The How the Data are Collected (methodology) section describes how the U.S. Census Bureau collects the Quarterly Summary of State and Local Government Tax Revenue data and how the published data were generated. Select the appropriate topic below for information on the sampling frame, collection periods, collection methods, editing, imputation, sampling error, response rates, and nonsampling error.
The population of interest for this survey covers the 50 state governments, all local government property tax collectors and local government non-property tax imposers. All state government tax collection is covered. A stratified cluster sample of local property tax collectors yields the estimates of property tax. A stratified sample of local tax imposers yields estimates of all nonproperty taxes. For the fourth quarter of 2010, the latter sample, which gives estimates of nonproperty taxes, replaced a nonprobability sample. A five-quarter bridge report, which compares results using the old methodology to results using the new sample and new methodology, accompanies the release of the data for the first quarter of 2012.
The state government tax data presented by the Census Bureau may differ from data published by state governments because the Census Bureau may be using a different definition of which organizations are covered under the term, "state government." For the purpose of the survey statistics, the term "state government" refers not only to the executive, legislative, and judicial branches of a given state, but it also includes agencies, institutions, commissions, and public authorities that operate separately or somewhat autonomously from the central state government but where the state government maintains administrative or fiscal control over their activities as defined by the Census Bureau.
For further information on the definition and organization of governments, see Chapter 1 of the Government Finance and Employment Classification Manual and the 2007 Census of Governments.
Confidentiality: : Data collected for the Quarterly Summary of State and Local Government Revenue are public record and are not confidential, as authorized by Title 13, United States Code, Section 9.
Dates of Collection: The following are important dates in the data collection process:
|First week following the reference period||Initial mail-out|
|Six weeks after initial mail-out||Non-response follow-up|
|10 weeks after initial mail-out||Completion of data compilation and editing|
Methods: Local government tax data (F-71 for local government property taxes and F-73 for local government non-property taxes) are collected by mail. Data are returned by mail or through web collection. The fifty state governments provide the Census Bureau with administrative records from their central accounting systems.
Editing: Editing is a process that identifies possibly inaccurate data for correction or verification. Efforts are made at all phases of collection, processing, and tabulation to detect and minimize errors.
Data are processed from several collection methods including direct response to survey forms from state government officials, as well as from the compilation of administrative records and supplemental sources. Regardless of the collection method, these data are edited using ratio edits of the current quarter's value to the value in the same quarter of the prior year.
Data are also subject to classification of administrative records. The fifty state governments provide the Census Bureau with administrative records from their central accounting system. These administrative records are unique to each state as each state is legally organized differently from every other state and, as such, each state has a unique organizational and accounting structure. It is the responsibility of the Census Bureau to classify the different accounting and organizational structures into uniform tax categories so that entities with different methods of government accounting can be presented on a comparable basis. The records represent the core, or central, state government and are limited to tax revenue. Statistics on state government tax revenues are compiled from state administrative records by Census Bureau employees, according to the Census Bureau's classification methodology as outlined in the Government Finance and Employment Classification Manual.
Imputation: Not all respondents answer every item on the questionnaire. There are also questionnaires that are not returned despite efforts to gain a response. Imputation is the process of filling in missing or invalid data with reasonable values in order to have a complete data set for analytical purposes.
Local government property tax non-response imputation utilizes two methods that minimize bias and give more accurate imputations. When historical data are available, imputes are calculated using a median growth rate multiplied by the data from the same quarter in the prior year. In cases where no historical data are available, the missing data are imputed using the adjusted cell mean of the property tax amount.
Local nonproperty taxes were not imputed for the first quarter of 2012. Due to low unit response rates, there were insufficient reported data to adequately impute for the nonrespondents. In order to obtain estimates for 2012, calibration methods were used. The 2007 Census of Governments: Finance component along with the intercensal annuals for both Employment and Finance data were used in the modeling and calibration.
For state government non-response, data are imputed using a national growth rate, after removing outliers, applied to the same quarter data from the prior year. There may be times when state records do not include full tax revenue detail or reporting units do not respond and supplemental data sources are required. Economic statistics for these supplemental sources are obtained and pro-rated from external financial reports or the Census Bureau's Annual Survey of State Government Finances and Annual Survey of State Tax Collections when all other data source investigations are exhausted. Supplemental records are merged with data from the state governments. Although every effort is made to obtain financial information from all state government entities, financial statements may not be available at the time the Census Bureau closes the processing, or governmental entities may not respond to our requests. As a result these data are subject to revisions for up to 7 prior quarters, under conditions of improved information.
