Logistic regression was used to estimate the odds ratios of asthma for the set of predictors shown in the graph. The same 6,201 children were analyzed in both regressions, so the only difference was in which definition of an “asthma case” was used: the mother’s report or a doctor’s diagnosis.

Odds ratios greater than 1.0 indicate higher risk of asthma, while those below 1.0 indicate reduced risk.

Variables marked * indicate that they were significant in model of maternal report of asthma but not doctor’s report. Those marked # were significant in both maternal report and doctor’s diagnosis of asthma.