The FREQ Procedure

specifies the name of the output data set. When you use an OUTPUT statement but do not use the OUT= option, PROC FREQ creates a data set and names it by using the DATAn convention.

specify the statistics to include in the output data set. Table 3.7 lists the output-options that are available in the OUTPUT statement, together with the TABLES statement options that are required to produce the statistics. Descriptions of the output-options follow the table in alphabetical order.

You can specify output-options to request individual statistics, or you can request groups of statistics by using output-options that are identical to the group options in the TABLES statement (for example, the CHISQ, MEASURES, CMH, AGREE, and ALL options).

When you specify an output-option, the output data set includes statistics from the corresponding analysis. In addition to the estimate or test statistic, the output data set includes associated values such as standard errors, confidence limits, p-values, and degrees of freedom. For more information, see the section Contents of the OUTPUT Statement Output Data Set.

To store a statistic in the output data set, you must also request computation of that statistic with the appropriate TABLES, EXACT, or TEST statement option. For example, the PCHI output-option includes the Pearson chi-square in the output data set. You must also request computation of the Pearson chi-square by specifying the CHISQ option in the TABLES statement. Or, if you use only one TABLES statement, you can request computation of the Pearson chi-square by specifying the PCHI or CHISQ option in the EXACT statement. Table 3.7 lists the TABLES statement options that are required to produce the OUTPUT data set statistics.

includes the following tests and measures of agreement in the output data set: McNemar’s test (for tables), Bowker’s symmetry test, the simple kappa coefficient, and the weighted kappa coefficient. For multiway tables, the AGREE option also includes the following statistics in the output data set: overall simple and weighted kappa coefficients, tests for equal simple and weighted kappa coefficients, and Cochran’s Q test.

The AGREE option in the TABLES statement requests computation of tests and measures of agreement. For more information, see the section Tests and Measures of Agreement.

AGREE statistics are computed only for square tables, where the number of rows equals the number of columns. PROC FREQ provides Bowker’s symmetry test and weighted kappa coefficients only for tables larger than . (For tables, Bowker’s test is identical to McNemar’s test, and the weighted kappa coefficient equals the simple kappa coefficient.) Cochran’s Q is available for multiway tables.

includes the continuity-adjusted chi-square in the output data set. The continuity-adjusted chi-square is available for tables and is provided by the CHISQ option in the TABLES statement. For more information, see the section Continuity-Adjusted Chi-Square Test.

includes all statistics that are requested by the CHISQ, MEASURES, and CMH output-options in the output data set. ALL also includes the number of nonmissing observations, which you can request individually by specifying the N output-option.

includes the Breslow-Day test in the output data set. The Breslow-Day test for homogeneity of odds ratios is computed for multiway tables and is provided by the CMH, CMH1, and CMH2 options in the TABLES statement. For more information, see the section Breslow-Day Test for Homogeneity of the Odds Ratios.

includes the binomial proportion estimate, confidence limits, and tests in the output data set. The BINOMIAL option in the TABLES statement requests computation of binomial statistics, which are available for one-way tables. For more information, see the section Binomial Proportion.

includes the following chi-square tests and measures in the output data set for two-way tables: Pearson chi-square, likelihood ratio chi-square, Mantel-Haenszel chi-square, phi coefficient, contingency coefficient, and Cramér’s V. For tables, CHISQ also includes Fisher’s exact test and the continuity-adjusted chi-square in the output data set. For more information, see the section Chi-Square Tests and Statistics. For one-way tables, CHISQ includes the chi-square goodness-of-fit test in the output data set. For more information, see the section Chi-Square Test for One-Way Tables. The CHISQ option in the TABLES statement requests computation of these statistics.

If you specify the CHISQ(WARN=OUTPUT) option in the TABLES statement, the CHISQ option also includes the variable WARN_PCHI in the output data set. This variable indicates the validity warning for the asymptotic Pearson chi-square test.

includes the following Cochran-Mantel-Haenszel statistics in the output data set: correlation, row mean scores (ANOVA), and general association. For tables, the CMH option also includes the Mantel-Haenszel and logit estimates of the common odds ratio and relative risks. For multiway (stratified) tables, the CMH option includes the Breslow-Day test for homogeneity of odds ratios. The CMH option in the TABLES statement requests computation of these statistics. For more information, see the section Cochran-Mantel-Haenszel Statistics.

