Data Dictionary of Variables

County Business Patterns 1990 CD-ROM [# of establishments]

bowl90: Bowling centers
civic90: civic and social associations
fitns90: Physical fitness facilities
golf90: Public golf courses
relig90: Religious organizations
sport90: sports clubs, managers and promoters
memspt90: Membership sports and recreation clubs
pol90: Political organizations
prof90: Professional organizations
bus90: Business associations
labor90: Labor organizations
memnec90: membership organizations not elsewhere classified
assn90: the aggregate for all of above variables

Census Mail Response Rates 1990

Respn90: Census Mail Response Rates 1990
(Suggestion: Missing values in each state may be replaced with the response rate for the state: See for state response rates)

Voter Turnout (USA Counties CD-ROM)

Pvote88: Vote cast for president in 1988/total population age 18 and over in 1990.
Pvote92: Vote cast for president in 1992/total population age 18 and over in 1990.

Tax-exempt not-for-profit organizations

National Center for Charitable Statistics:

Nccs90: Number of not-for-profit organizations
(Note: Organizations under CBP and NCCS may not be mutually exclusive)


Pop90: Resident population

Social Capital Index

ski90pcm: This composite index was created using principal components analysis. The variables used in the principal components analysis are: total associations (assn90) per 10,000 people, number of not-for-profit organizations (nccn90) per 10,000 people, census mail response rate for 1990, and vote cast for president in 1988 divided by total population of age 18 and over in 1990. The first principal component explained about 46% of the variation and is considered here as the index of social capital. The first principal component also had an eigenvalue of 1.824 and the other components had eigenvalues of less than one.

ski90std: This composite index was created using the same variables used to create ski90pcm. Each of these variables was standardized to have a mean of zero and a standard deviation of one and then the mean of the standardized variables was used to create the composite index of ski90std.

The simple correlation between the two composite indexes is 99 percent.