Imputation Changes (Effective December 2020, Quarter 4, 2020)
The QTAX survey data are being migrated from a legacy data processing system to the Standard Economic Processing System (StEPS II)-- the system used for many other Census economic surveys. The migration of historical data, and the use of StEPS II for future QTAX survey quarters, will require the use of a different data coding framework. QTAX survey analysts use certain codes to identify important characteristics, such as how the data were reported, whether the data required corrections, or if the data were imputed. The current set of codes used to process and edit the survey data will be simplified and streamlined in the transition to StEPS II.
Imputed data for the QTAX surveys are calculated at the government unit and tax item level. While the QTAX surveys will continue to use the same imputation methods, changes to the data processing codes may affect the values generated by these procedures. The imputation methodologies for F71 and F72 are dependent on the use of these codes, and therefore, we have carefully assessed the closest approximation to our procedures in the original processing system. The values of these codes are used to identify the data included in the imputation base data used, for example, to calculate growth rates or medians that will be applied to nonresponding units. The adaptations made to move to STEPS II may result in slight differences in the results of imputation.
Revisions: Revisions reflect tax collection amounts obtained from three general sources. State and local government respondents have submitted revisions to amounts as originally reported. In other cases, governments have reported data, which we used to replace data that were previously imputed or estimated. Finally, some of the revisions were compiled from government sources, both published and unpublished. Current revisions are noted in each table by an "R" or, for Table 3, by a label next to the applicable figure, for appropriate quarters.
Sampling Error:The survey to obtain tax information from the state governments is a complete enumeration of all state government imposers and as such does not have any sampling error. On the other hand, all local property and nonproperty estimates are subject to sampling error.
The survey of local property tax is a probability sample designed to yield a national coefficient of variation of under 3.0 percent. Likewise, the new survey of local nonproperty tax imposers was selected to yield national coefficients of variation of under 3.0 percent for major taxes (sales and individual income taxes).
Variance Data: Data that are derived from the annual sample survey are subject to sampling error. The statistics in this report that are based wholly or partly on data from the sample are apt to differ from the results of a census covering all governments. Estimates based on a sample survey are subject to sampling variability. The particular sample used is one of a large number of all possible samples of the same size that could have been selected using the same sample design. Each of the possible samples would yield somewhat different results.
The standard error is a measure of the variation among the estimates from all possible samples and thus is a measure of the precision with which an estimate from a particular sample approximates the average results of all possible samples. Table 1 is followed by Tables 1b and 1c, which give users the coefficients of variation and margins of error, respectively, for the estimates in Table 1. Estimates of sampling error are not given for nonproperty taxes prior to the fourth quarter of 2010 because these estimates resulted from a nonprobability sample and as such, the sampling error cannot be measured. The coefficient of variation is the estimated standard error expressed as a percent of the estimated total or proportion.
State government financial statistics result from a complete canvass of all state government agencies. Consequently, there is no associated measure of sampling error, such as the coefficient of variation. However, these statistics are subject to non-sampling error. Such error includes inaccuracies in classification, coverage, and processing.
Although efforts were made at all phases of collection, processing, and tabulation to minimize errors, the data were still subject to errors from imputing for missing data, errors from miscoding, and errors in coverage. Every effort was made to keep such errors to a minimum through examining, editing, and tabulating the data.
The CVs (coefficient of variation) presented in tables can be used to derive the standard error of the estimate. The standard error can then be used to derive interval estimates with prescribed levels of confidence that the interval includes the average results of all samples:
The user can calculate the standard error by multiplying the CV presented in the tables by the corresponding estimate. The CVs presented in the tables are in percentage form and must be divided by 100 before being multiplied by the estimate. This standard error estimate can then be used to get a 90 percent interval estimate by multiplying it by 1.645 and adding the result to the estimated total to get the upper bound and subtracting it from the estimated total to get the lower bound.
The margin of error, which is the half-width of the confidence interval described above, is given in Table 1c. The Census Bureau standard is the 90 percent interval.
Nonsampling Error: All aspects of the survey collections are subject to nonsampling errors. These errors are due to mistakes that the respondent may have made, errors due to not receiving a returned questionnaire, or errors due to various processing issues, such as keying or classification. Such errors are mitigated by editing the data using prior quarter data or data from four quarters prior. Errors of nonresponse are mitigated by imputing for units that did not supply complete information or by weight calibration.
Response Rate: The Total Quantity Response Rates (TQRR) are given in Table 1a. TQRRs represent the percentage of the final estimate for a given variable that resulted from survey respondents. For the first quarter of 2012, five state governments did not report data and had to be imputed. These units are denoted in Table 3.