Three Creeks
This study looks at the 23 resale homes sold in Three Creeks in the 365 days to October 25, 2026. A full year is used because only 13 resale homes sold in the last 180 days. New construction is left out and shown on its own in section 7. The homes are compared four ways: by price, by living area, by year built and with or without a view. The first three are split into four equal groups (quartiles), each named by its range.
Each comparison has its own section (3–6) with everything about it in one place: the key differences, a table of every measure at a glance, the charts and the numbers.
Each dot is one resale home, colored by its price range. Use the buttons to show one price range at a time. Click a dot for that sale: address, close date, original and last list price, net sale price, living area, net sale per SF and days on market. Scroll or pinch to zoom.
Sales ranked by net sale price and cut into four equal groups. This is the usual way to read a market by price.
- Homes over $399K are about 870 SF (56%) larger than homes under $273K.
3.1 Key differences
The top group against the bottom group. Only real differences are listed. Section 8 explains each measure.
Higher-priced homes Over $399K
- Are much larger.
Lower-priced homes Under $273K
- Are much smaller.
3.2 At a glance
Shaded rows are real differences; the darkest cell is the highest. Gray rows show no clear difference. Prices are averages; days on market and year built are medians.
| Under $273K | $273K–$304K | $319K–$399K | Over $399K | Real difference? | |
|---|---|---|---|---|---|
| Net price per SF | $152 per SF | $152 per SF | $143 per SF | $179 per SF | unclear |
| Living area | 1,567 SFSmallest | 1,892 SF | 2,527 SFLargest | 2,440 SF | ✓ large |
| Year built | 2020 | 2020 | 2018 | 2021 | unclear |
| Days on market | 76 days | 88 days | 116 days | 10 days | unclear |
| Discount from original list | 8.2% | 7.6% | 10.6% | 1.9% | unclear |
| Price cut before sale | 83% | 67% | 80% | 17% | unclear |
| Seller paid closing costs or repairs | 67% | 67% | 100% | 100% | unclear |
| Has a view | 0% | 17% | 40% | 33% | unclear |
How to read this section. Each heading gives the finding. The gold box says whether the gap is real (✓, with its size) or could be chance (–). The section ends with its numbers; the tests are in section 8.3.
3.3 Net price per SF: no clear difference
The average is $179 per SF for homes over $399K and $152 per SF for homes under $273K. It does not move in a straight line; homes over $399K are highest. Defined in 8.1.2.
3.4 Homes over $399K are about 870 SF (56%) larger than homes under $273K
The average is 2,440 SF for homes over $399K and 1,567 SF for homes under $273K. It does not move in a straight line; homes $319K–$399K are highest. Defined in 8.1.3.
3.5 Year built: no clear difference
The median is 2021 for homes over $399K and 2020 for homes under $273K. It does not move in a straight line; homes over $399K are highest. Defined in 8.1.4.
3.6 Days on market: no clear difference
The median is 10 days for homes over $399K and 76 days for homes under $273K. It does not move in a straight line; homes $319K–$399K are highest. Defined in 8.1.5.
3.7 Discount from original list: no clear difference
The average is 1.9% for homes over $399K and 8.2% for homes under $273K. It does not move in a straight line; homes $319K–$399K are highest. Defined in 8.1.6.
3.8 How often: price cut before sale, seller paid closing costs or repairs, has a view
Each bar is the share of homes in the group, in percent.
- 17% of homes over $399K had a price cut before sale, against 83% of homes under $273K. Defined in 8.1.7.
- 100% of homes over $399K had the seller pay closing costs or repairs, against 67% of homes under $273K. Defined in 8.1.8.
- 33% of homes over $399K have a view, against 0% of homes under $273K. Defined in 8.1.9.
