Statistical Power and Sample Size

Statistical power and sample size guidance and theory

Good general and understandable introduction to the concept of power.
http://www.socialresearchmethods.net/kb/power.htm
Good introduction from the U.S. Environmental Protection Agency.
http://www.epa.gov/bioindicators/statprimer/power.html
Russ Lenth's more advanced (and opinionated) discussion of how to conduct power and sample size calculations. Includes an interactive calculator that requires the Java plug-in (which may not be installable if you are on a networked computer).
http://www.stat.uiowa.edu/~rlenth/Power/index.html
Hypertext statistics textbook. If you like hypertext links, then you are going to love this book. Too many ads for my tastes.
http://davidmlane.com/hyperstat/
Statistics at Square One. An excellent introduction from the BMJ.
http://www.bmj.com/statsbk/

Worth mentioning again: Russ Lenth's interactive power/sample size calculator is not only helpful in designing your study but also educational in displaying the important determinants of statistical power. Does require the Java plug-in. http://www.stat.uiowa.edu/~rlenth/Power/index.html

Paired t-test of means (before/after comparison in the same subjects).
http://www.dssresearch.com/toolkit/spcalc/power_a1.asp

Paired comparison of a difference in proportions (before/after comparison in the same subjects).
http://www.dssresearch.com/toolkit/spcalc/power_p1.asp

Versatile calculator for means and proportions. It is absolutely necessary to read the online instructions. Not mentioned in the online instructions is the need to include a trailing zero when entering a proportion between 0 and 1 (i.e., enter 0.3, not .3). Power calculator: http://www.quantitativeskills.com/sisa/calculations/power.htm Sample size calculator: http://www.quantitativeskills.com/sisa/calculations/ samsize.htm Instructions: http://www.quantitativeskills.com/sisa/calculations/ powhlp.htm

Sample size and power for multiple linear regression analysis that requires an understanding of R2.
http://www.danielsoper.com/statcalc/

Sample size/power/minimum detectable difference for parallel and cross-over clinical trial designs with dichotomous, continuous, or failure-time outcomes. Not very user friendly but it does work once you get the hang of it.
http://hedwig.mgh.harvard.edu/samplesize/size.html

PS (PC only) is a super program with an easy to understand windows interface that performs basic power and sample size calculations in addition to more advanced procedures, such as regression and Cox proportional hazards analysis. *Warning* - there is a bug in installing the program. When asked to specify a directory, press "Enter" instead of clicking the menu button. http://biostat.mc.vanderbilt.edu/twiki/bin/ view/Main/PowerSampleSize

G*Power, versatile but user-unfriendly. If you liked DOS, you will love this. Also available in a Mac version, but unfortunately it does not work on newer models (OS X).
http://www.psycho.uni-duesseldorf.de/aap/projects/gpower/