Reliability and validity of observer ratings of neighborhoods

Objectives: The objectives were to examine the validity and reliability of a five-item neighborhood measurement scale. Methods: Respondents were enrolled from two catchment areas: a poor inner city and a heterogeneous suburban area. Items combine for a total score of 5 (best) to 20 (worst). The authors compared scales across catchment areas and respondent ratings and assessed inter-viewer effects and retest reliability. Results: Suburban neighborhood scale scores were 3 points lower (higher socioeconomic status, SES) than the inner-city scores. There was a strong relationship between scores and participants' neighborhood ratings. The retest correlation was substantial (.81), but only two of five items achieved kappas above .75. In regression models, interviewer experience and residence and individual interviewers contributed to different ratings, although there was still a marked difference between catchment areas. Discussion: Observer ratings of neighborhoods show promise as a measure of neighborhood SES, despite problems with interviewer effects. Future work should improve objective criteria for ratings.