3.9 The numbers: price and size
| Under $273K6 sales | $273K–$304K6 sales | $319K–$399K5 sales | Over $399K6 sales | All23 sales | Real difference? | |
|---|---|---|---|---|---|---|
| Net sale price average | $236,986 | $284,513 | $352,686 | $421,102 | $322,567 | by design |
| Net price per SF average | $152 | $152 | $143 | $179 | $157 | unclear |
| Living area average | 1,567 | 1,892 | 2,527 | 2,440 | 2,088 | ✓ large |
| Year built median | 2020 | 2020 | 2018 | 2021 | 2020 | unclear |
3.10 The numbers: selling
| Under $273K6 sales | $273K–$304K6 sales | $319K–$399K5 sales | Over $399K6 sales | All23 sales | Real difference? | |
|---|---|---|---|---|---|---|
| Days on market median | 76 | 88 | 116 | 10 | 75 | unclear |
| Discount from original list average | 8.2% | 7.6% | 10.6% | 1.9% | 6.9% | unclear |
| Price cut before sale | 83% | 67% | 80% | 17% | 61% | unclear |
| Seller paid closing costs or repairs | 67% | 67% | 100% | 100% | 83% | unclear |
| Has a view | 0% | 17% | 40% | 33% | 22% | unclear |
Prices, price per SF, living area and discount are averages; days on market and year built are medians. The test behind each verdict is in 8.3.1.
Sales ranked by living area and cut into four equal groups, from the smallest homes to the largest.
- Homes over 2,205 SF sell for about $143K (59%) more than homes under 1,680 SF.
4.1 Key differences
The top group against the bottom group. Only real differences are listed. Section 8 explains each measure.
Larger homes Over 2,205 SF
- Sell for much more.
Smaller homes Under 1,680 SF
- Sell for much less.
4.2 At a glance
Shaded rows are real differences; the darkest cell is the highest. Gray rows show no clear difference. Prices are averages; days on market and year built are medians.
| Under 1,680 SF | 1,680–2,078 SF | 2,090–2,205 SF | Over 2,205 SF | Real difference? | |
|---|---|---|---|---|---|
| Net sale price | $241KLowest | $325K | $344K | $384KHighest | ✓ large |
| Net price per SF | $156 per SF | $171 per SF | $159 per SF | $142 per SF | unclear |
| Year built | 2019 | 2020 | 2020 | 2020 | unclear |
| Days on market | 80 days | 81 days | 22 days | 20 days | unclear |
| Discount from original list | 8.0% | 6.5% | 7.8% | 5.7% | unclear |
| Price cut before sale | 83% | 67% | 60% | 33% | unclear |
| Seller paid closing costs or repairs | 67% | 67% | 100% | 100% | unclear |
| Has a view | 0% | 17% | 40% | 33% | unclear |
4.3 Homes over 2,205 SF sell for about $143K (59%) more than homes under 1,680 SF
The average is $384K for homes over 2,205 SF and $241K for homes under 1,680 SF. It rises step by step from the lowest range to the highest. Defined in 8.1.1.
4.4 Net price per SF: no clear difference
The average is $142 per SF for homes over 2,205 SF and $156 per SF for homes under 1,680 SF. It does not move in a straight line; homes 1,680–2,078 SF are highest. Defined in 8.1.2.
4.5 Year built: no clear difference
The median is 2020 for homes over 2,205 SF and 2019 for homes under 1,680 SF. It does not move in a straight line; homes 1,680–2,078 SF are highest. Defined in 8.1.4.
4.6 Days on market: no clear difference
The median is 20 days for homes over 2,205 SF and 80 days for homes under 1,680 SF. It does not move in a straight line; homes 1,680–2,078 SF are highest. Defined in 8.1.5.
4.7 Discount from original list: no clear difference
The average is 5.7% for homes over 2,205 SF and 8.0% for homes under 1,680 SF. It does not move in a straight line; homes under 1,680 SF are highest. Defined in 8.1.6.
4.8 How often: price cut before sale, seller paid closing costs or repairs, has a view
Each bar is the share of homes in the group, in percent.
- 33% of homes over 2,205 SF had a price cut before sale, against 83% of homes under 1,680 SF. Defined in 8.1.7.
- 100% of homes over 2,205 SF had the seller pay closing costs or repairs, against 67% of homes under 1,680 SF. Defined in 8.1.8.
- 33% of homes over 2,205 SF have a view, against 0% of homes under 1,680 SF. Defined in 8.1.9.
