1-Sample t-test

STAT>BASIC STATISTICS>1 SAMPLE t

This dialog box shows an example using summarised data. Use 10 as the hypothesised mean $$\mu_{0}$$, so that $$H_{0}: \mu = 10$$. Use 'Options' to change the default settings for the test.

One-Sample T Test of mu = 10 vs not = 10  N   Mean  StDev  SE Mean       95% CI          T      P 15  9.630  1.200    0.310  (8.965, 10.295)  -1.19  0.252

Using $$\alpha = 0.05$$, we would fail to reject $$H_{0}$$ because the p-value > 0.05.

Power of the test can be found by using:

STAT>POWER AND SAMPLE SIZE>1 SAMPLE t

In this example, the dialog uses power values of 0.8, 0.9 and 0.95 (power = probability of correctly rejecting $$H_{0}$$. 'Options' can be used to change the default settings, including $$\alpha$$.

 

The graph and session window output show the power for a given difference.

A 0.934 standard deviation difference can be detected,using 15 observations, with 80% power. i.e. there is an 80% chance of detecting this difference.

Power and Sample Size 1-Sample t Test Testing mean = null (versus not = null) Calculating power for mean = null + difference Alpha = 0.05  Assumed standard deviation = 1.2 Sample   Size  Power  Difference     15   0.80     0.93363     15   0.90     1.08105     15   0.95     1.20295

 

contents

close this window