PSYC 3000 Lecture Notes - Lecture 11: A Priori And A Posteriori, Statistical Hypothesis Testing
2 Aug 2016
School
Department
Course
Professor
Document Summary
Power is the ability to avoid type two error and way to see the true effect of a treatment. Significance, power and effect size are all inter related. Making calculation after the experiment has taken place. Now you want to find out how much power the test has. If you have insufficient power in a post hoc evaluation (<80%), then you want to take an a priori test. Tell the reader exactly what calculations made why. Very common to change sample size to change the power. Tells you what size your sample should be for a specific power. Related to cohens d. it is a parameter that is calculated that leads to power. Tells you how big your t score will be depending on your distribution. Non centrality parameter grows so does the power. Oen sample design leads to a bigger non centrality parameter. Which means a one sample t test is more powerful.