4.9 The numbers: price and size
| Under 1,680 SF6 sales | 1,680–2,078 SF6 sales | 2,090–2,205 SF5 sales | Over 2,205 SF6 sales | All23 sales | Real difference? | |
|---|---|---|---|---|---|---|
| Net sale price average | $240,819 | $325,422 | $344,078 | $383,534 | $322,567 | ✓ large |
| Net price per SF average | $156 | $171 | $159 | $142 | $157 | unclear |
| Living area average | 1,543 | 1,893 | 2,156 | 2,772 | 2,088 | by design |
| Year built median | 2019 | 2020 | 2020 | 2020 | 2020 | unclear |
4.10 The numbers: selling
| Under 1,680 SF6 sales | 1,680–2,078 SF6 sales | 2,090–2,205 SF5 sales | Over 2,205 SF6 sales | All23 sales | Real difference? | |
|---|---|---|---|---|---|---|
| Days on market median | 80 | 81 | 22 | 20 | 75 | unclear |
| Discount from original list average | 8.0% | 6.5% | 7.8% | 5.7% | 6.9% | unclear |
| Price cut before sale | 83% | 67% | 60% | 33% | 61% | unclear |
| Seller paid closing costs or repairs | 67% | 67% | 100% | 100% | 83% | unclear |
| Has a view | 0% | 17% | 40% | 33% | 22% | unclear |
Prices, price per SF, living area and discount are averages; days on market and year built are medians. The test behind each verdict is in 8.3.2.
Sales ranked by year built and cut into four groups, from the oldest to the newest. Homes built in the same year always stay in the same group, so group sizes can differ a little.
5.1 Key differences
The top group against the bottom group: no measure differs enough to rule out chance.
5.3 Net sale price: described, not tested
The average is $357K for homes built after 2021 and $310K for homes built before 2019. It does not move in a straight line; homes built after 2021 are highest. Defined in 8.1.1.
5.4 Net price per SF: described, not tested
The average is $178 per SF for homes built after 2021 and $150 per SF for homes built before 2019. It does not move in a straight line; homes built after 2021 are highest. Defined in 8.1.2.
5.5 Living area: described, not tested
The average is 1,994 SF for homes built after 2021 and 2,071 SF for homes built before 2019. It does not move in a straight line; homes built 2019 are highest. Defined in 8.1.3.
5.6 Days on market: described, not tested
The median is 70 days for homes built after 2021 and 22 days for homes built before 2019. It does not move in a straight line; homes built 2019 are highest. Defined in 8.1.5.
5.7 Discount from original list: described, not tested
The average is 8.1% for homes built after 2021 and 7.6% for homes built before 2019. It does not move in a straight line; homes built after 2021 are highest. Defined in 8.1.6.
5.8 How often: price cut before sale, seller paid closing costs or repairs, has a view
Each bar is the share of homes in the group, in percent.
- 75% of homes built after 2021 had a price cut before sale, against 43% of homes built before 2019. Defined in 8.1.7.
- 100% of homes built after 2021 had the seller pay closing costs or repairs, against 86% of homes built before 2019. Defined in 8.1.8.
- 0% of homes built after 2021 have a view, against 14% of homes built before 2019. Defined in 8.1.9.
5.9 The numbers: price and size
| Before 20197 sales | 20193 sales | 2020–20219 sales | After 20214 sales | All23 sales | |
|---|---|---|---|---|---|
| Net sale price average | $310,421 | $303,800 | $323,142 | $356,604 | $322,567 |
| Net price per SF average | $150 | $146 | $156 | $178 | $157 |
| Living area average | 2,071 | 2,226 | 2,098 | 1,994 | 2,088 |
| Year built median | 2018 | 2019 | 2021 | 2022 | 2020 |
5.10 The numbers: selling
| Before 20197 sales | 20193 sales | 2020–20219 sales | After 20214 sales | All23 sales | |
|---|---|---|---|---|---|
| Days on market median | 22 | 84 | 75 | 70 | 75 |
| Discount from original list average | 7.6% | 4.7% | 6.7% | 8.1% | 6.9% |
| Price cut before sale | 43% | 67% | 67% | 75% | 61% |
| Seller paid closing costs or repairs | 86% | 67% | 78% | 100% | 83% |
| Has a view | 14% | 0% | 44% | 0% | 22% |
Prices, price per SF, living area and discount are averages; days on market and year built are medians. The test behind each verdict is in 8.3.3.
Sales where the MLS lists a view (greenbelt or park, trees, golf course, hill country, water and similar) against those listing none or only the neighborhood.
6.1 Key differences
The top group against the bottom group: no measure differs enough to rule out chance.
