TL;DR

Coverage: 2026 assessments exist for 100,270 parcels (0 missing). Median increase vs the 1968 base is 16.99×; 99%+ exceed 3× (expected due to the 1968 baseline, not a fairness signal by itself).
Fairness vs recent sales (20,460 arms‑length sales since 2020):
- Median assessed-to-sale ratio: 1.076 (about +7.6%).
- Only 14.3% fall within ±5% of sale price; 25.8% within ±10%; 41.8% within ±20%.
- 40.5% are >20% above sale; 16.4% >100% above sale; 17.8% >20% below.
Municipal drift (median assessed-to-sale ratio):
- Higher medians include: Carbondale Twp (1.41), Carbondale City (1.18), Throop (1.16), Scranton (1.12).
- Dispersion is large in several places (p90s >3–4×; some extreme outliers).
Per‑sqft lens (homes with reasonable area): County median $110/sf; p90$ 175/sf; p99 $304/sf. ~**816** properties (\approx1.0%) exceed$ 300/sf.
Land share shift: Median land share rises from 20.0% to 22.3%. About 17.0% of parcels are all‑land (legitimate for vacant land, but still worth QA against structure presence).
Bottom line: These distributions are far broader than a “5% margin of error” narrative. On recent sales, fewer than 26% are within ±10%. A structured QA and a non‑uniformity‑aware appeal pathway are warranted before tax bills are set.

Explore the data yourself (Early Beta)

Our Parcel Explorer lets you browse parcels and ask questions in natural language. It's experimental and may have rough edges — please kick the tires and tell us what breaks.

Why this matters

Commissioners have suggested the process has “no issues” and that low appeal counts imply success. But taxpayers still don’t know 2026 tax bills (millage not set), making widespread appeals unlikely at this stage. Our countywide measurements show substantial dispersion relative to market proxies (recent arms‑length sales), which is the core of the constitutional uniformity standard.

Data and method

Data provenance & accuracy: We did our best to scrape official county sources, but scraping can introduce issues. We believe the compiled data are very accurate overall; please treat as best‑effort and report any problems so we can correct them.
Data provenance & accuracy: We did our best to scrape official county sources, but scraping can introduce issues. We believe the compiled data are very accurate overall; please treat as best‑effort and report any problems so we can correct them.
Data provenance & accuracy: We did our best to scrape official county sources, but scraping can introduce issues. We believe the compiled data are very accurate overall; please treat as best‑effort and report any problems so we can correct them.
Primary table: ultimate_property_data (100,270 records) with 2026 assessments, 1968 baselines, building attributes, and recent sale info. We scraped the Lackawanna County assessor ourselves and, for analysis/QA only, enriched certain fields with third‑party parcel/assessor datasets to check quality and detect anomalies.
Sales frame: Arms‑length sales since 2020 where most_recent_sale_price ∈ [ $10k,$ 5M].
Metrics computed directly in SQL (read‑only) against Supabase (vcxwapqhwjjtzsejcumb).
This is a countywide view; individual properties may reasonably vary from medians. Uniformity is about the distribution.

Parcel counts and coverage note

Our consolidated warehouse contains two relevant tables: ultimate_property_data = 100,270 parcels; parcels_analysis_ready = 100,490 rows. The difference reflects consolidation/quality‑control during the “ultimate” build.
The County has not published a single authoritative parcel count. We have heard differing totals in Tyler presentations and from Commissioners. Expect small discrepancies across sources/timestamps. All metrics in this article use ultimate_property_data (100,270).

Key findings

1) Increases vs 1968 baseline (context, not fairness)

Median assessment_multiplier (2026 ÷ 1968): 16.99×; p10 9.28×, p90 34.17×, p99 160.09×.
3× counts exceed 99% of parcels. This does not prove unfairness; it reflects the half‑century‑old base year.

2) Fairness vs recent arms‑length sales (the uniformity signal)

Sample size: 20,460 parcels (2020+ arms‑length sales).
Median assessed-to-sale ratio: 1.076.
Band shares: 14.3% within ±5%; 25.8% within ±10%; 41.8% within ±20%.
Tails: 40.5% >120% of sale; 27.0% >150%; 16.4% >200%. 17.8% are <80%.

Interpretation: In a “roughly uniform” system, we’d expect a far higher share clustering around parity for recent sales. Instead, nearly three in five recent sellers fall outside ±20%—evidence of dispersion and localized bias.

3) Municipal drift (median assessed-to-sale ratios)

Medians vary materially by municipality (n≥100 recent sales):

Municipality	Median ratio	>120% share	<80% share
640 – Carbondale Twp	1.407	56.7%	18.9%
680 – Carbondale City	1.176	49.5%	20.2%
380 – Throop	1.155	46.9%	11.3%
340 – Scranton	1.115	44.9%	17.3%
300 – Olyphant	1.096	42.9%	16.2%
250 – Moosic	1.087	38.4%	14.4%
290 – Old Forge	1.083	40.4%	14.1%
750 – Jefferson Twp	1.081	43.2%	20.9%
270 – Newton Twp	1.069	41.8%	16.9%
020 – Archbald	1.068	36.6%	14.2%

Note: Several places also show very high p90 ratios (3–4×+), indicating heavy outliers.

All municipalities: assessed-to-sale breakdown (2020+ arms‑length sales)

Municipality	Sales n	Median ratio	p10	p90	>120%	<80%
010 – Waverly Twp	166	1.013	0.481	1.536	23.5%	19.9%
020 – Archbald	745	1.068	0.730	2.823	36.6%	14.2%
030 – Benton Twp	196	1.060	0.600	3.165	35.7%	19.9%
040 – Blakely	505	1.047	0.613	2.623	37.2%	17.0%
070 – Clarks Green	140	0.974	0.745	1.524	24.3%	18.6%
080 – Clarks Summit	502	1.059	0.702	1.734	32.5%	14.3%
110 – Dalton	143	1.043	0.379	2.991	39.2%	23.1%
120 – Dickson City	520	1.039	0.435	2.119	36.2%	19.4%
130 – Dunmore	997	1.037	0.681	2.064	35.9%	16.0%
160 – Glenburn Twp	149	0.985	0.483	1.963	27.5%	26.8%
200 – Jessup	417	1.067	0.602	2.308	36.0%	20.1%
210 – LaPlume Twp	23	1.021	0.710	2.928	39.1%	17.4%
250 – Moosic	508	1.087	0.664	2.603	38.4%	14.4%
270 – Newton Twp	249	1.069	0.596	3.835	41.8%	16.9%
280 – North Abington Twp	68	1.088	0.689	4.459	39.7%	13.2%
290 – Old Forge	703	1.083	0.727	2.703	40.4%	14.1%
300 – Olyphant	452	1.096	0.621	3.118	42.9%	16.2%
310 – Ransom Twp	116	0.977	0.145	2.601	33.6%	33.6%
340 – Scranton City	6,592	1.115	0.603	2.948	44.9%	17.3%
350 – South Abington Twp	805	1.041	0.737	1.938	30.6%	13.3%
370 – Taylor	447	1.031	0.627	1.972	33.3%	16.1%
380 – Throop	390	1.155	0.769	4.340	46.9%	11.3%
400 – West Abington Twp	27	1.449	0.967	3.827	63.0%	0.0%
500 – Clifton Twp	512	1.041	0.608	2.533	38.9%	20.9%
510 – Vandling	98	1.090	0.293	2.076	39.8%	26.5%
520 – Madison Twp	176	1.154	0.535	3.268	42.6%	17.6%
530 – Elmhurst Twp	53	1.066	0.551	3.933	39.6%	20.8%
540 – Roaring Brook Twp	247	1.106	0.710	12.182	45.3%	14.2%
550 – Thornhurst Twp	212	1.171	0.567	3.028	49.1%	16.5%
570 – Spring Brook Twp	231	1.043	0.198	3.235	35.1%	26.4%
580 – Greenfield Twp	202	1.063	0.390	3.446	39.1%	22.8%
590 – Moscow	145	1.020	0.498	2.051	33.8%	18.6%
600 – Jermyn	170	1.050	0.665	2.445	36.5%	15.3%
640 – Carbondale Twp	127	1.407	0.546	4.093	56.7%	18.9%
680 – Carbondale City	944	1.176	0.339	3.434	49.5%	20.2%
720 – Covington Twp	1,066	1.055	0.688	1.774	35.7%	16.6%
730 – Scott Twp	456	1.056	0.280	3.175	38.6%	26.3%
740 – Fell Twp	305	1.017	0.233	2.877	39.3%	29.5%
750 – Jefferson Twp	472	1.081	0.446	4.803	43.2%	20.9%
900 – Mayfield	184	1.024	0.342	2.991	38.6%	22.8%

