State and Local Government Finance
Survey Methodology
Fiscal Year 2006
 
__
 2006 Census Technical Documentation     Back to Survey Methodology 
__

Sample Design
    Description of design:
  The 2006 Sample for the Annual Finance Survey (AFS) was developed in from the 2002 Census of Governments. It was designed to produce state by type of government estimates with a relative standard error of 3 percent or less for long-term debt, total revenue, and total expenditure and state estimates with a relative standard error of 5 percent or less on the following 11 major finance items: long-term debt, total revenue, total expenditure, criminal justice, education, highways, health/hospitals, housing, capital outlay, utilities, and welfare. The sample included all independent schools since their data were obtained from other education surveys. All other government units were selected into the sample with a probability proportional to their size, either total expenditure or long-term debt. These goals and strategies were chosen to keep the sample size under 12,000 among non-school governmental units.

    Births of general purpose governments and independent schools since the completion of the 2005 AFS processing cycle and identified by GOVS were assigned a weight of 1.

  Births of special district governments that met initial certainty were also included with a weight of 1. All other special district government births were sampled at a rate of 1 in 25.

  Deaths (dis-incorporated units) since the completion of the 2004 AFS processing cycle were taken out of the sample.

  The initial certainty criteria for the 2006 sample were as follows:

 

         * All county governments with a 2002 population of 100,000 or more.
         * All cities with a 2002 population of 75,000 or more.
         * All townships in New England and Middle Atlantic states with a population of 50,000 or more.
         * All special districts meeting any of the following three conditions:
                  * Full-time equivalent of 1,000 or more;
                  * A function code of 42(Mortgage Credit), 92(Electric Power), 93(Gas Supply), or 94                     (Transit);
                  * Had either 20+ million total revenue or expenditure.
         * Special districts that are the only Special district for a state and specific function code.


    Sample Frame:
 

The sampling frame for the 2006 AFS was the 2002 Census of Local Governments file, updated by births, deaths, and mergers that took place since 2002. Altogether, there were 79,360 local governments of Type 1, 2, 3, or 4.
    Weighting:
 

The weight for each unit in the sample is the inverse of that unit’s probability of being selected into the sample. For example, for units that were included in the sample with a probability of .02, the weight is (1/.02)=50. For units that were included in the sample with a probability of 1, the weight is 1.
    Sample size:
 

The sample size was around 12,000 non-school units. Of the total number of governments in the non-school sample, approximately 15% were counties, 26% were cities, 19% were townships, and 40% were special districts. All 50 state governments and District of Columbia were certainty units with a weight of 1.

Note, that the entire sample included all independent schools since their data were obtained from other education surveys. Of the total number of governments in the entire sample, approximately 7% were counties, 11% were cities, 8% were townships, 18% were special districts, and 56% were school districts.


__
 
Source: U.S. Census Bureau, Governments Division,
Created: September 12 2007
Last revised: November 30 2009