Mass Appraisal and Statistical Analysis
Mass Appraisal and Statistical Analysis
This domain tests how assessors value many parcels consistently using CAMA systems, stratification, ratio studies, and regression. Expect questions on measures of assessment level and uniformity, sample design, and common statistical bias.
Core rules
- Mass appraisal values a universe of properties as of one date using standardized data and models. It differs from single-property appraisal in scope and testing.
- A ratio study compares assessed values to sale prices. The median ratio measures assessment level; the coefficient of dispersion (COD) measures uniformity.
- COD = average absolute deviation from the median ratio ÷ median ratio × 100. Lower COD means better uniformity; acceptable ranges vary by property type and jurisdiction.
- Price-related differential (PRD) tests whether high-value and low-value properties are assessed regressively. PRD near 1.00 indicates vertical equity.
- Stratification by neighborhood, property type, or value can improve model fit but reduces sample size.
- Multiple regression can adjust for several variables at once; watch for multicollinearity, insufficient sales, and time-trend distortion.
- Sales chasing is improper: selecting only sales that support a target value biases the study.
Worked example
Five sales have ratios of 0.92, 0.96, 1.00, 1.05, and 1.07. The median is 1.00. Absolute deviations are 0.08, 0.04, 0.00, 0.05, and 0.07, totaling 0.24. Average deviation = 0.24 ÷ 5 = 0.048. COD = 0.048 ÷ 1.00 × 100 = 4.8%. That indicates tight uniformity; a COD of 20% would show much wider dispersion.
Traps
Using the mean when outliers distort it; dropping inconvenient sales; confusing COD with level; computing ratios with stale sale dates and no time adjustment; and treating a small sample as statistically reliable.