9.1 Paired Data Analysis

Where Paired Data Fits in the Sales Comparison Approach

AQB Content Area IV — Sales Comparison Approach (ECO effective April 1, 2026) weights this domain heavily—especially for Licensed Residential (28 of 110 scored items, 25.4%). After you select comparables, choose units of comparison, and identify elements of comparison (Chapter 8), you must support quantitative adjustments. Under IV.d Quantitative adjustments, the first listed technique is paired data.

Paired data analysis (also called paired sales analysis) compares two or more sales that are highly similar except for one primary difference. The price difference between the sales is attributed to that feature, producing a market-derived adjustment that can be applied to other comparables when they differ from the subject for the same element.

Paired data does not replace judgment. It is a market extraction tool. Weak pairs produce weak adjustments. Strong pairs—recent, local, arm’s-length, similar in other respects—produce the most defensible grid entries on a URAR-style sales comparison grid or a narrative SCA grid.

Core Logic and Adjustment Direction

Always adjust the comparable toward the subject, never the subject’s price toward the comparable.

Memory aid: Bring the comparable’s price to what it would have sold for if it were like the subject.

Simple Isolation Formula

When Sale A and Sale B are otherwise equivalent:

Indicated contribution of Feature X ≈ Sale price of property with X − Sale price of property without X

(Or, more carefully: price of the sale that has the superior quantity of X minus price of the sale with the lesser quantity, then scale to the subject’s difference.)

If you work in percentages:

Percentage difference ≈ (Price_superior − Price_inferior) / Price_inferior (or sometimes divided by a base both share—be consistent with how you will apply the adjustment).

Dollar adjustments are common for discrete items (garage, bath); dollars per square foot or percentage of price are common for continuous size differences (GLA).

Conditions for a Credible Pair

Before you trust a pair on exam day or in practice, check:

1. Arm’s-length, market sales — not REO forced sales mixed with retail without analysis, not related-party transfers.
2. Same market segment — similar location, price tier, design, and buyer pool.
3. Similar time or an already-supported market-conditions (time) adjustment so residual price difference is not pure appreciation.
4. One primary property difference after transactional adjustments (property rights, financing, conditions of sale, expenditures after sale, market conditions).
5. Physical/locational similarity on other major elements—style, quality, condition, lot utility, view, basement, etc.
6. Replication — ideally more than one independent pair pointing the same direction and magnitude.

Transactional adjustments are applied before you attribute residual differences to physical features. Sequence matters: if one sale had seller concessions or non-market financing, fix that first so the “pair difference” is not contaminated.

Worked Example 1: Garage Adjustment

Market facts: Two nearly identical 1,600 sq ft ranch homes on the same street, sold within 30 days, both conventional financing, no concessions, same condition and quality.

Extracted garage contribution:

$410,000 − $388,000 = $22,000 for a 2-car attached garage versus none.

Apply to a comparable used for the subject:

• Subject has a 2-car attached garage.
• Comparable 3 sold for $405,000 with no garage, otherwise similar after other adjustments.
• Garage line: comparable is inferior → +$22,000.
• Adjusted price contribution from garage: $405,000 + $22,000 = $427,000 (before other lines).

Partial garage case: If a second pair shows 1-car vs 2-car at a $9,000 difference in the same market, you may support a graduated garage schedule (none / 1-car / 2-car) rather than a single binary number. Exam stems often give one clean pair; use what is given.

Trap: Cost to build a 2-car garage might be $35,000. Contribution is $22,000 from the pair—not construction cost. Principle of contribution: market pays what the market pays.

Worked Example 2: Bathroom Adjustment

Market facts: Paired sales of similar 3-bedroom two-story homes, same subdivision, sales three weeks apart, both cash-equivalent after financing analysis.

Extracted half-bath / bath contribution:

$475,000 − $465,000 = $10,000 for the additional half bath (or for going from 2.0 to 2.5, as the stem frames it).

Subject application:

• Subject: 2.5 baths.
• Comparable 1: 2.0 baths, sale price $470,000, otherwise matched.
• Bath line: comparable inferior → +$10,000.

Full bath pair (second illustration): Suppose another pair isolates a full third bath:

Difference = $22,000 for a full additional bath versus 2.0. That does not mean every half bath is automatically $11,000—half baths and full baths are different utility packages. Do not invent linear rules the stem does not support.