If you specify the CMH(MANTELFLEISS) option in the TABLES statement, the CMH option includes the Mantel-Fleiss analysis in the output data set. The variables MF_CMH and WARN_CMH contain the Mantel-Fleiss criterion and the warning indicator, respectively.

includes the CMH statistics in the output data set, with the exception of the row mean scores (ANOVA) statistic and the general association statistic. The CMH1 option in the TABLES statement requests computation of these statistics. For more information, see the section Cochran-Mantel-Haenszel Statistics.

includes the CMH statistics in the output data set, with the exception of the general association statistic. The CMH2 option in the TABLES statement requests computation of these statistics. For more information, see the section Cochran-Mantel-Haenszel Statistics.

includes the Cochran-Mantel-Haenszel correlation statistic in the output data set. The CMH option in the TABLES statement requests computation of this statistic. For more information, see the section Correlation Statistic.

includes the Cochran-Mantel-Haenszel general association statistic in the output data set. The CMH option in the TABLES statement requests computation of this statistic. For more information, see the section General Association Statistic.

includes the Cochran-Mantel-Haenszel row mean scores (ANOVA) statistic in the output data set. The CMH option in the TABLES statement requests computation of this statistic. For more information, see the section ANOVA (Row Mean Scores) Statistic.

includes Cochran’s Q test in the output data set. The AGREE option in the TABLES statement requests computation of this test, which is available for multiway tables. For more information, see the section Cochran’s Q Test.

includes the contingency coefficient in the output data set. The CHISQ option in the TABLES statement requests computation of the contingency coefficient. For more information, see the section Contingency Coefficient.

includes Cramér’s V in the output data set. The CHISQ option in the TABLES statement requests computation of Cramér’s V. For more information, see the section Cramér’s V.

includes the test for equal simple kappa coefficients in the output data set. The AGREE option in the TABLES statement requests computation of this test, which is available for multiway, square () tables. For more information, see the section Tests for Equal Kappa Coefficients.

includes Zelen’s exact test for equal odds ratios in the output data set. The EQOR option in the EXACT statement requests computation of this test, which is available for multiway tables. For more information, see the section Zelen’s Exact Test for Equal Odds Ratios.

includes the test for equal weighted kappa coefficients in the output data set. The AGREE option in the TABLES statement requests computation of this test. The test for equal weighted kappas is available for multiway, square () tables where r > 2. For more information, see the section Tests for Equal Kappa Coefficients.

includes Fisher’s exact test in the output data set. For tables, the CHISQ option in the TABLES statement provides Fisher’s exact test. For tables larger than , the FISHER option in the EXACT statement provides Fisher’s exact test. For more information, see the section Fisher’s Exact Test.

includes the gamma statistic in the output data set. The MEASURES option in the TABLES statement requests computation of the gamma statistic. For more information, see the section Gamma.

includes the Gail-Simon test for qualitative interaction in the output data set. The CMH(GAILSIMON) option in the TABLES statement requests computation of this test. For more information, see the section Gail-Simon Test for Qualitative Interactions.

includes the Jonckheere-Terpstra test in the output data set. The JT option in the TABLES statement requests the Jonckheere-Terpstra test. For more information, see the section Jonckheere-Terpstra Test.

includes the simple kappa coefficient in the output data set. The AGREE option in the TABLES statement requests computation of kappa, which is available for square tables (where the number of rows equals the number of columns). For multiway square tables, the KAPPA option also includes the overall kappa coefficient in the output data set. For more information, see the sections Simple Kappa Coefficient and Overall Kappa Coefficient.

includes Kendall’s tau-b in the output data set. The MEASURES option in the TABLES statement requests computation of Kendall’s tau-b. For more information, see the section Kendall’s Tau-b.

includes the asymmetric lambda in the output data set. The MEASURES option in the TABLES statement requests computation of lambda. For more information, see the section Lambda (Asymmetric).

includes the symmetric lambda in the output data set. The MEASURES option in the TABLES statement requests computation of lambda. For more information, see the section Lambda (Symmetric).