6.2 At a glance
Shaded rows are real differences; the darkest cell is the highest. Gray rows show no clear difference. Prices are averages; days on market and year built are medians.
| View | No view | Real difference? | |
|---|---|---|---|
| Net sale price | $371K | $309K | unclear |
| Net price per SF | $166 per SF | $154 per SF | unclear |
| Living area | 2,237 SF | 2,047 SF | unclear |
| Year built | 2021 | 2020 | unclear |
| Days on market | 20 days | 81 days | unclear |
| Discount from original list | 3.8% | 7.8% | unclear |
| Price cut before sale | 40% | 67% | unclear |
| Seller paid closing costs or repairs | 100% | 78% | unclear |
6.3 Net sale price: no clear difference
The average is $371K for homes with a view and $309K for homes without a view. Defined in 8.1.1.
6.4 Net price per SF: no clear difference
The average is $166 per SF for homes with a view and $154 per SF for homes without a view. Defined in 8.1.2.
6.5 Living area: no clear difference
The average is 2,237 SF for homes with a view and 2,047 SF for homes without a view. Defined in 8.1.3.
6.6 Year built: no clear difference
The median is 2021 for homes with a view and 2020 for homes without a view. Defined in 8.1.4.
6.7 Days on market: no clear difference
The median is 20 days for homes with a view and 81 days for homes without a view. Defined in 8.1.5.
6.8 Discount from original list: no clear difference
The average is 3.8% for homes with a view and 7.8% for homes without a view. Defined in 8.1.6.
6.9 How often: price cut before sale, seller paid closing costs or repairs
Each bar is the share of homes in the group, in percent.
- 40% of homes with a view had a price cut before sale, against 67% of homes without a view. Defined in 8.1.7.
- 100% of homes with a view had the seller pay closing costs or repairs, against 78% of homes without a view. Defined in 8.1.8.
6.10 The numbers: price and size
| View5 sales | No view18 sales | All23 sales | Real difference? | |
|---|---|---|---|---|
| Net sale price average | $371,384 | $309,007 | $322,567 | unclear |
| Net price per SF average | $166 | $154 | $157 | unclear |
| Living area average | 2,237 | 2,047 | 2,088 | unclear |
| Year built median | 2021 | 2020 | 2020 | unclear |
6.11 The numbers: selling
| View5 sales | No view18 sales | All23 sales | Real difference? | |
|---|---|---|---|---|
| Days on market median | 20 | 81 | 75 | unclear |
| Discount from original list average | 3.8% | 7.8% | 6.9% | unclear |
| Price cut before sale | 40% | 67% | 61% | unclear |
| Seller paid closing costs or repairs | 100% | 78% | 83% | unclear |
Prices, price per SF, living area and discount are averages; days on market and year built are medians. The test behind each verdict is in 8.3.4.
Homes sold new by the builder in the same 365 days. They are left out of every other section and shown here as basic figures only: 8 sales, no groups, no tests.
| Measure | Average | Median |
|---|---|---|
| Sales | 8 | |
| Net sale price | $314,050 | $313,954 |
| Net price per SF | $158 | $159 |
| Living area (SF) | 1,998 | 1,962 |
| Days on market | 120 | 112 |
| Discount from original list | 3.6% | 3.6% |
| Price cut before sale | 0% of sales | |
| Seller paid closing costs or repairs | 100% of sales | |
| Year built | 2025–2025 | |
Same definitions as the rest of the study (section 8). New construction = NewConstructionYN true, or PropertyCondition “New Construction” or “Under Construction”.
Each measure and method three ways: in plain words, as statistics, and with the data from this study.
8.1 Measures
8.1.1 Net sale price
- In simple terms
- What the seller actually kept: the closing price, less money the seller gave back to the buyer for closing costs or repairs. It is the truest price for comparing homes.
- Statistically
- Net = ClosePrice − ACTRIS_BuyersClsgCostPdbySell − ACTRIS_RepairsAmount (a blank counts as $0). Summarized by the arithmetic mean in the slices.
- The data
- 23 resale sales. Average $322,567; median $304,000; standard deviation $75,640; middle half (25th–75th percentile) $261,750–$399,062; lowest $190,000, highest $465,000. 19 sales had a seller payment; their median payment was $10,000.
8.1.2 Net price per SF
- In simple terms
- Net sale price divided by the heated and cooled square feet. It puts homes of different sizes on the same footing: did buyers pay more for each square foot, not just for more square feet?
- Statistically
- psf = Net ÷ LivingArea, per sale; summarized by the mean of the per-sale ratios.
- The data
- 23 resale sales. Average $157; median $152; standard deviation $27; middle half (25th–75th percentile) $142–$177; lowest $102, highest $209.
8.1.3 Living area
- In simple terms
- The heated and cooled space, as listed in the MLS. Garages and porches are not counted.
- Statistically
- RESO LivingArea, in square feet. Sales under 400 SF are treated as entry errors and left out.
- The data
- 23 resale sales. Average 2,088; median 2,078; standard deviation 523; middle half (25th–75th percentile) 1,658–2,264; lowest 1,413, highest 3,509.
8.1.4 Year built
- In simple terms
- The year the home was finished, as listed in the MLS. It shows whether a group leans to older or newer phases of the community.
- Statistically
- RESO YearBuilt; summarized by the median, with the 25th–75th percentile (middle half) on the charts.
- The data
- 23 resale sales. Average 2020; median 2020; standard deviation 2; middle half (25th–75th percentile) 2018–2021; lowest 2016, highest 2022.
8.1.5 Days on market
- In simple terms
- How many days the home was on the market before the buyer's offer was accepted, counting any earlier listing periods (relists). This is the days on market agents and clients quote.
- Statistically
- CDOM = max(CumulativeDaysOnMarket, DaysOnMarket, PurchaseContractDate − ListingContractDate). The MLS counter can freeze at a price change, so the contract-date gap is a floor. Summarized by the median, because a few homes that sat for months pull an average up.
- The data
- 23 resale sales. Average 77; median 75; standard deviation 74; middle half (25th–75th percentile) 18–118; lowest 5, highest 271. 0 sales went under contract on their listing date (0 days).
8.1.6 Discount from original list
- In simple terms
- How far below the first asking price the home sold, after seller payments. It measures how much negotiating room sellers gave up. A negative figure means the home sold above its first asking price.
- Statistically
- Discount % = (OriginalListPrice − Net) ÷ OriginalListPrice × 100, per sale; summarized by the mean.
- The data
- 23 resale sales. Average 6.9%; median 5.8%; standard deviation 5.3%; middle half (25th–75th percentile) 3.0%–10.4%; lowest −0.7%, highest 20.1%. 1 sales closed above their first asking price.
8.1.7 Price cut before sale
- In simple terms
- The asking price was lowered at least once before the home went under contract.
- Statistically
- Indicator: ListPrice at close < OriginalListPrice. Reported as the share of sales (%).
- The data
- 14 of 23 resale sales (60.9%).
8.1.8 Seller paid closing costs or repairs
- In simple terms
- The seller paid some of the buyer's closing costs or repairs.
- Statistically
- Indicator: ACTRIS_BuyersClsgCostPdbySell + ACTRIS_RepairsAmount > 0. Reported as the share of sales (%).
- The data
- 19 of 23 resale sales (82.6%).
8.1.9 Has a view
- In simple terms
- The MLS says the home looks over something: a greenbelt or park, trees, golf course, hill country, water and similar.
- Statistically
- View = the View field holds any entry other than “None”, “Neighborhood” or “See Remarks”. No view = only “None” and/or “Neighborhood”. A record whose only entry is “See Remarks” cannot be classified and is left out of the view slice.
- The data
- 5 sales with a view, 18 without, 0 unclassified. Most common View entries: None 11, Neighborhood 10, Park/Greenbelt 3, Canyon 1, Garden 1, Trees/Woods 1, Hill Country 1, City 1.
8.2 Methods
8.2.1 Data source
- In simple terms
- Every figure comes from the Austin Board of REALTORS® MLS, read live by ALEX. No figure is estimated or typed in.
- Statistically
- Bridge Interactive RESO Web API (OData), ABOR dataset, every field returned. Filter: StandardStatus = Closed, PropertyType = Residential, City = Belton, SubdivisionName contains any of “three creeks”, “three crks” (case-insensitive). A record is kept when its subdivision name belongs to the Three Creeks family, the same rule comparable sales use (it includes Three Crks Ph Vi, proven adjacent and priced alike).
