in determining causation.