Seguridad alimentaria en Colombia y modelo Rasch

Abstract

This study subjected two scales used to determine food security in national nutritional status surveys in Colombia (ENSIN) to the Rasch model to assess their performance on the contradictory evidence of its usefulness. In practical terms, the EPSA measures food insecurity based on a single item, p1, which due to its low difficulty (8.46 logits, standard deviation of 1.83) is inappropriate to measure the surveyed households. Based on statistical calculation five of the twelve items presented over fitting and three of the twelve (including the first) presented measurement degradation. Like the EPSA, p1 defines the result to a large extent showing that the test is inadequate for the surveyed households. Based on statistics at least four of the fifteen items showed over-adjustment and four of the fifteen, including the first, showed measurement degradation. Measurements using tools that are not appropriate, but attractive, can be a new and serious problem that adds to other conditions associated with household food insecurity.