Health Sciences 3801A/B Lecture Notes - Lecture 8: Type I And Type Ii Errors, Bonferroni Correction, Analysis Of Variance

• School: Western University
• Department: Health Sciences
• Course: Health Sciences 3801A/B
• Professor: Andrew Johnson

Document Summary

Doing multiple t-tests increases the type i error rate. The alpha gets inflated dramatically - unacceptable to use. This creates an adjustment to the alpha. This still has a significant risk of error. You want to demonstrate that there is a significant treatment effect. Split the variance into the treatment and error variance. Use f-ratios to show this, then determine if there"s a statistically significant difference. As treatment variance gets larger, the f-ratio increases w/ it. The larger the f-ratio is, the more likely that the treatment effect is statistically significant. Not comparing the difference - you want to know if there"s a difference. You can"t belong to more than one group. Normality ensures that there is a normal distribution within each of the samples. When the group sizes are equal, there isn"t a significant effect on type i and type ii error rates. There should be similar variances b/w the groups. Use the sum of squares to determine that.