This webpage contains download links for scripts that enable you to compute confidence intervals and power based on the noncentral t, F, and Chi-square distributions. So far, these are available in SPSS, SPlus, and R. I will update this webpage with links to other freeware and/or scripts as they become available and known to me.
A general (but not expensive!) reference on confidence intervals that covers noncentral confidence interval estimation, as well as the more well-known confidence intervals, is:
Smithson,
M.J. (2003). Confidence
Intervals,
Quantitative Applications in the Social Sciences Series, No. 140.
Thousand Oaks, CA: Sage.
There is a PDF file based on a short workshop I’ve given on noncentral confidence intervals that includes instructions on how to use these scripts. This will have to do as an instruction manual until I write a more thorough-going version.
Download PDF file.
Noncentral F files |
Noncentral t files |
Noncentral Chi-square files |
Download NoncF3.sps |
Download NoncT2.sps |
Download NoncChi.sps |
Download NoncF.sav |
Download NoncT.sav |
Download NoncChi.sav |
Download F2R2.sps |
Download T2D.sps |
|
These scripts come with a "Readme" PDF file. Users not familiar with noncentral confidence intervals may also wish to download the workshop instructional PDF file mentioned above.
Download Readme file.
Noncentral F file |
Noncentral t files |
Noncentral Chi-square files |
Download NoncF2.sas |
Download Nonct.sas |
Download Noncchi.sas |
Download t2d1samp.sas | Download CramersV.sas | |
|
Download t2d2samp.sas |
These scripts come with a "Readme" PDF file. Users not familiar with noncentral confidence intervals may also wish to download the workshop instructional PDF file mentioned above.
Download Readme file.
Noncentral F file |
Noncentral t files |
Noncentral Chi-square file |
Download Fscript.SSC |
Download Nonct.SSC |
Download Chiscript.SSC |
|
Download Rknonct.SSC |
Some great software that runs under Excel dealing with confidence intervals in general, and also with confidence intervals for effect sizes, is the Exploratory Software for Confidence Intervals (ESCI--pronounced "es-key") software developed by Geoff Cumming:
Jim Steiger and Rachael Fouladi (1992) have provided very useful program called "R2" that computes confidence intervals and power for multiple and partial R-squared. This program runs in DOS and can be downloaded from:
http://www.interchg.ubc.ca/steiger/homepage.htm
For Statistica users, information about an add-on module for computing noncentral confidence intervals and power may be found at:
http://www.statsoft.com/power_an.html