Audit of the survey of residential buildings, site I, wave 2
Assesses the completeness and reliability of 1975 survey data on the observable characteristics of properties and buildings in Brown County, Wisconsin. Observations were completed for 2,075 buildings, 93.3 percent of the sample. Validation checks during fieldwork exposed so many data discrepancies that it was decided to repeat the observations for 1,156 buildings. Refielding improved the quality of the data. However, on the basis of the final validation comparison, analysts are cautioned to infer real differences between two evaluations of the same building or property only when the discrepancy exceeds one rating interval (the ratings are on a four-point scale). The incidence of errors and of missing or unusable responses presents no further data-quality problems. Record weights are computed for use in estimating the county's property and building characteristics. Application of the weights overestimates the properties in certain isolated strata.