Medical Education Online Resource Section (Software)
This section contains software resources.
Statistical Power Calculator
This program will perform power calculations for six common statistical tests:
Two independent sample t-test
This program is written in the 32 bit version of Visual Basic and will ONLY RUN UNDER WINDOWS 95 or 98. All calculations are performed in double precision based on algorithms obtained from power tables in Hulley SB, Cummings SR (Eds ) Designing Clinical Research. 1988, Williams & Wilkins, Baltimore. The program has been carefully checked against tabled power values but THE AUTHOR MAKES NO GUARANTY AS TO THE ACCURACY OF THE CALCULATIONS.
The program is copyrighted by David J. Solomon, Ph.D., but may be freely distributed at no charge.
The installation files for the program are contained in a self-extracting archive, R0000007.EXE. Download the archive into an empty folder (subdirectory). Execute the program (archive) either from the "Run Command" on the Start button or by double clicking on it from the Windows Explorer. This will unpack the archive. Once unpacked, there should be a file SETUP.EXE in the folder. Execute this program which will install the Statistical Power Calculator. Follow the instructions on the screen.
When completed, the Statistical Power Calculator should be listed among the programs in the Program List accessed from the Start Button. At that point you can get rid of folder with the installation files. The program can be uninstalled and removed by Add/Remove Programs button in the Windows 95 Control Panel.
Please send comments, bug reports and suggestions to David J. Solomon, Ph.D. via electronic mail at
Note: The archive is just under two megabytes in size and may take an extended period to download.
Press here to download the Statistical Power Calculator.
This program will estimate the inter rater reliability of a set of ratings based on a formula presented by Ebel:
Ebel RL. Estimation of the reliability of ratings. Psychometrika 1951;16:407-424.
It generates what is essentially an intraclass correlation and is very flexible in that it makes no assumptions about the number of judges rating each person or object rated nor which judges rate each person or object. All that is required are an alphanumeric ID indicating the person or object rated and a numeric rating, one rating per line in a standard ASCII file comma delimited.
This program is written in PHP and is available for public use at:
The program has been checked against examples provided by Ebel and tried on a number of other data sets. THE AUTHOR HOWEVER MAKES NO GUARANTY AS TO THE ACCURACY OF THE CALCULATIONS.
The program is free software distributed under the GNU General Public License and the source code may be downloaded at:
The program consists of three files,
The program will also create a scratch file "tmp.text" which gets overwritten with the rating data each time the program is run.
Installation notes: The program requires that PHP and the must be versions of PHP 3 after PHP 3.0.16, or versions of PHP 4 after 4.0.2. Be sure to place all three files in the same directory as the hyperlinks are relative.
Please send comments, bug reports and suggestions to David J. Solomon, Ph.D. at
By Karl Dockray, M.D.
AN ELECTRONIC PEARL-BOOK FOR CLASS-NOTES AND THE FACTOIDS OF
MEDICAL TRAINING THAT COULD BE CARRIED INTO PRACTICE
From: A*A DATA, 1808 19th St. Lubbock, TX 79401, USA Free to students,
residents, registrars, teachers and researchers. (Commercial apps (DOCU*MENTOR)
The self-extracting archive is approximately 1.57 MB and may take a substantial amount of time to download.
Press here to download STAT*WRITER
Press here for detailed installation instructions.
Web Survey Mailer System (WSMS)
WSMS is an integrated system for conducting Web-based surveys. The system will send out personalized cover letter emails requesting respondents complete a Web-based survey with the URL to the survey embedded in the email. The system tracts whether respondents have completed the survey while maintaining the anonymity of their responses. It will send out follow-up reminders to only those subjects who have not completed the survey and blocks access to the survey by people who have not been sampled as well as only allowing a sampled respondent to complete the survey once. Survey data are written to a tab-delimited ASCII data file that can be easily be downloaded and analyzed with most common data bases and statistical packages.
WSMS consists of a series of HTML documents and PHP scripts. The respondents' names and email addresses are stored in a MySQL database. PHP and MySQL are widely available open source packages and must be installed on the Web server on which the WSMS is used. Documentation is provided and the scripts are heavily commented. A user however must make some modifications in the PHP scripts and setup the MySQL data base to use the system. A user will likely need a good working knowledge of HTML and some background in PHP or a similar server-based programming language such as Pearl to make the necessary modifications and install the system.
WSMS is distributed under the terms of the GNU General Public License as published by the Free Software Foundation. Information about the GNU General Public License can be found at: http://www.gnu.org.
The link below will download a self-extracting archive of the WSMS system and documentation. Send comments and suggestions for enhancements to David J. Solomon, Ph.D. at firstname.lastname@example.org. I am sorry but I can not provide help for installing WSMS on your server.
This disclaimer is not meant to sidestep the responsibility for the material we will share with you, but rather is designed to emphasize the purpose of the Resource Section, which is to provide information for your own purposes. The submissions contained in the Resource Section have been chosen for their usefulness to medical educators. The material has been provided by other medical educators for distribution through MEO's Resource Section. The information should not be considered to be completely error-free or to include all relevant information; nor should it be used as an exclusive basis for the purpose for which it is being provided. The user understands and accepts that if the Journal were to accept the risk of harm to the user from use of this material, it would not be able to make the material available because the cost to cover the risk of harms to all users would be too great. Thus, use of the information is strictly voluntary and at the user's sole risk.