4) Per‑square‑foot lens (homes with reasonable area)

County median: $109.6/sf; p90$ 175.2/sf; p99 $303.8/sf.
Outliers: ~816 properties (≈1.03%) exceed $300/sf.
Highest municipal medians by $/sf (n\geq100): Waverly Twp$ 158/sf, South Abington $152/sf, Clifton$ 147/sf, Clarks Summit $139/sf, North Abington$ 139/sf, Moscow $134/sf.

This lens helps normalize differences in house size; the long right tail deserves QA review.

All municipalities: assessed $/sf medians

Municipality	Homes n (area filter)	Median $/sf
010 – Waverly Twp	731	$157.77
020 – Archbald	3,110	$118.92
030 – Benton Twp	960	$108.28
040 – Blakely	2,521	$111.85
070 – Clarks Green	631	$126.37
080 – Clarks Summit	2,053	$139.14
110 – Dalton	548	$125.31
120 – Dickson City	2,442	$109.77
130 – Dunmore	4,823	$119.16
160 – Glenburn Twp	605	$124.33
200 – Jessup	1,853	$102.85
210 – LaPlume Twp	148	$99.15
250 – Moosic	2,217	$117.89
270 – Newton Twp	1,106	$118.13
280 – North Abington Twp	289	$138.54
290 – Old Forge	3,386	$104.54
300 – Olyphant	2,044	$115.36
310 – Ransom Twp	598	$84.98
340 – Scranton City	22,293	$104.65
350 – South Abington Twp	3,225	$151.74
370 – Taylor	1,926	$97.06
380 – Throop	1,709	$113.45
400 – West Abington Twp	137	$127.90
500 – Clifton Twp	1,236	$146.68
510 – Vandling	322	$82.64
520 – Madison Twp	942	$103.27
530 – Elmhurst Twp	317	$103.79
540 – Roaring Brook Twp	873	$122.43
550 – Thornhurst Twp	630	$127.29
570 – Spring Brook Twp	1,077	$100.54
580 – Greenfield Twp	1,067	$114.19
590 – Moscow	750	$133.99
600 – Jermyn	815	$92.43
640 – Carbondale Twp	505	$90.22
680 – Carbondale City	3,199	$92.72
720 – Covington Twp	2,116	$48.06
730 – Scott Twp	2,254	$103.50
740 – Fell Twp	1,080	$85.99
750 – Jefferson Twp	1,704	$114.91
900 – Mayfield	766	$100.61

Parcel counts by municipality (warehouse coverage)

