The Four Horsemen

No, it’s not the latest Hollywood action movie. It refers to the four factors that influence the outcome of a research study — Type 1 error risk, statistical power, effect size, and sample size.

If you keep two Horsemen constant, increasing the third causes the fourth to decrease, and vice versa. For example, if we keep our Type I error rate and sample size constant, then if we decrease the effect size, we increase the changes of a Type II error (by decreasing statistical power). In other words, hold two Horsemen constant and you can seesaw the remaining two.

Why does this matter? Because one of the fundamentals of planning a study is conserving resources and maximizing the likelihood of detecting an effect if there is one.

Still confused? Sign up for Essentials of Statistical Analysis (EOSA): Complete, EOSA: Part 1, EOSA: Part 2, or EOSA: Part 3 to learn about this and other basics to prepare you for a full course in statistics.