NotesFAQContact Us
Search Tips
Peer reviewed Peer reviewed
PDF on ERIC Download full text
ERIC Number: EJ863505
Record Type: Journal
Publication Date: 2008
Pages: 9
Abstractor: As Provided
Reference Count: 14
ISSN: ISSN-8756-5943
A Tutorial on Calculating and Interpreting Regression Coefficients in Health Behavior Research
Stellefson, Michael L.; Hanik, Bruce W.; Chaney, Beth H.; Chaney, J. Don
Health Educator, v40 n1 p12-20 Spr 2008
Regression analyses are frequently employed by health educators who conduct empirical research examining a variety of health behaviors. Within regression, there are a variety of coefficients produced, which are not always easily understood and/or articulated by health education researchers. It is important to not only understand what these coefficients are, but also understand what each represents and how each is interpreted. By correctly interpreting regression coefficients (Pearson r, Pearson r[superscript 2], Multiple R[superscript 2], a, b, [beta], r[subscript s], r[superscript 2][subscript s], partial correlation and semi-partial correlation), the informed health behavior researcher can better understand the dynamics of his/her data. The purpose of this manuscript is to describe and explain some of the coefficients produced in regression analysis. Specifically, the manuscript will describe (a) why and when each regression coefficient is important, (b) how each coefficient can be calculated and explained, and (c) the uniqueness between and among specific coefficients. A data set originally used by Holzinger and Swineford (1939) will be referenced throughout the manuscript to tangibly illustrate how coefficients should be calculated and interpreted in both simple and multiple regression analyses. (Contains 5 figures.)
Eta Sigma Gamma Inc. 2000 University Avenue CL 325, Muncie, IN 47306. Tel: 800-715-2559; Tel: 765-285-2258; Fax: 765-285-3210; Web site:
Publication Type: Journal Articles; Reports - Descriptive
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A