Friday, March 11, 2011

HB of Survey Research: Chapter 7: Sample Size, Optimum Allocation and Power Analysis


-          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: