Despite the significant correlations with which we began this report and the research cited above, a reader should not conclude, per se, that civic engagement alone boosts employment. The following cautions are important:

• There are other plausible hypotheses that we have not been able to test, because civic engagement was not included in the Current Population Supplement or other federal surveys until recently. For example, perhaps the housing bubble, which tripled housing prices in some states between 1991 and 2006, eroded civic health in those states by drawing in many new residents who had not had time to put down roots. When the bubble burst after 2005, those states were especially badly hurt by the recession, but their civic health had already declined. In that case, the relationship between civic engagement and employment that we found during 2006-10 would be misleading.

• Other unknown events may have lowered both civic health and employment between 2006 and 2010. An example would be a specific economic policy that was implemented in some states but not in others.

• Although the hypothesis is that civic health in 2006 affected unemployment change from 2006-10, the inverse is also entirely possible—that unemployment has affected civic engagement. In fact, by most measures, the civic health of the nation has declined since 2006, and a leading explanation of that decline is the economic recession and its aftermath. Thus, even if the relationship between civic engagement and unemployment is meaningful, the causal arrow could point either way, or could point both ways at once. Reciprocal relationships are very common in the social sciences and still important to examine in more detail.