PairedSamples Proportions
The PairedSamples Proportions procedure provides tests and confidence intervals for the difference in two related or paired binomial proportions. The data are assumed to be from a simple random sample, and each hypothesis test or confidence interval is a separate test or individual interval. Output includes observed proportions, estimates of differences in population proportions, asymptotic standard errors of population differences under null and alternative hypotheses, specified test statistics with twosided probabilities, and specified confidence intervals for differences in proportions.
 Example
 Statistics
 AgrestiMin, BonettPrice, Newcombe, Wald, Wald (continuity corrected), Exact Binomial, Midp Adjusted Binomial, McNemar, McNemar (continuity corrected).
Data Considerations
 Data

 A variable list containing at least two variables is required.
 If a single list of variables is specified, each member of the list is paired with every other member of the list.
 Assumptions

 If two lists of variables are separated by WITH without the (PAIRED) keyword, each member of the first list is paired with each member of the second list.
 If two lists of variables are separated by WITH and the second list is followed by (PAIRED), members of the two lists in order are paired: the first member of the first list is paired with the first member of the second list, the second members of each list are paired, etc. Unmatched variables are ignored and a warning message is issued.
Obtaining PairedSamples Proportions tests
This feature requires the Statistics Base option.
 From the menus choose:
 Select one or more quantitative test variables.
 Optionally, you can:
 Select success criteria settings under the Define
Success section:
 Last Value
 The last or highest value among the sorted distinct values in the data is used. This applies to numeric or string variables. This is the default setting.
 First Value
 The first or lowest value among the sorted distinct values in the data is used. This applies to numeric or string variables.
 Value(s)
 One or more parenthesized specific values. Multiple values must be separated by spaces. This applies to numeric or string variables. String variable values should be enclosed in single quotes.
 Midpoint
 Values at or above the middle of the range of observed values in the data. This applies only to numeric data.
 Cut Point
 Values at or above a specified value. This applies only to numeric data.
 Click Confidence Intervals... to specify which types of confidence intervals are displayed, or to suppress all confidence intervals.
 Click Tests... to specify which types of test statistics are displayed, or to suppress all tests.
 Click Missing Values... to control the treatment of missing data.
 Click Bootstrap... for deriving robust estimates of standard errors and confidence intervals for estimates such as the mean, median, proportion, odds ratio, correlation coefficient or regression coefficient.
 Select success criteria settings under the Define
Success section:
 Click OK.
This procedure pastes PROPORTIONS command syntax.