Municipality	Total parcels	With 2026 value	Sales sample n
010 – Waverly Twp	858	858	166
020 – Archbald	3,832	3,832	745
030 – Benton Twp	1,293	1,293	196
040 – Blakely	2,995	2,995	505
070 – Clarks Green	660	660	140
080 – Clarks Summit	2,291	2,291	502
110 – Dalton	696	696	143
120 – Dickson City	2,834	2,834	520
130 – Dunmore	5,448	5,448	997
160 – Glenburn Twp	903	903	149
200 – Jessup	2,165	2,165	417
210 – LaPlume Twp	214	214	23
250 – Moosic	2,710	2,710	508
270 – Newton Twp	1,435	1,435	249
280 – North Abington Twp	594	594	68
290 – Old Forge	3,851	3,851	703
300 – Olyphant	2,403	2,403	452
310 – Ransom Twp	877	877	116
340 – Scranton City	26,751	26,751	6,592
350 – South Abington Twp	3,833	3,833	805
370 – Taylor	2,254	2,254	447
380 – Throop	2,006	2,006	390
400 – West Abington Twp	189	189	27
500 – Clifton Twp	1,831	1,831	512
510 – Vandling	429	429	98
520 – Madison Twp	1,247	1,247	176
530 – Elmhurst Twp	429	429	53
540 – Roaring Brook Twp	1,247	1,247	247
550 – Thornhurst Twp	1,070	1,070	212
570 – Spring Brook Twp	1,607	1,607	231
580 – Greenfield Twp	1,611	1,611	202
590 – Moscow	898	898	145
600 – Jermyn	955	955	170
640 – Carbondale Twp	881	881	127
680 – Carbondale City	3,948	3,948	944
720 – Covington Twp	4,696	4,696	1,066
730 – Scott Twp	3,245	3,245	456
740 – Fell Twp	1,617	1,617	305
750 – Jefferson Twp	2,501	2,501	472
900 – Mayfield	966	966	184

5) Land share shift and all‑land valuations

Median land share rises from 20.0% (1968) to 22.3% (2026). p90 new land share hits 100%.
17.0% of parcels are all‑land (new land = new total). Many are legitimately vacant; however, a systematic cross‑check against structure presence (multiple sources) is a prudent QA step to prevent unintentional zeroing of improvement value.

Legal context: uniformity is the standard

In the 2023 Seminary Hill ruling, a Lackawanna three‑judge panel held that the taxpayers showed their properties were assessed at “significantly higher percentages of fair market value than comparable properties,” and ordered reductions to bring assessments in line with the common level (average ratio 71.05%) across comparable neighborhoods.

“Accordingly, the taxpayers are entitled to have their assessments reduced to comport with assessment‑to‑value ratios averaging 71.05% for the assessment periods from July 1, 2021, to the present and continuing forward.”

That decision confirmed three essentials:

Uniformity turns on ratios to market, not raw assessment levels.
Similar properties in like neighborhoods must be treated roughly alike.
When dispersion is pronounced, relief (and, if needed, countywide corrective action) is appropriate.

This matters for 2026 because it’s the same constitutional criterion—just on new values.

Addressing public comments and claims

“There are no issues; appeals are low.” Tax bills (millage) aren’t set, so low appeals do not prove fairness. Our sales‑based lens shows only 25.8% within ±10% of sale; 58.2% fall outside ±20%. That is not “no issues.”
“We have no choice—court order.” The 2018 stipulation sets an implementation date (Jan 1, 2026). Courts expect good‑faith compliance. If material QA issues remain, seeking appropriate relief (with a concrete QA plan) is legally cleaner than shipping known dispersion.
“People had plenty of time to appeal.” Without bills or uniformity‑aware guidance, most owners can’t benchmark. A transparent, uniformity‑oriented QA plus a robust appeal runway after bills are modeled would materially improve outcomes.

Methodological notes

Sales used: most recent arms‑length transactions since 2020 where consideration ∈ [ $10k,$ 5M].
$/sf computed on homes with building area between 600 and 20,000 sf.
Land share based on assessed land ÷ total for both 1968 and 2026 values.
All calculations were done directly in SQL against the consolidated warehouse and can be reproduced on request.