This page exists so that any figure we publish can be checked, challenged, or reproduced. It is published before the measures themselves, so the rules can be argued with first rather than written to fit a result.
Every figure on a Market Data page is one of three things, and each is labelled on the page where it appears.
No number on this site is generated, inferred, rounded into existence, or filled in by a model. A measure is computed from records by one deterministic program, or it is cited to a named source, or the page says it is unavailable. There is no fourth case.
Closed-sale records come from the Garden State Multiple Listing Service and the New Jersey Multiple Listing Service. We publish aggregates only: medians, counts, ratios and distributions over a stated set of sales. We do not publish individual addresses, listing detail, photographs, or any searchable or interactive record of a property, and raw records are never committed to our systems of record. Where a sale is named anywhere on this site it is because it is a matter of public record and relevant to a decision, not because it came out of a data file.
Tax and assessment figures come from the New Jersey Division of Taxation and the relevant county board of taxation. School information links to the New Jersey Department of Education's own reports; we do not publish school ratings or scores of our own and do not rank districts. Mortgage rates come from the Freddie Mac Primary Mortgage Market Survey, quoted with its survey week.
Licensed vendor data informs our reading of the market. It is not republished here.
A figure that cannot state all five of these does not get published:
A small sample does not produce a cautious number. It produces no number. Where a measure's sample falls below the threshold for that measure, the page states that the measure is unavailable for that period and geography, and says why, instead of printing a figure with a hedge attached. A median over four sales is not a median of anything; publishing it with a footnote invites exactly the use the footnote warns against.
Suppression is applied by the program that computes the measures, from a stated threshold, not by editorial judgement after the fact. A suppressed measure is visible as suppressed. It is never silently omitted, because an absent row and a thin row look identical to a reader and only one of them is honest.
Measures are computed for a closed period and left alone. When later data revises a figure, we do not quietly rewrite the published page: the correction is recorded in the corrections log with what changed, when we learned of it, and what the figure now is. An article keeps its original text so that the record of what we said, and when, stays intact.
Comparisons across unmatched periods are labelled as such. A full year against a nine-month window is not a like-for-like comparison, and a change in a median reflects both price movement and the mix of homes that sold, not per-home appreciation.
A mortgage rate appears here as what it does to a monthly payment at a stated price, with the survey week named. It does not appear as a running number, because a rate with no price attached to it does not inform a decision.
We publish effective tax rates to compare one town with another. We do not compute anyone's tax bill, and a figure here should not be used as one: an individual assessment, exemptions and appeals all sit between a town rate and a bill. Assessment practice varies by municipality and changes on its own schedule, independent of any particular sale.
Town and neighbourhood content here describes housing stock, closed-sale behaviour, commuting infrastructure, tax mechanics and official school sources. It does not characterise who lives anywhere, and it does not rank places as better or worse. Which town suits a particular household is that household's judgement, and the useful service is showing the record clearly rather than scoring it.
Each Market Data page carries a one-line "Cite as" with the measure, its period, its geography and its source, so a figure can be quoted accurately without reading the whole page. Please quote the period and the sample size along with the number; a median without its period and its n is not a fact about anything.
That line states how to attribute a figure, and grants nothing. It is not a licence to republish the underlying closed-sale records, and we do not offer one: those records are licensed to us under multiple listing service terms, and we have no standing to pass that permission on. If you want to use more than a cited figure, ask us and we will tell you what we can and cannot authorise.
If a figure here looks wrong, tell us and we will check it against the records and publish the result either way. A disagreement we cannot resolve gets stated as a disagreement. Send a correction or a question.