Study limitations were noted. NESARC/NESARC-III lacked biological testing for substances and excluded the literally homeless86 and some institutionalized individuals, as do most large general population surveys. Also, AUDADIS-5 interviewers were not clinicians. However, AUDADIS DSM-IV marijuana use diagnoses are related to considerable disability and comorbidity.22,87,88 Further, a NESARC-III substudy comparing AUDADIS and clinician diagnoses of 12-month marijuana disorder showed nearly identical prevalence and good concordance,50 suggesting valid diagnostic assessment of a clinically meaningful condition. Additionally, the NESARC-III response rate was acceptable (60.1%) but lower than for NESARC. Weighting that compensated for nonresponse facilitated comparisons between the surveys. However, surveys with lower response rates may miss more substance abusers,89 potentially leading to lower prevalence. If this happened in NESARC-III, then the NESARC/NESARC-III differences we found may actually underestimate the true differences. Additionally, employers of NESARC and NESARC-III interviewers differed (Census and West at, respectively); whether having different employers of the interviewers affected participant responses is unknown. However, NESARC and NESARC-III were both presented to respondents as voluntary surveys conducted under the auspices of the US government, possibly mitigating this difference. Finally,
