What is the "significance level" of a statistical test, and how does it relate to our decision about whether an independent variable had an effect? How does our choice of significance level impact the likelihood of making a Type I and a Type II error?