The program (STTT.EXE) may be downloaded. Also by the same author: MEDSEQ, SSD and TABCHI
Also: Group sequential experiment plans a.m. Pocock: An adaptation for medicine, with power trials using computer simulation
Student's t-test
Author: Egil H. Lehmann, GP, Haugesund, Norway
Email: egilhl@online.no
Welcome to a program which I hope you will like. To profit from it, one should be familiar with the standard deviation as a statistical tool, and the meaning of 'statistical significance' and P value. Brief explanations are given within the program. Also recommended is the book Statistics at square one by TDV Swinscow, published by the British MedicalAssociation, chapters two to seven.
Student's t test is probably still the most popular of all statistical tests. The test is a classic, with an interesting history.
- Simple hypothesis: The test compares two mean values, to judge if the two groups from which the averages were calculated, could stem from a common population. Random selection is postulated.
- A more general hypothesis may also be tested: That the two samples stem from two populations where the difference (Mean of population B minus Mean of population A) has a SET VALUE which may be ZERO, POSITIVE or NEGATIVE.
- For input is needed data computed from the observations of the two groups: The two sample sizes, the respective mean values, and the respective standard deviations. If more practical, instead of the SDs, the program will accept the respective standard errors of the mean - SEMs.
- If you have only raw data, the present program will help you compute mean values, SDs and SEMs.
If the measurements are skew, i.e. tend to cluster in the upper or lower range, it pays off to use a rank test. Wilcoxon's two-sample test (also performed as Mann-Whitney's U-test) is recommended. It is in fact a Student's t-test performed on ranks. Ranks are order numbers after measurements have been sorted by magnitude. The smallest measurement is given rank 1, and so on.
The present program offers this test in a computer friendly version. If the sum of the sample sizes are 30 or smaller, the calculated p values are only approximations. The user is then recommended to find p values by hand in a table, using the rank sums computed by the program. (A reminder will be given by the program.)
This is an unregistered program copy. Copying and all use of this program is free. If you find the program useful, letting the author know would encourage further work. Most places in the program, entering a negative number, zero or blank leads to immediate exit. Please read codelists carefully - and good luck!
Download Program (STTT.EXE)
Home page Department
Home page University
Department of Public Health and Primary Health Care, last updated 14.12.00
Hogne.Sandvik@isf.uib.no