Lunenburg by the numbers

Each of those is an estimate from a sample, not a count, and the number after the ± is how far out it may be. Two figures whose ranges overlap are not different from each other. That matters here more than it would in a city: the American Community Survey reaches a small number of households in a town this size, so 65 and over is 1,979 ± 309 — somewhere between 1,669.614 and 2,288.386 people. The 1,568 children in the schools, by contrast, is a count: the district counted them.

What this page establishes

Those are the answers this report could reach from the documents behind it. If the one you came for is not among them, ask us — no name or email needed.

The town, by age

Five age groups, each with its own margin. 65 and over and under 18 are the two the budget argument is about, and they are closer in size than either side usually says: 1,979 ± 309 against 2,736 ± 429.

Census table B01001, 2019–2023, for Lunenburg town. The whiskers are each group’s own margin of error; the label above each bar is its share of the town. A group is a sum of twelve or more published cells, so its margin is the root of the sum of theirs — wider than any one of them.

Households, and who is in them

This is the denominator under the thing people say at meetings — that the schools are for about a third of homes. On this measure the claim is fair. But it counts a child under 18, which includes children at Monty Tech, at private and charter schools, and children not yet old enough for school, so it is a ceiling on homes with a child in the Lunenburg public schools rather than that figure. Nothing published joins a child to a household to a school; that limit is a registered gap, not an oversight of this page.

What households earn

Median household income overall, and by the age of the person who heads the household. The last of those is the number the “fixed incomes” argument has never had: a household headed by someone 65 or over has a median income of $72,935 ± $8,677, which is 56.5% of the $129,183 a household headed by someone 45 to 64 has — a gap of $56,248 a year.

Census table B19049, 2019–2023. Four separate medians of four separate sets of households, measured in the same five years — not one household’s income changing as it ages, and not a series through time. The under 25 band is not drawn because the Census published no estimate for it.

Where Lunenburg sits in Massachusetts

Lunenburg’s median household income is $109,753 ± $12,275. Ranked against the other 350 municipalities in Massachusetts that puts it 167 of 350 — and 205 of those 350 have an interval that overlaps this town’s, spanning everything from 37th to 336th place. The rank is real arithmetic on real figures and it is not a ranking anybody should carry to a meeting. The honest sentence is that Lunenburg is an ordinary Massachusetts town on income, and no amount of sorting will make it otherwise.

Census table B19013, 2019–2023, every Massachusetts municipality. The statewide median municipality is $107,705. 4 municipalities publish no usable margin — the Census top-codes a median above its highest bracket — so they are counted in the distribution and in neither of the two counts above; every one of them is far above Lunenburg. On the Census’s own significance test, which is stricter than an overlap of intervals, 152 of 350 municipalities still cannot be told apart from this town.

And the same town, ranked on what it spends for each pupil

Both of these are true at once. On income Lunenburg is in the middle of Massachusetts and indistinguishable from most of it. On spending for each pupil it is 310 of 318 districts in FY2025, and in the bottom quarter in 17 of 17 published years. The two ranks are not the same kind of rank — one is municipalities on a sample estimate with a margin, the other is school districts on a figure the state computes from returns, including charter and virtual districts — so they sit side by side and are never differenced. The per-pupil comparison in full.

Did anything change between 2014–2018 and 2019–2023?

Mostly, nothing that can be shown. 15 measures were compared across the two releases and 3 of 15 clear their own margins: Population; Households; Median income, householder 65 and over. The other 12 moved by less than the combined margin of their own two estimates, which is a difference this survey cannot see. That includes the share of the town 65 or over and the share of households with a child — the two figures an argument about a changing town would most want.

2014–2018 against 2019–2023: two five-year windows that do not overlap, which is what makes the comparison permissible at all. Consecutive releases share four years of sample and may not be differenced. Every difference is drawn in multiples of its own combined margin, because the15 measures are counted in people, households, dollars and percentage points, and one axis cannot carry four units. Population clears by far the largest multiple because the survey measures the town total to ±19 where a single age group inside it is measured to ±309.

The record, and what it does not carry

What the town has said about who lives here

Searched 8,899 of the 12,015 meeting documents this archive holds (74%), covering 2009-01-05 to 2026-11-17. The other 3,116 carry no text a search can match — 3,096 of them are image scans awaiting OCR. An empty result above is a statement about what can be read, never about what was said.

