Food Insecurity Experience Scale

A raw count treats all items as equally severe and all surveys as equivalent, which they are not. The FIES uses a Rasch model so that each item carries an estimated severity and each respondent receives a latent-trait score, then equates that score onto a global reference scale. This means a given level of food insecurity is interpreted the same way across countries and over time, and prevalence is computed by integrating measurement uncertainty above fixed global thresholds rather than by a simple headcount of affirmations.

It measures the severity of food insecurity as experienced — the spectrum running from anxiety about access, through compromising on the quality and quantity of food, to skipping meals and going a full day without eating. It does not measure dietary quality, micronutrient adequacy, caloric intake, or nutritional status. For those dimensions you would pair the FIES with dietary diversity scores, dietary recalls, or anthropometric data; the FIES is best understood as an access-and-experience indicator.

The Rasch model assumes items differ in severity but share a common discrimination, which makes the raw score a sufficient statistic for the latent trait and greatly simplifies cross-country equating. This parsimony is a feature for a global monitoring instrument: it yields stable, interpretable item severities that can be aligned across surveys with a simple linear transformation. The trade-off is that items must be screened for fit, and items that discriminate atypically in a given context are flagged or dropped from that calibration.