skip banner Southern Forest Resource Assessment    Draft Report

Search this site:


Home > Draft Report > SOCIO-7   

Previous PageNext Page

4 Data Sources

Four sources of data were used in the analyses. All units of observation were at the county level.

First, data on forest variables were obtained from the Forest Inventory and Analysis (FIA) unit of the Southern Research Station. Data were not available for all southern counties at the same point in time, so data from the most recent survey were used to provide the most current representation of forest conditions. For one variable, change in plantation acreage, the two most recent forest surveys were used to compute the percentage change. Because Kentucky was not included in forest surveys conducted by the FIA unit of the Southern Research Station, forest variables for this State were not directly comparable with those of other Southern States. Kentucky data, therefore, were not included in the analyses.

The reader should be alerted to the fact that FIA data were sampled in a way to meet sampling error standards at the State level. As data are subdivided into smaller geographical units, such as the county level used in this Chapter, the sampling errors increase and the reliability of the estimates decrease. This condition may impact the analysis reported in this Chapter primarily by increasing the size of the standard errors associated with the Pearson correlation coefficients where such correlations were estimated using forest variables. Increases in the standard errors associated with correlation coefficients suggest that some relationships that may exist did not meet the 10-percent significance threshold for reporting in this Chapter. However, we do not believe that this effect biased the estimated correlations or caused some correlations to appear statistically significant when, in fact, they are not.

Data on employment and income were obtained from the IMPLAN Database (Minnesota IMPLAN Group 1997). To make these data as comparable as possible with data from the most recent available Census data that were available when the analysis was undertaken (1990), we used 1993 IMPLAN data. Employment data in the IMPLAN Database were created from the Bureau of Labor Statistics ES202 data, the County Business Pattern data provided by the U.S. Department of Census, and the Regional Economic Information System data provided by the Bureau of Economic Analysis. It should be noted that these data are based on where people worked (where the industrial sectors were located), not on where they resided. However, across the entire South, a discrepancy between the county where people worked and where they resided should not be an important issue.

Data on a number of social and economic indicators were obtained from 1990 Bureau of Census data sets. These indicators included median household income, unemployment rate, poverty rate, and percent of owner-occupied housing. Of course, these data were based on where people resided.

Finally, data on a number of other social variables were obtained from the State and County Data Book, which was available on the Internet ( In an attempt to align these data with the Census data, we chose the most recent data that were closest in date to the 1990 Census. Thus, data from the 1994 State and County Data Book were obtained for the following indicators: crime rate (serious crimes per 100,000 population; 1991), percent graduating high school (persons 25 years and older, percent high school graduate or higher; 1990) infant mortality rate (deaths of infants under 1 year per 1,000 live births; 1988), percent voting in the most recent presidential election (votes cast for president; 1992, divided by voting-age population; 1992) and percentage population change (population, percent change 1980 to 1992).

Previous PageNext Page

Glossary | Sci.Names | Process | Comments | Final Report


content: Thomas P. Holmes
webmaster: John M. Pye

created: 21-NOV-2001