Methodology and data sources

Statistics tables

Most popular cars counts licensed cars in Great Britain in DfT table VEH0120 at the end of the latest quarter, grouped by DfT's generic model (so every Fiesta version counts as a Fiesta). Cars declared off the road (SORN) are left out. Rows where DfT records the model or make as missing are left out of the rankings but kept in the totals.

Best-selling new cars adds up the four quarters of the last full calendar year in DfT table VEH0160, cars registered for the first time in Great Britain.

Fuel types uses the same two tables, by DfT's fuel type. Engine sizes groups DfT's 100cc bands in table VEH0220 (UK, end of year) into wider bands; cars with no engine size recorded are fully electric.

Mileage by age is the median odometer reading at MOT tests in 2025 for class 4 vehicles (cars and light vans), by whole years since first use, from DVSA's anonymised MOT data. The median is the middle reading, so a few very high mileages do not pull it up.

Car colours come from the same 2025 MOT test results. We use class 4 tests only (cars, taxis and light vans up to 3,000kg) and count each vehicle once, taking its colour, make and first-use year from its first test of the year. Cars under three years old do not need an MOT, so the newest cars are under-represented, and the few vehicles with no colour stated are left out of the shares.

CO2 emissions come from DfT tables VEH0156 (the average reported CO2 of cars registered for the first time), VEH0256 (new cars by CO2 band) and VEH0206 (licensed cars by CO2 band), all for the United Kingdom. "Reported" CO2 is the figure from the official test when the car was new. The test changed from NEDC to WLTP between 2018 and 2020; WLTP gives higher figures for the same car, and DfT's table says which test each year uses. Shares by band leave out cars with no CO2 figure.

Most economical new cars uses official WLTP figures from the Vehicle Certification Agency (VCA), Crown copyright: the latest car fuel and emissions data, downloaded on 31 October 2026 with figures updated to 29 October 2026. Official figures are measured in a laboratory and are not real-world mpg. We show each model once, at the best combined figure of its versions, and group models into 5 mpg bands listed A to Z, because VCA says its figures should not be used to rank cars with similar figures. Plug-in hybrids and range extenders are left out, as their official mpg assumes a charged battery; fully electric cars use no fuel. Model-year codes in VCA's model names are removed so that one model appears once.

Driving test pass rates

Figures are car practical tests from DVSA table DRT122A for April 2025 to March 2026, by the test centre where the test was taken. The pass rate is passes divided by tests conducted. First-attempt rates come from DRT122C. Each centre's five-year trend uses its yearly total in DRT122A from 2021/22 to 2025/26; a year is missing where DVSA published no figures for the centre under that name. The national first-attempt rate is DVSA's own figure from DRT122C. We rank only centres with at least 1,000 tests in the year, because at small centres a few results move the rate a lot. DVSA says its figures are provisional and can change slightly.

Theory test pass rates come from DVSA table DRT112A (car theory tests by test centre, April 2025 to March 2026) and the latest national quarters from DRT111A. The pass rate is passes divided by tests conducted, and we rank only theory test centres with at least 1,000 tests in the year.

Road safety

Collisions are DfT's STATS19 road safety open data for 2021 to 2025: road collisions on public roads that involved at least one vehicle, injured at least one person and were reported to the police. Collisions where nobody was hurt, and many where the injury was not reported to the police, are not in the data.

Each collision is counted for its current highway authority (the council responsible for the road, on today's boundaries), so councils that have merged show their whole area. Severity is the police's own classification (fatal, serious or slight); we do not use DfT's adjusted severity. A fatal collision is one where someone died within 30 days.

The rate per 100,000 people uses ONS mid-2025 population estimates for the same council. It measures how many fatal and serious collisions happen on a council's roads against how many people live there, not how risky it is to live there: councils with busy through-roads and few residents show high rates.

People killed or injured come from DfT's STATS19 casualty file, one row per person, linked to each collision so they are counted in the council where the collision happened, not where they live. We group DfT's casualty types into road users; e-scooter riders are those DfT flags as using an e-scooter. Severity is the police's assessment at the time and is not adjusted.

Vehicle recalls

Recalls come from DVSA's recalls file (all recalls since 1992), downloaded from DVSA's recall service on 31 October 2026. We use road vehicle recalls only: cars, vans, lorries and buses, and motorcycles; recalls for parts, trailers and tyres are left out. The file has one row for each model a recall covers, so we group rows by recall number and count each recall once, in the year it was launched. DVSA records the maker's UK company name, so we bring the names together under one brand (for example, three Mercedes-Benz company names become Mercedes-Benz). Models are grouped the same way as on our MOT pages. The number of vehicles affected is DVSA's figure, shown only where the file gives one number for the whole recall. We do not publish chassis number or build date ranges. A model appearing in the list does not mean every car of that model is affected: check a specific car on GOV.UK.

Cars in each council

Licensed cars come from DfT table VEH0105 and plug-in cars from VEH0142, at the end of June 2026, for the council where the registered keeper lives. Company cars are counted where the company is registered, which can make a lease firm's home council look high. Cars per 1,000 people use ONS mid-2025 population.

The postcode finder uses the ONS Postcode Directory (August 2026). Each postcode district is matched to the council that holds most of its live postcodes. Postcodes in Northern Ireland are not covered.

How we process the files

The files are processed on our own computer into small summary files, and the pages are built from those. We do not change any figure; we add them up, divide them and round them for display (percentages to one decimal place). Where we group or leave out rows, this page says so. If you spot a mistake, tell us.

All the figures contain public sector information licensed under the Open Government Licence v3.0. The government publishers do not endorse Motoring Insider.