- The data
- Pulled October 26, 2026: 32 closed Three Creeks records over the prior year, in ZIP 76513. 10 subdivision spellings kept: Three Creeks Ph I (11); Three Creeks (11); Three Crks Ph Vi (2); Three Crks Ph Iv (2); THREE CREEKS PHASE IV (1); THREE CREEKS PHASE I (1); THREE CREEKS (1); Three Crks Ph V (1); Three Creeks Ph II (1); Three Creeks Ph Ii (1). No other names were found by the search.
8.2.2 Study window
- In simple terms
- The most recent 365 days of closed sales: recent enough to describe today's market, long enough to give each group plenty of sales. The usual window is 180 days; it holds only 13 resale sales here (under 80), so the full year is used.
- Statistically
- CloseDate from October 26, 2025 to October 25, 2026 inclusive.
- The data
- 31 single-family sales in the window: 23 resale, 8 new construction.
8.2.3 Resale and new construction
- In simple terms
- A resale home was previously owned. New construction is sold by the builder and behaves differently, so it is kept out of every analysis and reported once on its own.
- Statistically
- New construction = NewConstructionYN true, or PropertyCondition contains “New Construction” or “Under Construction”. Everything else is resale.
- The data
- 23 resale; 8 new construction.
8.2.4 Sales left out
- In simple terms
- A few records cannot be used as entered and are left out before any analysis.
- Statistically
- Excluded: PropertySubType other than Single Family Residence; no OriginalListPrice; LivingArea under 400 SF; seller payments above 10% of ClosePrice (treated as entry errors).
- The data
- 1 sale(s) with concessions above 10% of price
8.2.5 Quartiles and groups
- In simple terms
- Sales lined up from lowest to highest and cut into four groups of about equal size. Each group is named by its range, for example “Under $273K”.
- Statistically
- Sort by the slicing variable; cut k sits at round(nk/4), k = 1…3. A cut that would split tied values moves to the nearer edge of the tie. A group under 3 sales merges into its smaller neighbor. Range names use the next group's lowest value (“Under”) and the previous group's highest value (“Over”/“After”).
- The data
- Quartiles by net sale price: Under $273K: $190,000–$250,500 (6 sales); $273K–$304K: $273,000–$304,000 (6 sales); $319K–$399K: $318,855–$398,675 (5 sales); Over $399K: $399,450–$465,000 (6 sales)
Quartiles by living area: Under 1,680 SF: 1,413–1,635 SF (6 sales); 1,680–2,078 SF: 1,680–2,078 SF (6 sales); 2,090–2,205 SF: 2,090–2,205 SF (5 sales); Over 2,205 SF: 2,323–3,509 SF (6 sales)
Quartiles by year built: Before 2019: 2016–2018 (7 sales); 2019: 2019–2019 (3 sales); 2020–2021: 2020–2021 (9 sales); After 2021: 2022–2022 (4 sales)
View and no view: View: MLS lists a view (5 sales); No view: None or Neighborhood (18 sales)
8.2.6 Average and median
- In simple terms
- The average adds the values and divides by how many; one very expensive home can pull it up. The median is the middle value, with half above and half below.
- Statistically
- Mean x̄ = (1/n) Σ xᵢ; median = 50th percentile. Price, price per SF, living area and discount use the mean; days on market and year built use the median (skewed or whole-year values).
- The data
- Net sale price: average $322,567, median $304,000. Days on market: average 77, median 75: the average is pulled up by a few long listings.
8.2.7 Kruskal–Wallis test
- In simple terms
- Checks whether a measure really differs across the four groups of a slice. It uses each sale's rank, not its value, so a few extreme sales cannot drive the answer.
- Statistically
- H = [12/(n(n+1))] Σ Rⱼ²/nⱼ − 3(n+1), compared with χ² on k − 1 degrees of freedom (Kruskal & Wallis, 1952). Effect size ε² = H/(n − 1): ≥ .01 small, ≥ .06 medium, ≥ .14 large (Tomczak & Tomczak, 2014). A slice is tested only when every group has at least 5 sales; smaller groups are described, not tested.
- The data
- Run on every measure in sections 3–5; each result is in 8.3.
8.2.8 Mann–Whitney U test
- In simple terms
- The same check when there are only two groups: view against no view.
