Skip to main content
ARS Home » Research » Publications at this Location » Publication #165785
Title: FINDING A FIT MODEL FOR FOODS

Author
item JO, CHAN-HEE - DELTA NIRI
item SIMPSON, PIPPA - DELTA NIRI
item GOSSETT, JEFF - DELTA NIRI
item Bogle, Margaret
item HUANG, EMMA - UNIV NORTH CAROLINA, BIOS

Submitted to: International Society for Behavioral Nutrition and Physical Activity
Publication Type: Abstract Only
Publication Acceptance Date: 4/23/2004
Publication Date: 6/10/2004
Citation: Jo, C.H., Simpson, P., Gossett, J., Bogle, M., Huang, B. 2004. Finding a fit model for foods [abstract]. Third Annual Conference of the International Society of Behavioral Nutrition and Physical Activity. p. 23.

Interpretive Summary:

Technical Abstract: Purpose: To show that there are many extensions of linear regression which allow a better understanding of the interrelationship of nutrition, health and socioeconomic status. Background: Often linear regression is used to explore the relationship of variables to an outcome. For example, in the Lower Mississippi Delta, it is of interest to see whether the total health eating index (HEI) is related to socioeconomic variables, age and health. R-square is usually defined as the proportion of variance of the response that is predictable from (that can be explained by) the independent (regressor) variables. When the R-square is low, it may be that there is little relationship of the independent variables to the dependent variable. Alternatively, the model fit may be poor. Methods: We investigate sources and remedies of poor fit. These include outliers, incorrect form of the independent variables and incorrect assumptions about the errors. We show that the errors may not be normally distributed and indeed the independent variables may not truly be fixed and have measurement error of their own. In fact the low R-square may be due to a combination of the above aspects. Conclusions: As a supplement to usual analyses, considering specific steps in fitting nutrition regression models enhances investigation of the complex interrelationship of factors affecting nutrition. Acknowledgements: This work was funded under the Lower Mississippi Delta Nutrition Intervention Research Initiative, USDA ARS grant #6251-53000-003-00D.