The codes that are not numbers

The Census publishes sentinels where it has no figure, and each is a large negative number that would read as data to anything that did not know better. They are kept verbatim in the archive and never rendered as figures here: 2 cells in this page’s tables carry one.

What this page cannot answer

• How many Lunenburg homes have a child in the LUNENBURG PUBLIC SCHOOLS — as opposed to a child under 18?The Census answers the second and not the first. B11005 puts 1,477 of 4,529 households — 32.6% — with a child under 18, which is what people mean when they say "about 30% of taxpaying homes", and on that reading the claim is fair. But that count includes children at Monty Tech, at private and charter schools, and children not yet of school age, so it is a CEILING rather than the figure. The arithmetic suggests the gap is real: 1,477 households against roughly 1,745 school-age children is about 1.2 per household, low enough that a meaningful share of those homes have only pre-school children. Nothing published joins a child to a household to a school. — closes: a count of DISTINCT HOUSEHOLDS with at least one child enrolled, which the district can produce from its own student information system and DESE cannot, because DESE counts children and the Census counts households and neither knows about the other.
• Who actually lives in Lunenburg — how many are over 65, how many households have a child in the schools, and what each group can afford?ANSWERED IN PART, 11 September 2026, and the part that remains is the part the argument actually runs on. The Census tables named below were fetched and are loaded as `census_acs`: 11,804 ± 19 residents, 1,979 ± 309 of them 65 or over, 4,529 ± 235 households, 1,477 ± 198 with a child under 18, median household income $109,753 ± $12,275 and $72,935 ± $8,677 where the householder is 65 or over. See /lunenburg-by-the-numbers, where every figure carries its margin. WHAT SURVIVES is the third clause — what each group can AFFORD. Income is not wealth, not savings, not equity and not a tax bill, and nothing published joins a property tax payment to the age of the person who paid it; the two rows below carry the two halves of what is still missing. — closes: the town census, which counts every resident by name, age and household rather than sampling them, held by the Town Clerk.
• How much of the property tax levy is paid by households headed by someone 65 or over?The Census now gives income by age of householder — a household headed by someone 65 or over has a median income of $72,935 ± $8,677, against $129,183 ± $21,150 for a 45-to-64 household — and that is income, not tax. The override argument is about a BILL: what a $3,000,000 override adds to a household is quoted at meetings as a town-wide average, and the town publishes no split of the levy by who pays it, by age or by anything else. Nor does it publish how many households take the statutory elderly exemptions and abatements that would change the answer. So "seniors on fixed incomes cannot carry it" cannot be checked in either direction, and neither can its opposite. — closes: the Board of Assessors’ commitment list with the Clause 41A, 41C and 17D exemptions granted, joined to the town census by age of owner — both held by the Town and neither published.

What this does not establish

• That any of these figures is a count. Every one is a five-year SAMPLE estimate with a margin of error, and two figures whose intervals overlap are not different. The town’s own population total is measured to ±19; a single age band inside it is not.
• That 1,477 ± 198 households have a child in the LUNENBURG PUBLIC SCHOOLS. The Census counts a child under 18, which includes Monty Tech, private and charter children and children not yet of school age, so the figure is a ceiling.
• That Lunenburg is richer or poorer than a town it out-ranks. 205 of the 350 municipalities have an income interval that overlaps this town’s, so a rank of 167 of 350 is arithmetic rather than a finding.
• That the income rank and the per-pupil rank measure the same population. One is 350 municipalities on a sample estimate; the other is 318 school districts on a figure DESE computes from returns, and it includes charter and virtual districts that are not municipal.
• Anything about what a household can AFFORD. Income is not wealth, not savings, not equity in a house and not a tax bill, and nothing in this archive joins a property tax payment to the age of the person who paid it.
• Anything in real terms. This archive holds no price index, so no figure here is deflated and no change described is a change in purchasing power.
• Anything at all from 2 published sentinel(s) — the Census’s codes for "no estimate" and "no computable margin". They are not zeros and are not printed as figures.
• Why any of this is so. Every figure here is a measurement of who lives in this town; a reason for it would be a hypothesis, and this page states none.

Where every figure came from

The Census answers a query with JSON, and an API answer is a document like any other: each raw response is saved, hashed and published, and the figures on this page are read from those files rather than from a query nobody kept.

Built by scripts/build_lunenburg_by_the_numbers.py, and recomputed from the CSVs by a second route in scripts/verify_lunenburg_by_the_numbers.py.