Response Errors in Estimating the Value of Homes

Abstract

In the 1950 Survey of Consumer Finances home owners were asked to estimate the market value of their houses. Estimates for these same homes were later made by professional appraisers. These two estimates for each of 568 homes comprise the data analyzed here. The proportion of discrepancies between the two estimates is great: only 37 per cent of the estimates by respondents are within plus or minus 10 per cent of the appraisers' estimates. However, the errors tend to be offsetting, and in none of the ten price classes used is the difference in the relative frequencies for owners and appraisers statistically significant. Similarly, although the root-mean-square difference between the two measurements is high (an average of $3,100), the mean of the respondents' estimates is only $350 higher than the mean of $9,200 for the appraisers' estimates. The amount of variability is found to be rather similar for several sub-populations. However, for houses worth over $10,000 the mean-square difference between the measurements is found to increase with the value of the home. In the Appendix a model is developed for the statistical investigation of the data.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 100 citations exist upstream.