Skip to main content
. 2021 Dec 28;14(1):122. doi: 10.3390/nu14010122

Figure 2.

Figure 2

Linear regression progression (backward elimination). For the food group-based model, all food groups were kept in the original model. Nutrients and confounders were pre-selected based on the literature and on bivariate correlation analysis. Variables and confounders were eliminated stepwise in order to keep only significant variables or those where a trend was apparent (p < 0.1).