Statistics graphite how to stroke it

Interpret all statistics and graphs for Selectivity and sample size for testing proportions, 2 samples


You can use the power curve to determine the appropriate sample size or power for a test.

The power curve represents any combination of power and comparative percentage for each sample size, if the significance level is kept constant. Each symbol on the power curve represents a value calculated based on the values ​​entered. For example, if you enter a value for sample size and a value for power, Minitab calculates the corresponding comparative percentage and displays the calculated value on the graph.

Examine the values ​​on the curve to determine the difference between the comparative proportion and the proportion that can be seen at a given power and sample size. A power value of 0.9 is generally considered suitable. On the other hand, some experts consider a power level of 0.8 to be suitable. If a hypothesis test has insufficient power, a difference with practical significance may not be able to be detected. As you increase the sample size, the power of the test also increases. It is recommended that there are a sufficient number of observations in the sample to achieve adequate power. However, the sample size should not be so large that time and money are wasted in sampling or that irrelevant differences are recognized as statistically significant. If you reduce the size of the difference to be detected, the selectivity also decreases.

In this graphic, the power curve shows that with a sample size of 1000 and a power of 0.8, the test can detect a difference between the comparative proportion and the proportion (p0) of approximately 0.006 in both directions. The selectivity curve shows that the difference detected by the test approaches 0.007 when the selectivity is increased to 0.9.