PSYC2009 Study Guide - Quiz Guide: Effect Size, Type I And Type Ii Errors, Null Hypothesis

42 views2 pages
14 Jun 2018
School
Department
Course
Professor
Effect Size and Power
Statistical power:
Likelihood of detecting an effect/significant difference
Higher power decreases the likelihood of committing a type II error
Determining power of a study:
1. Determine CL
2. Find the z-score that corresponds to an area beyond z that is equivalent to the level (use
tables in back of textbook, or standard z-score table online):

3. Convert z-score into units of the original scale:
4. Add X to the null hypothesis value to find the minimum sample mean that would allow
rejection of null hypothesis
5. Calculate z-score of this value relative to the distribution of the sample mean
6. Determine the area below Z for using Table A.2.
o This is the power
Power can be increased by obtaining larger samples or lowering CL in order to make the CI
narrower
o Larger samples preferable: can increase power without increasing risk of type I error
Can work out how big the sample needs to be in advance once CL is selected and desired
power is determined as well as the smallest possible effect-size that needs to be detected
o Effect size calculated using
and N
o Rearranged to make d subject:

o Rearranged again to make N the subject: this will determine exactly how many
participants needed in order to reach requirements for power and effect size
o Is determined by establishing tge desired power and then finding the corresponding z-
score in Table A.2 with an area beyond z of
Effect size:
How large the difference is between the null hypothesis value and the sample mean
Effect size helps us to understand how meaningful the result is
Measured relative to the range of the scale (scale-based evaluation) or relative to dispersion of
data (distribution-based evaluation)
o Scale-based: measured in same units as dependant variable
o Distribution-based: measured in terms of standard deviations
Cohen's d
Distribution based evaluation
Gives the difference between the sample mean and the population parameter in standard
deviation units
o 
find more resources at oneclass.com
find more resources at oneclass.com
Unlock document

This preview shows half of the first page of the document.
Unlock all 2 pages and 3 million more documents.

Already have an account? Log in