- Statistically
- Two-sided U test (Mann & Whitney, 1947). Effect size = rank-biserial correlation |1 − 2U/(n₁n₂)|: ≥ .10 small, ≥ .30 medium, ≥ .50 large (Kerby, 2014; Cohen, 1988).
- The data
- Run on every measure in section 6; results in 8.3.
8.2.9 Chi-square test
- In simple terms
- Checks whether a yes/no share (price cut, seller payment, view) differs between groups more than chance would.
- Statistically
- Pearson χ² on the 2 × k table of counts, no continuity correction (Pearson, 1900). Effect size Cramér's V = √(χ²/n): ≥ .10 small, ≥ .30 medium, ≥ .50 large (Cramér, 1946).
- The data
- Run on the three shares in every slice; results in 8.3.
8.2.10 Real difference (significance)
- In simple terms
- “Real difference” means a gap this large would be very unlikely if the groups were really alike. Because many measures are checked in each slice, the bar is set higher than usual.
- Statistically
- Significant when p < 0.05 ÷ m, where m is the number of measures tested in that slice (Bonferroni; Dunn, 1961). The slicing variable itself differs by design and is not tested.
- The data
- Quartiles by net sale price: m = 8, threshold p < 0.0063; Quartiles by living area: m = 8, threshold p < 0.0063; View and no view: m = 8, threshold p < 0.0063
8.2.11 Size of a difference
- In simple terms
- A difference can be real but small. The size word (small, medium, large) says how much the groups pull apart.
- Statistically
- ε² for Kruskal–Wallis, rank-biserial r for Mann–Whitney, Cramér's V for chi-square, with the thresholds above; below the small threshold the size is “negligible”.
- The data
- 2 real differences across the four slices; sizes are listed in 8.3.
8.2.12 Rounding in the text
- In simple terms
- Headings and sentences round figures so they read easily; the tables at the end of each section give them in full.
- Statistically
- Prices in headings round to the nearest $1K; price per SF and days on market to the nearest whole number; living area to the nearest 10 SF; shares and percentage gaps to whole percents.
- The data
- Days on market is always shown in days.
8.3 Test results
Every test behind every gold box. Statistic, degrees of freedom (df), p-value, the threshold it must beat, and the size of the difference.
8.3.1 Quartiles by net sale price
| Measure | Test | Statistic | df | p-value | Threshold | Real difference? | Effect size | Size |
|---|---|---|---|---|---|---|---|---|
| Net price per SF | Kruskal–Wallis | H = 5.38 | 3 | 0.1462 | p < 0.0063 | Unclear | ε² = 0.244 | large |
| Living area | Kruskal–Wallis | H = 16.79 | 3 | 0.0008 | p < 0.0063 | Yes | ε² = 0.763 | large |
| Year built | Kruskal–Wallis | H = 2.66 | 3 | 0.4476 | p < 0.0063 | Unclear | ε² = 0.121 | medium |
| Days on market | Kruskal–Wallis | H = 6.24 | 3 | 0.1004 | p < 0.0063 | Unclear | ε² = 0.284 | large |
| Discount from original list | Kruskal–Wallis | H = 8.64 | 3 | 0.0344 | p < 0.0063 | Unclear | ε² = 0.393 | large |
| Price cut before sale | Chi-square | χ² = 7.05 | 3 | 0.0704 | p < 0.0063 | Unclear | Cramér’s V = 0.553 | large |
| Seller paid closing costs or repairs | Chi-square | χ² = 4.44 | 3 | 0.2178 | p < 0.0063 | Unclear | Cramér’s V = 0.439 | medium |
| Has a view | Chi-square | χ² = 3.21 | 3 | 0.3602 | p < 0.0063 | Unclear | Cramér’s V = 0.374 | medium |
8.3.2 Quartiles by living area
| Measure | Test | Statistic | df | p-value | Threshold | Real difference? | Effect size | Size |
|---|---|---|---|---|---|---|---|---|
| Net sale price | Kruskal–Wallis | H = 13.13 | 3 | 0.0044 | p < 0.0063 | Yes | ε² = 0.597 | large |
| Net price per SF | Kruskal–Wallis | H = 2.79 | 3 | 0.4246 | p < 0.0063 | Unclear | ε² = 0.127 | medium |
| Year built | Kruskal–Wallis | H = 1.91 | 3 | 0.5914 | p < 0.0063 | Unclear | ε² = 0.087 | medium |
| Days on market | Kruskal–Wallis | H = 1.07 | 3 | 0.7851 | p < 0.0063 | Unclear | ε² = 0.048 | small |
| Discount from original list | Kruskal–Wallis | H = 2.02 | 3 | 0.5673 | p < 0.0063 | Unclear | ε² = 0.092 | medium |
| Price cut before sale | Chi-square | χ² = 3.27 | 3 | 0.3522 | p < 0.0063 | Unclear | Cramér’s V = 0.377 | medium |
| Seller paid closing costs or repairs | Chi-square | χ² = 4.44 | 3 | 0.2178 | p < 0.0063 | Unclear | Cramér’s V = 0.439 | medium |
| Has a view | Chi-square | χ² = 3.21 | 3 | 0.3602 | p < 0.0063 | Unclear | Cramér’s V = 0.374 | medium |
8.3.3 Quartiles by year built
Too few sales in a group to test.
