A "Begin" button will appear on the left when the applet is finished
loading. This may take a minute or two depending on the speed of your internet connection
and computer. Please be patient. If no begin button appears, it is probably because
you do not have Java installed or Java is not enabled.
Starting the Applet and setting the conditions
Press the "Begin" button to start the applet in another window.
This applet simulates a two-condition experiment. You specify the number of subjects per sample (n), and the population means for each condition (Mean A and Mean B). You can also specify the correlation between the two groups in the population (rho), and the population standard deviation (sd, which is the same in each condition).
You also determine whether to use a within-subjects or a between-subjects design. Note that in a within-subjects design, the subjects participate in both conditions, so that if your sample size is 8, 8 subjects are used. In a between-subjects design, a separate group of 8 subjects is needed for each condition, so a sample size of 8 requires 16 subjects.
Also, if you choose a between-subjects design, rho is automatically set to 0.
Simulating a single experiment
A single experiment is simulated with you press the "Simulate" button. The data for the simulated experiment is graphed on the lower right side of the applet. If you are using a within-subjects design, the two scores for each subject are connected.
The formula for a t test is shown on the lower right side of the window. The lower formula has the sample mean and other statistics are filled in for you. The value of t for this sample is displayed, along with the degrees of freedom (df) and the critical value (for the .05 level). Note that the same formula is used for both the within- and between-subjects designs, but that in the between-subjects design, the value of r (correlation between the two conditions) is set to 0.
In the upper right corner, a tally is kept of the number of significant and non-significant findings for the current population settings. If the two population means are different, then the percent rejected is an estimate of power. If the population means are equal then the percent rejected is an estimate of the Type I error rate.
Simulating a large series of experiments
The "Simulate 1000" and "Simulate 10,000" buttons can be used to simulate many experiments and see how many have significant differences. Look at the differences in the percent rejected depending on changes in the population means, standard deviations, rho, and between-subjects versus within-subjects designs.