An alpha risk is the risk of concluding that something is significant when in fact it really is not.

Use: Understanding the concept of the alpha risk is important when making decisions based upon the results of statistical tests. There is always a risk that a statistical finding leading to a decision (e.g., and F-test, a t-test, a chi-square test, etc.) is found to be significant even though in reality it is not significant. Let’s say that two methods are tested. Statistical data is collected for both method A and method B. And, let’s say that it is determined through the use of sound testing methods that method B is found to be significantly better than method A.

The team concludes that the two methods were significantly different, so based upon this finding, the decision is made to install method B and get rid of method A. Six months later, to the surprise of the team, it turns out that in reality method B is actually no better than method A at all. The two methods tested are not significantly different. That is known as a Type I error. A Type I error is to conclude that something is significant when in fact, it is not significantly different. The risk that a team makes when concluding that something is significant (based upon some statistical test), is that there is a small chance that they are going to make a Type I error. The risk involved is referred to as ‘the alpha risk.’ Typical alpha risks are 0.10, 0.05, and 0.01 that correspond to a 90%, 95%, and 99% level of confidence respectively.

While we’re on the topic, a beta risk is to conclude that something is not significantly different, when in fact, it really is. If that were to happen, the team will have made a Type II error.

Six Sigma Terminology