- statistical power – probability of detecting an effect
- power of statistical test is function
o of sensitivity of study
§ sample size key factor
o effect size in popukation
o criteria used to test statistical hypotheses.
- sample size
o sensitivity can be increased by
§ using more accurate sampling design
§ better measures
§ controlling noise and unnecessary variability in data
§ increase sample size
- Effect size
o how much impact explanatory variables have on outcome variable
- Decision Criterion / Statistical Significance Level
o alpha
o one vs two tail
- other factors'
o measurement
o research design
o statistical models
o missing data
Applications of Power Analysis
- 4 parameters of power analysis
o power (1-beta)
o sample size
o effect size
o significance criterion
-
No comments:
Post a Comment