Proportions Test
The proportions parameters test uses a Pearson's chi-squared statistic to assess whether a binomial sample has a specified proportion parameter p, or to assess whether two or more samples have the same proportion parameter.
As the proportions parameters test uses a normal approximation to the binomial distribution, it is less powerful than the exact binomial test, and hence the exact binomial test is usually preferred. The advantages of the proportions parameters test are that it provides a confidence interval for the proportions parameter and that it may be used with multiple samples.
To perform a test of proportions
Choose Statistics Compare Samples
Counts and Proportions
Proportions Parameters. The dialog shown below appears.
The Proportions Test has the following options:
Data
Data Set
Specify a data set, although this is not required (see the Options field).
Success Variable
Specify the variable that contains the observed counts of successes. Hypothesized Proportion
Trials Variable
Specify the variable that contains the corresponding number of trials.
Hypotheses
Specify the variable that contains the hypothesized values for p, the probabilities of success.
Specify the alternative hypothesis.
Options
Confidence Level
Enter the confidence level desired for the returned confidence interval. The default of 0.95 yields a 95% confidence interval.
Apply Yates' Continuity Correction
Select to correct for continuity. See the online Help for prop.test for an algebraic definition of the continuity correction.
Enter the name for the object in which to save the results of the analysis.
Print Results
Select this to print out the results of the analysis in the designated output window.
Related S-PLUS language functions
prop.test, print.htest, menuProp