# Representativeness of Social Surveys among Older Individuals Living in Poverty: Who Were Left Behind?

## TL;DR

This study examines how well social surveys represent older people in poverty, finding that those living alone are more likely to respond.

## Contribution

The study identifies response biases in social surveys among impoverished older individuals using linked administrative and survey data.

## Key findings

- Only 48.8% of sampled older public assistance recipients responded to the JAGES survey.
- Older individuals living alone were more likely to respond to the survey (adjusted IR 1.48).
- Seven public assistance recipients were misclassified as non-recipients in the JAGES survey.

## Abstract

As considering only the data of survey respondents overlooks the opinion of voiceless people, acknowledging the direction of response biases of social survey data is crucial. We aimed to examine the representativeness of social surveys among impoverished older individuals. We conducted a cross-sectional study. Using linkage data from the Japan Gerontological Evaluation Study (JAGES) and residential registry data of public assistance (PA), we examined the validity of responses on receiving PA, as documented by the JAGES. Furthermore, we assessed the sociodemographic factors associated with the response to the JAGES using information on age (65-74/75-84/≥85 years), sex (male/female), household composition (living alone or not), nationality (Japanese/others), and level of long-term care needs (not applicable/support required/care needed) using complete data of PA recipients. A multiple modified Poisson regression analysis was performed to calculate the adjusted incidence ratio (IR) for responses to each explanatory variable. Among the sampled 162 older PA recipients, 79 (48.8%) responded to the JAGES. Seven recipients were misclassified as non-recipients in JAGES survey. There was no misclassification among non-recipients. PA recipients living alone were more likely to respond to the JAGES (IR 1.48, 95% confidence interval, 1.02-2.13). Using linkage data of the PA and JAGES databases from different departments within the municipality, we observed that social survey data represented a greater proportion of individuals living alone than those living with others among older impoverished people. To provide equitable policies, stakeholders should collect further information on older impoverished individuals not living alone.