Recommendation 5-2: USDA should undertake the following additional analyses in the development of the underlying latent variable model:
Fitting models that allow for different latent distributions for households with children and those without children and possibly other subgroups of respondents.
Fitting models that allow for different item parameters for households with and without children for the questions that are appropriate for all households in order to study the possibility and effects of differential item functioning.
Studying the stability of the measurement system over time, possibly using the methods of differential item functioning.
Recommendation 5-3: To implement the underlying latent variable model that results from the recommended research, USDA should develop a new classification system that reflects the measurement error inherent in latent variable models. This can be accomplished by classifying households probabilistically along the latent scale, as opposed to the current practice of deterministically using the observed number of affirmations. Furthermore, the new classification system should be more closely tied to the content and location of food insecurity items along the latent scale.
Recommendation 5-4: USDA should study the differences between the current classification system and the new system, possibly leading to a simple approximation to the new classification system for use in surveys and field studies.
Recommendation 5-5: USDA should consider collecting data on the duration of spells of food insecurity in addition to the currently measured intensity and frequency measures. Measures of frequency and duration spells may be used independently of the latent variable measuring food insecurity.