Use the significance level to decide whether an effect is statistically significant. Because the significance level is the threshold for statistical significance, a higher value also increases the chance of making a type I error. A type I error is the incorrect conclusion that an effect is statistically significant.
Usually, a significance level (denoted as α or alpha) of 0.05 works well. A significance level of 0.05 indicates that the risk of concluding that an effect exists—when, actually, no effect exists—is 5%.
- Choose a higher significance level, such as 0.10, if you are willing to increase the risk of declaring that an effect is statistically significant—when, actually, no effect exists—so that you have greater power to detect an effect that is important. For example, a chemical engineer designs an experiment to study the effect of 12 factors on the yield of a substance. The engineer prefers to consider small or insignificant effects further rather than to remove a factor that could be important. Therefore, the engineer chooses a significance level of 0.10 to be more certain of detecting a factor that is important.
- Choose a lower significance level, such as 0.01, to be more certain that you do not conclude that an effect exists when no effect exists. For example, a scientist at a pharmaceutical company designs an experiment to study the effect of 24 factors on a new drug. For further experimentation, the scientist prefers to study only factors that are important. The scientist chooses a significance level of 0.01 to be more certain not to conclude that an effect that does not exist is statistically significant.