Exam tip: If two pairs conflict ($8,000 vs $14,000 for similar bath differences), reconcile by quality of match, recency, and location—or report a range and weight the better pairs. Blind averaging of bad pairs is still bad analysis.

Worked Example 3: GLA (Gross Living Area) Adjustment

Size is continuous, so pairs often produce a $/sq ft of GLA difference rather than a lump sum for “having GLA.”

Market facts: Two highly similar homes, same quality/condition/garage/bath count; only GLA differs materially.

Price difference: $450,000 − $430,000 = $20,000
GLA difference: 1,800 − 1,600 = 200 sq ft
Indicated GLA adjustment rate: $20,000 ÷ 200 = $100 per sq ft of GLA difference

Apply to subject vs a comparable:

• Subject GLA: 1,700 sq ft
• Comparable 2: 1,850 sq ft, sale price $455,000
• GLA difference: comparable has 150 sq ft more than subject → comparable superior
• Adjustment: 150 × $100 = $15,000 downward on Comparable 2

Check reasonableness: $100/sq ft of difference is not the same as overall price per square foot of the whole house ($430,000 ÷ 1,600 ≈ $269/sq ft). Overall unit prices include land and fixed components; marginal GLA adjustments from pairs are often lower than average price per square foot. Exam items love that distinction.

Mini Grid Using Extracted Adjustments

Assume subject: 1,700 sq ft, 2.5 baths, 2-car garage. One comparable for illustration:

(Other elements assumed equal for this teaching grid.)

Building a Pair Set, Not a Single Pair

Best practice and exam reasoning:

If the market is thin, appraisers may combine paired data with cost, income, regression, or qualitative methods (next section)—not invent precision from a single noisy pair.

Common Traps (High-Yield Exam Content)

1. Insufficient Pairs

One pair can reflect buyer eccentricity, an undetected defect, or measurement error. Insufficient pairs means you lack a replicated market signal. On the exam, the correct critique of an appraiser who uses one odd sale pair for a $50,000 “view” adjustment with no other support is that the sample is inadequate and the difference may not isolate view alone.

2. Correlated Differences (Confounded Pairs)

Correlated differences occur when two (or more) features change together, so you cannot isolate one cause.

Example of a bad “pair”:

Price gap = $45,000. Is that GLA? Garage? Condition? All three? You cannot extract a pure GLA rate from this pair. The differences are correlated (confounded).

Exam recognition phrase: If the stem says two sales differ in size and basement and updates, any single-feature extraction is unsupported unless other analysis separates the effects.

3. Ignoring Transactional Differences

A “pair” where one sale had $15,000 in financing concessions and the other did not will overstate or understate the physical feature if concessions are not adjusted first.

4. Wrong Direction on the Grid

Students often add when they should subtract. Re-read: adjust the comparable to the subject. Superior comparable → lower adjusted price.

5. Using List Prices or Asking Prices as Pairs

Paired data for market value adjustments should rest on closed sales (or other market evidence of what was paid), not unmatched listings, unless the problem is specifically about listing behavior.

6. Cost Equals Value Fallacy

Builder cost for a feature is not automatically the paired-data adjustment. Contribution can be less (or occasionally more in a shortage) than cost.

7. Cross-Market Transfer Without Support

A $25,000 pool adjustment extracted in a luxury waterfront submarket may not transfer to a starter-home inland market where pools add little or even hurt some buyers.

Qualitative Note (Bridge)

When pairs cannot isolate a difference cleanly, you may still know the comparable is superior/inferior/equal and handle the residual in reconciliation or with ranking / relative comparison (Section 9.2) rather than forcing a fake dollar line. Quantitative paired data is preferred when the market gives you clean isolation; it is not mandatory to invent a number when isolation fails.

Process Checklist for Exam Items

1. Identify what feature the question wants isolated.
2. Confirm the two sales are alike on other major points (or adjust transactional items first).
3. Subtract prices in the correct order for contribution.
4. Convert to $/unit if size or count differs from the subject’s gap.
5. Apply to the comparable, with correct sign.
6. Ask: Could another difference explain the gap? If yes, the pair is weak.

Master paired data and you own the heart of ECO IV.d.1. The next section adds the other quantitative tools and the qualitative comparison methods when pairs are thin or features are hard to dollarize.