How the index collects, labels and dates its numbers

Every number on this site follows the same rules. Each one states its source, its edition, the date we checked it, its unit, whether it is weighted, and whether it was measured or modelled. Every page carries its published date and its modified date from its record, never from the build.

This page sets out those rules once. Which source we use for each kind of claim is a separate question, and the sourcing policy answers it. See which source we use for each kind of claim.

Weighted and unweighted figures are labelled, never blended

The Cross-Canada Survey of Radon Exposure in the Residential Buildings of Urban and Rural Communities (the 2024 Cross-Canada Radon Survey) weights its results for some provinces and reports others unweighted. We never blend the two.

• A weighted result is shown as the survey reports it.
• An unweighted result is labelled unweighted.
• Nunavut has no survey data and shows as “no data”, never as zero.

Survey tables here reproduce the published figures exactly, with credit and page references. Any column we add is marked ours. The survey asks that its figures and tables not be altered; we keep to that. We cite its current edition, Version 1.1 (CCSR.24.1.1), October 2024, ISSN 2818-8195.

Its headline figure reads: “approximately 17.8% of residential properties in Canada have radon levels at or above 200 Bq/m³”. That describes properties of three types (single detached, semi detached, row), never people. Any companion figure from the survey is quoted as the survey phrases it. See the survey tables with their weighting labels and page references.

Test outcomes and surveyed homes are different units, and the page says which one it shows

Every display says which unit it shows. The 2024 survey report counts 69,478 unique long-term test outcomes. The radon dashboard reports “over 35,791 surveyed homes”. We never merge them; where one headline number is needed we use the report’s test outcomes.

A dashboard figure carries the dashboard’s own note: “Results displayed below are simulated to assume all research participants tested their homes on the same floor.” We name which of its four outcome states produced it. When credible sources disagree, we follow four rules:

• Name each source and give its date or edition.
• Explain the difference where we can.
• Use the newest dated primary edition and note the difference.
• Never blend incompatible figures or pick the most alarming one.

Health Canada’s provinces and territories action guide cites the survey’s “about 1 in 5 homes”. Its municipalities action guide still says “about 7%”. Both pages are live, so every prevalence claim here names its page and date. See what the 2024 figures mean and why they differ from 2012.

Every number carries its source and the date it was checked

Each statistic record carries its source URL and date checked; one past its review date is listed for re-checking. Each dataset display states all of these.

• Year and publication version.
• Methodology and sample size.
• Geographic level.
• Weighting status.
• Testing location assumption.
• Measured or modelled.
• Date last verified.

Radon is reported in becquerels per cubic metre (Bq/m³). Where a source uses picocuries per litre we show both, at 37 Bq/m³ per pCi/L. A dataset’s HTML table, CSV and JSON come from one source file. The data sits in the HTML, not a script.

Our original datasets and studies carry a CC BY 4.0 licence. A dataset derived from third-party data carries no licence statement until its terms are cleared. The coverage, price observations, response times and service terms datasets are our own original data. Each method is published with its data on that dataset’s page.

• See every dataset with downloads and column notes
• See what the published, modified and review dates mean
• See the citation form for a dataset, edition or fact
• See what you may do with the data

Results are grouped by census division, the unit the surveys use

Place results are grouped by census division, keyed to Statistics Canada 2021 census division codes. Maps use the matching Statistics Canada boundary file; the legend carries “No data” and “Insufficient data”.

A provider record is one business, one page, national, name variants merged. The datasets hold every certified organization with its type. Directory, provider pages and place counts cover only those offering services to the public. Provider and credential records come from a dated pull of the C-NRPP registry:

• Every count carries the pull date, and every listing its last-checked date.
• A lapsed credential shows as lapsed as of its date; the provider leaves the directory and counts but stays in the datasets with its status.
• Status comes from the credential fields, never by hand.

Our tools help you understand radon information; neither they nor a map can tell you the level in your home. We make no medical claim and never turn a number into an instruction. See the provider list with its dated registry check.