8.3.4 View and no view
| Measure | Test | Statistic | df | p-value | Threshold | Real difference? | Effect size | Size |
|---|---|---|---|---|---|---|---|---|
| Net sale price | Mann–Whitney U | U = 67.00 | — | 0.1109 | p < 0.0063 | Unclear | rank-biserial r = 0.489 | medium |
| Net price per SF | Mann–Whitney U | U = 58.00 | — | 0.3630 | p < 0.0063 | Unclear | rank-biserial r = 0.289 | small |
| Living area | Mann–Whitney U | U = 65.00 | — | 0.1504 | p < 0.0063 | Unclear | rank-biserial r = 0.444 | medium |
| Year built | Mann–Whitney U | U = 54.00 | — | 0.5207 | p < 0.0063 | Unclear | rank-biserial r = 0.200 | small |
| Days on market | Mann–Whitney U | U = 23.50 | — | 0.1173 | p < 0.0063 | Unclear | rank-biserial r = 0.478 | medium |
| Discount from original list | Mann–Whitney U | U = 23.00 | — | 0.1109 | p < 0.0063 | Unclear | rank-biserial r = 0.489 | medium |
| Price cut before sale | Chi-square | χ² = 1.17 | 1 | 0.2798 | p < 0.0063 | Unclear | Cramér’s V = 0.225 | small |
| Seller paid closing costs or repairs | Chi-square | χ² = 1.35 | 1 | 0.2461 | p < 0.0063 | Unclear | Cramér’s V = 0.242 | small |
8.4 Data notes
Left out of the analysis: 8 new-construction sales (basic statistics only, section 7); 1 sale(s) with concessions above 10% of price. Statistics: Python, NumPy 2.5.3, SciPy 1.18.1.
8.5 References
- Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.). Erlbaum.
- Cramér, H. (1946). Mathematical Methods of Statistics. Princeton University Press.
- Dunn, O. J. (1961). Multiple comparisons among means. JASA, 56(293), 52–64.
- Kerby, D. S. (2014). The simple difference formula: an approach to teaching nonparametric correlation. Comprehensive Psychology, 3, 11.IT.3.1.
- Kruskal, W. H., & Wallis, W. A. (1952). Use of ranks in one-criterion variance analysis. JASA, 47(260), 583–621.
- Mann, H. B., & Whitney, D. R. (1947). On a test of whether one of two random variables is stochastically larger than the other. Annals of Mathematical Statistics, 18(1), 50–60.
- Pearson, K. (1900). On the criterion that a given system of deviations… Philosophical Magazine, 50(302), 157–175.
- Tomczak, M., & Tomczak, E. (2014). The need to report effect size estimates revisited. Trends in Sport Sciences, 21(1), 19–25.
This is a market study of closed sales, not an appraisal or an opinion of value for any home. The analysis covers 23 closed MLS resale sales; the 8 new-construction sales appear only in their own section, as basic statistics. Sale and listing data: ABOR MLS via secure APIs, pulled October 26, 2026. Map data © OpenStreetMap contributors. ALEX Companies · Licensed by the Texas Real Estate Commission · Broker: Alexander Tiffany Southwest, LLC (ALEX Real Estate), #565194-BB · Diane Hart Alexander, Designated Broker, #574877-B.