No, the treatments were all paid for by the government of North Carolina in this case, and were based on levels of lead concentration in the blood, not socioeco

No, the treatments were all paid for by the government of North Carolina in this case, and were based on levels of lead concentration in the blood, not socioeconomic status. So by comparing those with slightly above the threshold (who got treatment and thus ended up with significantly lower lead levels in their blood) and those with slightly below the threshold (who did not receive treatment and therefore ended up with higher levels of lead in their blood) you can make an estimate of the effect of lead in the blood that should cut across SES. (This is described on page 316 of the pdf- actually page 2- if you tested above the 10 microgram/deciliter concentration twice you got treatment, if you tested above only once you did not get treatment. They did a matched control group where the initial tests matched a test in the treatment group and generally matched the SES and geography, but the second test was in the 5-10 mcg/dcl range. Thus there is strong evidence that the control kid had a blood-lead problem but just missed the cut-off for government treatment.)