Statistics Definitions > Beta Level
What is a Beta Level?
Alpha levels and beta levels are related:
- An alpha level is the probability of a type I error, or rejecting the null hypothesis when it is true.
- A beta level, usually just called beta(β), is the opposite*; the probability of of accepting the null hypothesis when it’s false. Note that we informally say “accept the null,” but in strict statistical terms we only “fail to reject” the null, because it’s rarely proven true.
You can also think of beta as the incorrect conclusion that there is no statistical significance (if there was, you would have rejected the null). Note that rejecting or failing to reject the null hypothesis (i.e., statistical significance) does not necessarily speak to how practically important the findings are in a real-world context.
*Informally, we often say that beta is “the opposite” of alpha. However, strictly speaking, they are different error probabilities—alpha concerns rejecting a true null (Type I), and beta concerns failing to reject a false null (Type II). But it’s common to compare them as opposites because both are directly tied to test performance.
Beta and Power
Beta is directly related to the power of a test. Power relates to how likely a test is to distinguish an actual effect from one you could expect to happen by chance alone. Beta plus the power of a test is always equal to 1. Usually, researchers will refer to the power of a test (e.g. a power of .8), leaving the beta level (.2 in this case) as implied.
How do I Lower Beta?
In theory, the lower beta, the better.
You could increase the power of a test to lower the beta level. However, there’s an important trade-off. Alpha and beta levels are connected: for a fixed sample size and effect size, lowering one will raising the level of the other. For example, a Bonferroni correction reduces the alpha level (i.e. the probability of making a type I error) but inflates the beta level (the probability of making a type II error). False positives are minimized, but with the payoff that the possibility of false negatives are increased.
In practice, researchers often try to mitigate this trade‐off by increasing sample size, adjusting study design, or choosing an appropriate effect size to detect, so that both alpha and beta remain acceptably low. Many fields use a 5% alpha levels and aim for a power of .80 (i.e., a beta level of .20). Some fields, such as clinical trials, require higher power (0.90 or above).