Chapter 11 - PROC FREQ - Analyzing Tables and Counts, Crosstabs, Chi Square, Kappa, Odds Ratio скачать в хорошем качестве

Chapter 11 - PROC FREQ - Analyzing Tables and Counts, Crosstabs, Chi Square, Kappa, Odds Ratio

Chapter 11 - PROC FREQ - Analyzing Tables and Counts, SAS Essentials, Alan Elliott & Wayne Woodward Summary (auto generated) SAS Essentials: Analyzing Counts and Tables (Chapter 11) Introduction to PROC FREQ Multi-purpose SAS procedure for analyzing count data Used for frequency counts of individual variables and two-way tables Basic syntax: PROC FREQ [options]; followed by TABLES [requested tables] / [options]; One-Way Frequency Tables Used to produce tables of counts by categories Common options: ORDER=FREQ arranges categories by descending frequency ORDER=DATA arranges by order in the dataset ORDER=FORMATTED arranges by formatted values Using Summarized Data Weight variables can be used when data is already summarized Example: WEIGHT number; where number represents the count of each category Goodness of Fit Testing Tests if observed frequencies match expected theoretical distributions Syntax: TABLES variable / CHISQUARE; TEST P=(value1 value2...); Example: Mendel's pea plant experiment testing 9:3:3:1 genetic ratio Two-Way Tables (Cross-Tabulations) Format: TABLES A*B / CHISQUARE; or TABLES A BY B / CHISQUARE; Tests two types of hypotheses: Test of independence: no association between variables Test of homogeneity: same distribution across populations Analyzing Cross-Tabulation Results Tables display count, total percent, row percent, and column percent Chi-square test determines if there's a significant relationship Example: Analysis of relationship between drinking status and crime types Calculating Relative Risk Used in two-by-two tables to examine risk measures Syntax: TABLES variable1*variable2 / RELRISK; Interpretation: Odds ratio/relative risk GT 1: exposure is harmful Odds ratio/relative risk LT 1: exposure is beneficial Example: Comparing rash incidence between two cleaners Inter-Rater Reliability (Cohen's Kappa) Measures agreement between two raters for categorical measurements Syntax: TABLES rater1*rater2 / AGREE; TEST KAPPA; Kappa interpretation scale: 0.61-0.80: Substantial agreement 0.41-0.60: Moderate agreement 0.21-0.40: Fair agreement Below 0.20: Slight or poor agreement Weighted Kappa Used when categories have a meaningful order Syntax: TABLES rater1*rater2 / AGREE; TEST WTKAP; Accounts for degree of disagreement in ordinal categories Example: Rating airline passengers for security threat levels