Statistics

Question 1

Type 1 error

Answer 1

Concluding there is a significant difference when there is none

Question 2

Type 2 error

Answer 2

Concluding there isn’t a significant difference where there IS one

Question 3

Intention-to-treat vs per-protocol analysis

Answer 3

Intention-to-treat analysis looks at all data even if did not complete entire study => minimize compliance bias VS per-protocol analysis looks only at people who completed the study