pctdiff is a program that allows you to create two sample proportions, at different stages of the randomizing cycle. The observed proportions are subtracted and compared to the population proportions. The program displays a Z score, based on the net difference between the samples and the population.
The purpose of this program is to show that, the randomizing routine is consistent at different stages of the randomizing cycle.
pctdiff has the following parameters:
Examples of how to run the program.
pctdiff -2 -a 0.5 -b 0.4 -m 100 -n 400 -c 95.0
pctdiff -1 -t ut -a 0.5 -b 0.4 -m 100 -n 400 -c 90.0
In the first example, the program makes 100 observations for sample 1 with a probability of 50 percent and a confidence level of 95 percent. The program makes 400 observations for sample 2 with a probability of 40 percent and a confidence level of 95 percent. This test is a two tail test.
In the second example, the program makes 100 observations for sample 1 with a probability of 50 percent and a confidence level of 90 percent. The program makes 400 observations for sample 2 with a probability of 40 percent and a confidence level of 90 percent. This test is a one tail test based on the upper tail of the normal curve.
The calculation of the Z score for the comparison is based on the following book:
Lincoln L. Chao
Statistics for Management
2nd Edition
Palo Alto: The Scientific Press, 1984
Chapter 10, Hypothesis Testing
Section 5, Testing for the difference between two proportions
Lincoln L. Chao published his book while at California State University, Long Beach.