|
Activity Number:
|
65
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #306294 |
|
Title:
|
Simple Power Calculations: How Do We Know We Are Doing It the Right Way?
|
|
Author(s):
|
Michael Vorburger*+
|
|
Companies:
|
RTI International
|
|
Address:
|
3040 Cornwallis Road, Research Triangle Park, NC, 27709,
|
|
Keywords:
|
power analysis ; sample size ; minimum detectable difference ; arcsine transformation
|
|
Abstract:
|
Sample size and power determination is a crucial step in setting up efficient studies. It ensures that there are enough sampling units to enable the detection of anticipated effects. After determining the hypothesis test on a parameter, the significance level, estimates of parameters needed to compute the power function; the researcher explores several scenarios by considering different power levels and effect size (also known as minimum detectable difference). We are interested in the problem of estimating power and the effect size for specific sample sizes when assessing the difference between two proportions. We compare the accuracy of alternative methods and discuss the applicability of these results for practitioners.
|
