ANALYSIS OF NOMINAL DATA

• NCJ Number: 50952
• Author(s): H T REYNOLD
• Editor(s): E M USLANER
• Date Published: 1977
• Length: 82 pages
• Corporate Author: Sage Publications, Inc
• Sale Source: Sage Publications, Inc
• Publication Type: Statistics
• Language: English
• Country: United States of America
• Note: ONE OF SAGE UNIVERSITY PAPER SERIES ON QUANTITATIVE APPLICATIONS IN THE SOCIAL SCIENCES

Annotation

TECHNIQUES FOR SIMPLE AND MULTIVARIATE ANALYSIS OF NOMINAL DATA--FACTS THAT CAN BE SORTED INTO CATEGORIES (E.G., VOTING BEHAVIOR, MARITAL STATUS, RACE, SEX, RELIGION)--ARE EXPLAINED AND ILLUSTRATED.

Abstract

NOMINAL DATA DIFFER FROM OTHER KINDS OF INFORMATION IN THAT THEY ARE DISCRETE: A PERSON EITHER IS OR IS NOT POLITICALLY CONSERVATIVE. NOMINAL DATA DO NOT PERMIT THE RESEARCHER TO HANDLE SUBTLE COMPARATIVE GRADATIONS (E.G., MORE OR LESS CONSERVATIVE) OR TO MEASURE COMPARISONS (HOW MUCH MORE OR LESS CONSERVATIVE). ORDINAL AND INTERVAL SCALES ARE REQUIRED FOR SUCH OPERATIONS. COMPILATION OF NOMINAL DATA IS A MATTER OF COLLECTING DISCRETE FACTS AND CREATING MODELS WITH ONE PREDICTOR AND A DEPENDENT VARIABLE, OR MORE COMPLEX MODELS WITH MULTIVARIATE PREDICTORS. ANALYSIS OF NOMINAL DATA REQUIRES THAT THE RESEARCHER FIRST DETERMINE WHICH TYPE OF MEASURE OF ASSOCIATION IS MOST DESIRABLE (E.G., LINEAR OR CURVILINEAR). TO HELP RESEARCHERS, TEACHERS, AND ADVANCED STUDENTS IN THE SOCIAL SCIENCES MAKE THIS DECISION, ALTERNATIVE MEASURES OF ASSOCIATION ARE USED TO TEST DIFFERENT HYPOTHESES ABOUT THE EFFECT OF IDEOLOGY ON THE PARTY IDENTIFICATION OF VOTERS, AND AN EXAMPLE OF SIMPLE ANALYSIS OF NOMINAL DATA IS GIVEN. A SEPARATE SECTION EXAMINES TECHNIQUES OF MULTIVARIATE ANALYSIS, TOUCHING ON SUCH TOPICS AS CHI-SQUARE VALUES, PROPORTIONATE REDUCTION-IN-ERROR MEASURES, LOG-LINEAR MODELS, AND PARTIAL MEASURES OF ASSOCIATION. THE DISCUSSION PROVIDES THE MATERIALS NEEDED TO MASTER THE BASICS OF THE TECHNIQUES DESCRIBED. READERS ARE ADVISED TO PURSUE SPECIFIC TECHNIQUES IN GREATER DETAIL IN THE APPROPRIATE LITERATURE. POTENTIAL USERS OF NOMINAL DATA TECHNIQUES INCLUDE POLITICAL SCIENTIST INTERESTED IN THE IDEOLOGY-PARTY IDENTIFICATION AND OTHER VOTER-RELATED QUESTIONS; SOCIOLOGISTS AND PSYCHOLOGISTS CONCERNED WITH THE EFFECTS OF SEX, RACE, AND/OR INCOME ON SELF-ESTEEM; PUBLIC POLICY ANALYSTS WHO WISH TO DETERMINE THE LIKELIHOOD THAT A GIVEN POLICY WILL BE ADOPTED BY A LEGISLATURE OR REGION; AND COMMUNICATION RESEARCHERS INTERESTED IN THE EFFECTS OF SEX, RACE, AND/OR ORGANIZATION ON GROUP ATTITUDES TOWARD THE MEDIA. TABLES, MATHEMATICAL FORMULATIONS, AND A LIST OF REFERENCES ARE INCLUDED. (AUTHOR ABSTRACT MODIFIED--LKM)