Having Trouble locating Social Cohesion+Non Profit Density Rankings

[Andrew Jawitz]

Hello All,

   I came across this challenge while doing research for a presentation our Code for America Brigade (Code for Maine) will be giving to a committee of government/private sector leaders on this coming Tuesday 6/28/2013.  Basically, the committee invited us to discuss the potential for civic hacking to help address troubling numbers concerning the tech industry in Maine.  However, when we discovered the report titled “Civic Health and Unemployment II: The Case Builds”  the numbers present Maine in a very different light...  According to the report, Maine is consistently listed amongst the top five states in critical civic health measures of non-profit density and social cohesion.  Armed with such data we'd be able to make a convincing case that Civic Hacking can be used to leverage civic cohesion in Maine as a tool for addressing economic problems in the tech sector.

   However, when we've tried to drill deeper into the data all we've found are the raw data files hosted on the Civic Data Challenge Website  which do not appear to include aggregate state by state rankings on social cohesion or non-profit density. After, reviewing the report again and again, we've found nothing that indicates the metrics used to determine "social cohesion".  Are we missing something?  

       To be clear, I'm not asking anybody to crunch the numbers for us, but rather to help locate what appears to be a missing link between the raw data and the metrics used in the "Civic Health and Unemployment" report.

    While our presentation is on Tuesday, the outcome may well involve participation in the Civic Data Challenge.  But first we need to locate the missing pieces...

                                                                                                                         Thank You Very Much,

                                                                                                                          Andrew Jawitz

                                                                                                                           Brigade Captain

                                                                                                                           Code for Maine        

[Alice Murphy]

Hi Andrew,

Thanks so much for the question and for your interest in the economic research.

NCoC collaborated with our partners at the Center for Information and Research on Civic Learning and Engagement (CIRCLE) at Tufts University who analyzed the Current Population Survey data (the raw data provided for the Civic Data Challenge) for this report (www.ncoc.net/unemployment2). 

The definition of social cohesion, and indicators used to inform it, are available in the Glossary of the report (http://www.ncoc.net/Gloassary-of-Terms-and-Data-Sources). Furthermore, our research lead offered this insight:

“Social cohesion was calculated as an index score based on a factor analysis and the items used are listed in the executive summary. Using three items (trust in neighbors, seeing or hearing from friends, and giving or receiving favors to/from neighbors), we tested whether these three items make up a cohesive construct (or a “concept”) that makes theoretical sense. We then calculated a factor-score, which puts people on a normal curve of social cohesion level based on their response to each of these three items.  Someone that is absolutely average on all three items would have a factor score of 0, and someone that is above average would have a positive factor score and someone below average would have a negative score.  Once we had the social cohesion factor score for everyone, we were then able to calculate state averages by averaging everyone in that state’s social cohesion factor score.  The state averages were then ranked.”

As for the NPO density data, the data used in the report was from the National Center for Charitable Statistics categories of nonprofits. The source is included in endnote 7 of the report—the National Taxonomy of Exempt Entities. Our research lead said that they then ran regression using each category (e.g., social benefits) as a predictor in multiple regression models which controlled for various factors (specified in endnote 11).  

Let us know if you have further questions!