Abstract
The authors used community-based participatory research to survey 238 residents in public housing. They developed 9 exposure and 2 health symptom indexes through exploratory factor analyses. The univariate analysis showed that most environmental risk factors were associated with the symptom indexes. The authors found a statistically significant intercorrelation between 29 of 36 risk factor pairs. Because of this, most of the univariate associations lost statistical significance in multivariate models. The authors found a normal distribution of risk factors across household; however, risk factors per apartment were associated with symptoms. Environmental risk factors accounted for 2% to 26% of the variation in symptoms. The analysis suggests that correlation between factors could affect surveys in which researchers assess a small number of housing risk factors or assess larger numbers of factors without testing intercorrelation. The analysis is consistent with the possibility that building quality or housekeeping might drive the development of risk factors.