Editing is a process that ensures survey data are accurate, complete, and consistent. Efforts are made at all phases of collection, processing, and tabulation to minimize errors. Although some edits are built into the Internet data collection instrument and the data entry programs, the majority of the edits are performed after the case has been loaded into the Census Bureau's database.
Edits consist primarily of two types: consistency and a ratio of the current year's reported value to the prior year's value. The consistency edits check the logical relationships of data items reported on the form. For example, if a value exists for the number of retirees receiving benefits because of age or length of service then there must be a value reported for the amount paid. The current year/prior year edits compare by item code the data reported for the current year with data reported for the prior year. If data fall out of acceptable tolerance levels, the item is flagged for review.
For both types of edits, the edit results are reviewed by analysts and adjusted when needed. When the analyst is unable to resolve or accept the edit failure, contact is made with the respondent to verify or correct the reported data.
Not all respondents answer every item on the questionnaire. Imputation is the process of filling in missing or invalid data with reasonable values in order to have a complete data set.
The imputations were based on either a prior year annual survey or the most recent Census of Governments. All but six missing variables (Z90, Z95 and four relatively new variables: Z13, Z14, Z15, and Z16) were imputed using one of the following methods: cell median or donor distribution of Z81, cell mean, or reported prior year or census year data which was multiplied by a growth factor. If the non-respondent to Z90 and Z95 does not write in anything in the Other categories for Z90 and Z95, we impute those variables to be zero.
After the data were edited, the survey data was aggregated to yield the viewable and downloadable files that are available on the website.
The 2014 Annual Survey of Public Pensions: State- and Locally-Administered Defined Benefit Data released data for Fiscal Year 2014 on July 28, 2015. Users should note that this release also includes revisions to Fiscal Years 2012 and 2013. The revised data are accessible through all viewable and downloadable data files on the survey's website.
Because all 230 state government pension systems are included, these data are not subject to sampling error or any sampling variability. Of the 3,742 local government-administered pension systems, 1,568 were sampled in 2014. Sampling error is available as a coefficient of variation (CV).