Who works for the town

How many people each part of the town employs, which are growing and which have shed staff. The seats people volunteer for are board composition.

How many people each part of the town employs

Department of Public Works and Board of Assessors report an ESTABLISHMENT rather than a headcount — the posts the department states it has, not the people standing in them. A post can sit vacant and still be printed here, and the Assessing office says so itself: in FY2024 it reported being “fully staffed for the first time in over a year”.

These are not identical measures — a named roster, a career count plus the low end of an on-call range, and an establishment of posts — and they are answers to two different questions: how many people are there, and how many posts the department says it has. Where a department states a range the LOW end is used, so none is flattered by its own vagueness.

change is measured over each department’s own record, and those records do not line up: the Schools record spans 14 years and the Board of Assessors’s 2, so +32 and +1 are not the same claim. No year is published by all 9 — the fullest is FY2019, with 6 of them — and standardising on that would drop the three that STOPPED publishing, which is the part worth seeing. a year is the change divided by the years it spans, and that column compares.

A year whose count reads under half the year before it is dropped as a short read rather than published as a cut: Police Department FY2014. The Police roster comes back as seven officers in FY2024 against twenty-five in FY2023, which is a page this reader did not find, not three quarters of a police force.

Headcount over time

On one scale the relation is plain, and it is the one the panels below deliberately hide: every other department that publishes a headcount fits inside the schools several times over. The panels answer how each one is MOVING; this answers how big each one is.

Schools moves between 170 and 272 across 15 published years; Fire Department moves between 36 and 48 across 15 published years; Police Department moves between 13 and 33 across 13 published years. Library and Board of Assessors publish too few years to show a trend at all.

The schools, in detail

Six times the next employer, so worth breaking out. FY2025.

The school columns come to 244 and the district employs 232. That is not an error to tidy away: somebody who teaches at two schools is staff at both and one employee of the district, so the two answer different questions and are counted differently.

Which parts of the budget have a headcount behind them

Town Meeting votes the budget in 12 groups. 6 of them now have at least one department publishing how many people it employs; 6 have none.

A group with a count is not a group that is counted: General Government is voted as one figure covering half a dozen offices, and two of them state a number.

The ones with no count, and why

Three different answers, and only one of them is a department being quiet.

The gross-wages list named a department beside each employee through FY2016 and stopped, so there has been no town-wide headcount by department since.

How long people stay, and how many leave

A person’s tenure here is the number of years their name appears under a department in the years that department published anything — not calendar years, because 27 of 86 bodies skip at least one year inside their own span and a naive count reads a missed report as the whole staff leaving and coming back.

Half of these people are CENSORED and their tenure is a lower bound, not a length. 579 of 1290 are present in their department’s first published year — so they started before the archive begins and nothing here can say when — or in its last, so they are still there as far as anything here can see. That is why two medians are given: the first counts everybody, the second only the people whose arrival AND departure both fall inside the record, which is the only one that means what it says.

A name leaving a roster is a name leaving a roster. It is not a resignation, a retirement or a cut post. Somebody may have moved between departments, been left off a page, or held a post the town stopped printing. The last column is the rate at which names stop appearing, and it is not a separation rate.

A roster is also a point in time and undated within its year, so N appearances is at most N years of service and at least N−1.

What this cannot show

• What anybody is paid. No salary is read here, and none is inferred from a post’s name.
• How many people the town employs. Nine departments state a figure, in eight different forms, and no year has all of them — the wage list that would give one town-wide headcount stopped naming departments after FY2016.
• FTE. A career post and an on-call post are not the same job and cannot be netted against each other.
• Whether a post was CUT. A roster is people IN post and a cut removes a POST, and those come apart both ways: someone retires and the post sits vacant but funded, so the count falls and nothing was cut; or a post is eliminated and its holder moves to another vacancy, so the count holds and something was. The town’s own words for it are in the Assessing office’s FY2024 report — “we are fully staffed for the first time in over a year” — a year of posts that existed and were empty. A rise or a fall here is a change in PEOPLE PRESENT, never a decision about establishment.
• When in the year anybody was counted. A roster is a point in time and is undated within its year, so a September departure and a June one are the same figure.
• A town total. The departments that describe their staffing do it in whichever form that year’s department head chose — a count, a range, an establishment post by post, a list of names — and those are different quantities.

Where it comes from

• Every appointed post and its holder, FY2016 to FY2025 — the APPOINTED OFFICIALS listing in each annual town report. Read by scripts/extract_personnel.py. A post that states no membership and no term is an officer rather than a board seat — the town’s own distinction, not ours.
• What each department says it employs — the prose of each department’s own report; no heading names it in any year. Read by scripts/extractdepartmentstaffing.py. Stored as printed, because the forms do not agree with each other and may not be summed. Eight forms so far, and the department head chooses: a list of names, an establishment post by post, the same establishment written one article at a time, a career count beside an on-call range, a number spelled out in words, and — the Library’s only one — a headcount in an aside inside a sentence about turnover.
• The Police and Fire rosters, by name — the Police Department Personnel and Fire Roster of the Lunenburg Fire Department pages, both set in two columns. Read by scripts/extractdepartmentrosters.py, from the WORD geometry because a line box spans both columns.