Power

Internet
Demonstrations

Up
Power Answer

You can link to a power applet to determine how the power of a hypothesis test varies with effect size and also with sample size.

• Click on Introduction to Statistical Power.
• Read Simple Instructions for using the power applet.
• Notice that the default values specify a null (hypothesized) population mean of 100; an alternative (true) population mean of 115; and a one-tailed hypothesis test, upper tail critical, at the 0.05 level of significance. Also notice that sectors under the hypothesized sampling distribution are shown as shades of blue, while those under the true sampling distribution are shown as shades of red.
• Increase and decrease power (the pink sector under the true sampling distribution) by clicking on and dragging the true sampling distribution to the right and left.
• Next, for a fixed difference between hypothesized and true population means (effect size), observe how power changes with increases in sample size (by clicking on the n thermometer).
• When finished, answer the following question.

Question: How does power vary with increases in effect size and with increases in sample size?  Answer.