Response to Invited Commentary Wight et al. Respond to ‘‘Considerations about Multilevel Thinking’’

Galea and Ahern (1) have provided an insightful com-mentary on our article (2), raising several noteworthy ob-servations pertaining to the current state of and future directions for multilevel epidemiologic research. We concur that theory-driven multilevel studies are needed, along with explication of the mechanisms that link levels of influence on health, particularly since these mechanisms may be more readily amenable to intervention than structural deter-minants of health. For example, in our analyses of the as-sociation between neighborhood education and late-life cognition (2), we recommend promoting educational attain-ment in the general population in disadvantaged neighbor-hoods because it may prove cognitively beneficial to aging residents of these neighborhoods. An argument could be made instead for promoting alteration to the hypothesize