Sonia
This is a really good point about the challenge of assessing reliability of the discrimination measure. I suppose this is a special case of a "retest" effect. But if this is going on, it implies that the first assessment is not good, because people cannot remember all of the discriminatory experiences they have encountered. In this case, you need to modify the measurement approach to give people more time to think back and remember their own life experiences.
With respect to absolute vs relative measures: I meant the characterization of the magnitude of the inequality, not the characterization of the SES predictor per se. In this case, it would be useful to have the absolute magnitude of inequality, and they could convert their odds ratios to absolute effects (e.g., children in the lowest SES category had an x% chance of out-of-control asthma, whereas children in the highest SES category had a y% chance of out of control asthma). This is important because the OR is very much influenced by the probability of the outcome overall and without knowing the absolute values and risk difference, we cannot well understand if this is an important public health burden for low SES children or not.
Maria