Researchers often use register data-based measures of neighborhood characteristics to estimate neighborhood effects. However, the underlying causal mechanisms might be based on the perception of such characteristics. The mismatch between the measures and perceptions is likely influenced by individuals’ characteristics and attitudes. This paper investigates the relationship between the measured percentages of low-income and foreign background neighbors and their perceptions. Using Dutch register data merged with survey data, we found that higher education is associated with the underestimation of both characteristics, but after controlling for individual institutional trust it becomes less significant to the perceived percentage of foreign background neighbors, and insignificant to that of low-income neighbors. Individuals with lower institutional trust are likely to overestimate both percentages. Older age, greater wealth and greater social embeddedness are associated with underestimation of both characteristics, and higher household income with the underestimation of the share of low-income neighbors. We find that trust in democratic and public institutions is associated with biased perceptions of social reality, which can be an important finding for the spatial inequality studies. Our results also suggest that urban research would benefit from augmenting administrative datasets with surveys and interviews.