includes the asymmetric lambda in the output data set. The MEASURES option in the TABLES statement requests computation of lambda. For more information, see the section Lambda (Asymmetric).

includes the logit estimate of the common odds ratio in the output data set. The CMH option in the TABLES statement requests computation of this statistic, which is available for tables. For more information, see the section Adjusted Odds Ratio and Relative Risk Estimates.

includes the logit estimate of the common relative risk (column 1) in the output data set. The CMH option in the TABLES statement requests computation of this statistic, which is available for tables. For more information, see the section Adjusted Odds Ratio and Relative Risk Estimates.

includes the logit estimate of the common relative risk (column 2) in the output data set. The CMH option in the TABLES statement requests computation of this statistic, which is available for tables. For more information, see the section Adjusted Odds Ratio and Relative Risk Estimates.

includes the likelihood ratio chi-square in the output data set. The CHISQ option in the TABLES statement requests computation of the likelihood ratio chi-square. For more information, see the section Likelihood Ratio Chi-Square Test.

includes McNemar’s test (for tables) in the output data set. The AGREE option in the TABLES statement requests computation of McNemar’s test. For more information, see the section McNemar’s Test.

includes the following measures of association in the output data set: gamma, Kendall’s tau-b, Stuart’s tau-c, Somers’ , Somers’ , Pearson and Spearman correlation coefficients, lambda (symmetric and asymmetric), and uncertainty coefficients (symmetric and asymmetric). For tables, the MEASURES option also includes the odds ratio, column 1 relative risk, and column 2 relative risk. The MEASURES option in the TABLES statement requests computation of these statistics. For more information, see the section Measures of Association.

includes the Mantel-Haenszel chi-square in the output data set. The CHISQ option in the TABLES statement requests computation of the Mantel-Haenszel chi-square. For more information, see the section Mantel-Haenszel Chi-Square Test.

includes the Mantel-Haenszel estimate of the common odds ratio in the output data set. The CMH option in the TABLES statement requests computation of this statistic, which is available for tables. For more information, see the section Adjusted Odds Ratio and Relative Risk Estimates.

includes the Mantel-Haenszel estimate of the common relative risk (column 1) in the output data set. The CMH option in the TABLES statement requests computation of this statistic, which is available for tables. For more information, see the section Adjusted Odds Ratio and Relative Risk Estimates.

includes the Mantel-Haenszel estimate of the common relative risk (column 2) in the output data set. The CMH option in the TABLES statement requests computation of this statistic, which is available for tables. For more information, see the section Adjusted Odds Ratio and Relative Risk Estimates.

includes the number of nonmissing observations in the output data set.

includes the number of missing observations in the output data set. For more information, see the section Missing Values.

includes the odds ratio (for tables) in the output data set. The MEASURES, OR, and RELRISK options in the TABLES statement request this statistic. For more information, see the section Odds Ratio.

includes the Pearson chi-square in the output data set for two-way tables. For more information, see the section Pearson Chi-Square Test for Two-Way Tables. For one-way tables, the PCHI option includes the chi-square goodness-of-fit test in the output data set. For more information, see the section Chi-Square Test for One-Way Tables. The CHISQ option in the TABLES statement requests computation of these statistics.

If you specify the CHISQ(WARN=OUTPUT) option in the TABLES statement, the PCHI option also includes the variable WARN_PCHI in the output data set. This variable indicates the validity warning for the asymptotic Pearson chi-square test.

includes the Pearson correlation coefficient in the output data set. The MEASURES option in the TABLES statement requests computation of the Pearson correlation. For more information, see the section Pearson Correlation Coefficient.

includes the phi coefficient in the output data set. The CHISQ option in the TABLES statement requests computation of the phi coefficient. For more information, see the section Phi Coefficient.

includes the polychoric correlation coefficient in the output data set. For tables, this statistic is known as the tetrachoric correlation coefficient. The PLCORR option in the TABLES statement requests computation of the polychoric correlation. For more information, see the section Polychoric Correlation.

includes the column 1 risk difference (row 1 – row 2) in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences, which are available for tables. For more information, see the section Risks and Risk Differences.

includes the column 2 risk difference (row 1 – row 2) in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences, which are available for tables. For more information, see the section Risks and Risk Differences.

