Paired Data Analysis
~11 min read · Isolate one difference between sales to extract a defensible adjustment amount.
Paired data analysis is where adjustment dollars come from: find two sales identical except for ONE feature, and the price difference IS that feature's market value. The exam gives you small sale sets and asks you to isolate the variable.
The method
Paired sales (matched pairs): locate sales alike in all respects but one — same model with and without the pool, same house sold before and after the arterial opened — and attribute the price difference to that lone difference. The extracted amount is the feature's contributory value, the only legitimate source for grid adjustments (cost is not value; the pool's invoice proves nothing). Multiple pairs pointing to a range beat one pair pointing to a number; reconcile the indications.
- Two sales, one difference → the difference prices the feature
- Output = contributory value, the adjustment amount
- Several pairs → a supported range; one pair → a hint
Multi-step isolation
Perfect pairs are rare; practical pairing subtracts KNOWN adjustments first: if two sales differ by a garage AND three months of a rising market, time-adjust the older sale with the extracted monthly rate, THEN read the residual as the garage. Each layer of inference adds noise — the exam's multi-step problems test whether you clean the pair before reading it. Sale-resale pairs of the SAME property isolate market conditions the same way (no physical difference → all difference is time).
- Adjust for known differences first; read the residual
- Sale-resale of one property isolates the time variable
- Every extra inference widens the error band
Alternatives and support
Where pairs run out: grouped data analysis (average price differences across grouped sales with/without the feature), statistical/regression support (the GLA coefficient as a per-square-foot adjustment), cost-less-depreciation as a ceiling check, and income capitalization of rent differences for income property. USPAP cares that adjustments are SUPPORTED — paired data is the gold standard, the others corroborate.
Worked example
Extract the value of a third-car garage bay: Sale 1 — two-car garage, sold January at $460,000. Sale 2 — same model, three-car garage, sold May at $486,000. Sale-resale evidence in the tract shows the market rising 0.5%/month. Isolate the bay.
The pair differs in TWO ways: the bay and four months of market. Clean the time layer first: Sale 1 time-adjusted to May = 460,000 × (1 + 0.005 × 4) = 460,000 × 1.02 = $469,200. Now the pair differs only by the bay: 486,000 − 469,200 = $16,800 — the third bay's contributory value, call it $16,500–$17,000 for the grid pending another pair. The classic error: reading the raw $26,000 difference as the garage — overstating it by the market's four-month climb. Clean first, read second.
Common exam pitfalls
Reading raw pair differences with a second variable in play.
Adjust for known differences (time, concessions) BEFORE attributing the residual to the target feature.
Using construction cost as the adjustment.
Paired data measures what the market PAYS — contribution, which routinely diverges from cost.
Betting the grid on a single pair.
One pair is an indication; multiple pairs and corroborating methods make it support.
Same-same-except-one: clean the pair, read the gap, and let several gaps agree before you believe them.
Recap
- Matched pairs isolate one feature's price effect
- Extracted difference = contributory value = the adjustment
- Time-adjust and concession-clean pairs before reading
- Sale-resale pairs extract market-conditions rates
- Grouped data, regression, cost, and rent capitalization corroborate
- Support is the standard; pairing is the gold version
VIDEO LESSON
Watch this lesson on video — free
Prove it: 10 questions on this topic
Every lesson ends with a ten-question check in the free course — your progress syncs between the web and the EstatePass app.
Studying for the appraiser licensing exam? Track every lesson free — progress syncs with the app.
Start freeMore in Sales Comparison
Study smarter in the free dashboard
- Every lesson tracked, synced with the iOS app
- Ten-question checks after each lesson
- Lesson videos, flashcards and mock exams
No credit card required.