1. What's so bad about making a Type I error?
2. How does the Bonferroni adjustment control for Type I error when many statistical tests are conducted?
3. Why do researchers use ANOVA rather than t-tests to analyze data from experiments that have more than two groups?