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.