Do you see it how I see it?

For the study of neighbourhood effects, researchers often use register data-based measures of neighbourhood characteristics. However, the causal mechanisms explaining such neffects might be based on the perception of neighbourhood characteristics, rather than the objectively measured characteristics. The mismatch between the objective measures and subjective perceptions is likely influenced by individuals’ characteristics and attitudes. This paper investigates the mismatch in the measured percentage of low-income and foreign background neighbours and the perceptions of these neighbourhood characteristics. We use a unique Dutch neighbourhood perceptions survey merged with population register data to create neighbourhoods coinciding with the areas the respondents were asked to consider. The perceptions and register data-based measures correlate, although not strongly for the percentage of low-income neighbours. Older age and greater social embeddedness lead to underestimation of both characteristics, and higher household income to underestimation of the share of low-income neighbours. Higher education also leads to underestimation, but after controlling for individual institutional trust it becomes less significant to the perceived percentage of foreign background neighbours, and insignificant to the perceived percentage of low-income neighbours. Individuals with lower institutional trust are likely to overestimate both the percentages of their foreign background and low-income neighbours; in the case of low income, the same can be said for generalised trust. Our results could explain inconclusive results of neighbourhood studies using register-based variables, and suggest that urban research should benefit from augmenting administrative datasets with surveys and interviews highlighting feelings and values.