includes the column 1 and column 2 relative risks (for tables) in the output data set. The MEASURES and RELRISK options in the TABLES statement request these statistics. For more information, see the section Relative Risks.

includes risks (binomial proportions) and risk differences for tables in the output data set. These statistics include the row 1 risk, row 2 risk, total (overall) risk, and risk difference (row 1 – row 2) for column 1 and column 2. The RISKDIFF option in the TABLES statement requests computation of these statistics. For more information, see the section Risks and Risk Differences.

includes column 1 risks (binomial proportions) and risk differences for tables in the output data set. These statistics include the row 1 risk, row 2 risk, total (overall) risk, and risk difference (row 1 – row 2). The RISKDIFF option in the TABLES statement requests computation of these statistics. For more information, see the section Risks and Risk Differences.

includes column 2 risks (binomial proportions) and risk differences for tables in the output data set. These statistics include the row 1 risk, row 2 risk, total (overall) risk, and risk difference (row 1 – row 2). The RISKDIFF option in the TABLES statement requests computation of these statistics. For more information, see the section Risks and Risk Differences.

includes the column 1 relative risk in the output data set. The MEASURES and RELRISK options in the TABLES statement request relative risks, which are available for tables. For more information, see the section Odds Ratio and Relative Risks for 2 x 2 Tables.

includes the column 2 relative risk in the output data set. The MEASURES and RELRISK options in the TABLES statement request relative risks, which are available for tables. For more information, see the section Odds Ratio and Relative Risks for 2 x 2 Tables.

includes the overall column 1 risk in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences, which are available for tables. For more information, see the section Risks and Risk Differences.

includes the column 1 risk for row 1 in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences, which are available for tables. For more information, see the section Risks and Risk Differences.

includes the column 2 risk for row 1 in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences, which are available for tables. For more information, see the section Risks and Risk Differences.

includes the overall column 2 risk in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences. For more information, see the section Risks and Risk Differences.

includes the column 1 risk for row 2 in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences, which are available for tables. For more information, see the section Risks and Risk Differences.

includes the column 2 risk for row 2 in the output data set. The RISKDIFF option in the TABLES statement requests computation of risks and risk differences, which are available for tables. For more information, see the section Risks and Risk Differences.

includes the Spearman correlation coefficient in the output data set. The MEASURES option in the TABLES statement requests computation of the Spearman correlation. For more information, see the section Spearman Rank Correlation Coefficient.

includes Somers’ in the output data set. The MEASURES option in the TABLES statement requests computation of Somers’ D. For more information, see the section Somers’ D.

includes Somers’ in the output data set. The MEASURES option in the TABLES statement requests computation of Somers’ D. For more information, see the section Somers’ D.

includes Stuart’s tau-c in the output data set. The MEASURES option in the TABLES statement requests computation of tau-c. For more information, see the section Stuart’s Tau-c.

includes the Cochran-Armitage test for trend in the output data set. The TREND option in the TABLES statement requests computation of the trend test. This test is available for tables of dimension or . For more information, see the section Cochran-Armitage Test for Trend.

includes Bowker’s symmetry test in the output data set. The AGREE option in the TABLES statement requests computation of Bowker’s test. For more information, see the section Bowker’s Symmetry Test.

includes the uncertainty coefficient (symmetric) in the output data set. The MEASURES option in the TABLES statement requests computation of the uncertainty coefficient. For more information, see the section Uncertainty Coefficient (Symmetric).

includes the asymmetric uncertainty coefficient in the output data set. The MEASURES option in the TABLES statement requests computation of the uncertainty coefficient. For more information, see the section Uncertainty Coefficients (Asymmetric).

includes the asymmetric uncertainty coefficient in the output data set. The MEASURES option in the TABLES statement requests computation of the uncertainty coefficient. For more information, see the section Uncertainty Coefficients (Asymmetric).

includes the weighted kappa coefficient in the output data set. The AGREE option in the TABLES statement requests computation of weighted kappa, which is available for square tables larger than . For multiway tables, the WTKAP option also includes the overall weighted kappa coefficient in the output data set. For more information, see the sections Weighted Kappa Coefficient and Overall Kappa Coefficient.