(1) Goldman Sachs Global Economics, Commodities and Strategy Research in its US Daily: State of the States (Pandl) report for August 16, 2011, based on a regression model. This report was cited in the Washington Post online and elsewhere: Ezra Klein, “Why Are Some States Doing Better than Others?” Washington Post, August 24, 2011. Separately, David E. Rapach and Jack K. Strauss run models that use (1) state real income growth, (2) state real housing price growth, (3) state housing building permit growth, and (4) average change in unemployment in adjacent states to forecast unemployment change. See Rapach and Strauss, “Forecasting U.S. State-Level Employment Growth: An Amalgamation Approach,” International Journal of Forecasting, forthcoming (preprint version at http://forecasters.org/submissions10/StraussEmploymentstates.pdf.)

(2) The eight measures were: housing price inflation (1991-2006); age dependency ratio (people age 18-65/who population); nonwhite population as percentage of whole population; percent of workforce in professional jobs; per capita GDP; oil and gas extraction per GDP (2003); percent living at current address, 5 yrs (all residents); percent of adults age 25+ holding a high school degree or greater.

(3) We refer to a regression model with state unemployment change (2006-10) as the dependent variable and the eight independent variables listed in note 2 plus volunteering (2006), attending public meetings (2006), working with neighbors (2006), registering to vote in 2006, and voting in the 2006 election. All the civic measures are from the Current Population Supplement. R2=..639.

(4) In this model, r2=.551. For working with neighbors, the unstandardized coefficient is -.256 and p<.005.

(5) In this model, r2=.480. For the meeting attendance variable, the unstandardized coefficient is -.239 and p<.05.

(6) In this model, r2=.546. For the volunteering variable, the unstandardized coefficient is -.192 and p<.001.

(7) In this model, r2=.505. For registering to vote, the unstandardized coefficient is -.105 and p<.005. Using this method, voting misses being statistically significant (p=.228).

(8) The correlations between state volunteering and meeting attendance rates in 2006, 2008, and 2010 exceeded 0.9. For working with neighbors, the correlations were somewhat lower at 0.78-.81. In the models using 2006 and 2008 data, the confidence intervals overlap for the three community engagement variables that were measured in both years.

(9) The method is to regress these factors, as measured in 2008, individually against the unemployment change from 2006-2010, with models that control for the same eight 2006 economic factors listed in note 2. We report the results that are significant at p < .05. Factors that missed being statistically significant predictors of unemployment change were: voting in the 2008 election; talking regularly with neighbors; regularly discussing politics; regularly listening to radio news; obtaining news regularly from a newspaper; boycotting or “buycotting” products; belonging to church, school or “other” groups; regularly eating dinner with family members; and communicating with friends and family online. Obtaining news regularly from magazines, the Internet, and television were positively associated with unemployment change: higher rates of consuming these media were associated with greater growth in unemployment to statistically significant degrees.

(10) For a broad overview of benefits, see John Wilson, “Volunteering,” Annual Review of Sociology, vol. 26, (2000), pp. 231‐3

(11) Abt Associates, JoAnn Jastrzab et al., Serving Country and Community: A Longitudinal Study of Service in AmeriCorps (Washington, DC: Corporation for National and Community Service, December 2004, Updated April 2007), pp. 34, 64.

(12) Dávila, A., and M. T. Mora. 2007. “Civic Engagement and High School Academic Progress: An Analysis Using NELS Data.” CIRCLE working paper 52, Jonathan M. Tisch College of Citizenship and Public Service, Center for Information and Research on Civic Learning and Engagement, Tufts University, Medford, MA. http://civicyouth.org/PopUps/ WorkingPapers/WP52Mora.pdf.

(13) A recent study with helpful literature review is Kim Mannemar Sønderskov, “Does Generalized Social Trust Lead to Associational Membership? Unravelling a Bowl of Well-Tossed Spaghetti,” European Sociological Review (2011), vol. 27, no. 4, pp.419-434.

(14) Francis Fukuyama, Trust: The Social Virtues and The Creation of Prosperity (New York: Free Press, 1996).

(15) Robert D. Putnam, “Community-Based Social Capital and Educational Performance,” in Diane Ravitch and Joseph P. Viteritti, eds., Making Good Citizens: Education and Civil Society (New Haven: Yale University Press, 2001), pp. 58-95.

(16) Jeffrey Berry, Kent Portney, and Ken Thomson, The Rebirth of Urban Democracy (Washington, DC, Brookings Institution Press, 1993).

(17) Robert D. Putnam, Making Democracy Work: Civic Traditions in Modern Italy (Princeton: Princeton University Press, 1994). Although the topic is Italy rather than the United States, this book is important for its strong methodology and contribution to basic theory.

(18) Soul of the Community survey (2010), Knight Foundation and Gallup, via http://www.soulofthecommunity.org

(19) We acknowledge Robert Putnam for this suggestion.


*These groups are not ranked because often the differences among the states and metropolitan areas that are listed are not statistically significant. By the same token, often the differences between the listed states and the runners-up would not be significant.