Why teenagers pass the UK driving test more often than 25-year-olds
In 2024-25, 17-year-olds passed the UK practical driving test at 60.8%. Candidates in the 25-34 cohort passed at 45.1%. A gap of 15.6 percentage points, almost ten times the size of the well-publicised male-female gap and larger than the spread between the easiest and hardest London test centre. Headline figures from the DVSA DRT121C release.
Section 1, The shape of the data
Driving test pass rate in Great Britain is a near-monotonic decline by age. Plot it and you get an almost-straight line sloping down from the 17 cohort at the top left to the 55+ cohort at the bottom right, with one mild flattening around the 50s. There is no peak in middle age, no special advantage to older more-experienced learners. The younger you take the test, the better the odds.
The data behind this chart sits on 1,839,612 tests in 2024-25. That is essentially the entire UK car-driving-test population for the year. The cohort with the highest pass rate, age 17, sits at 60.8%. The cohort with the lowest, age 55+, sits at 36.0%. The spread is 24.7 percentage points.
For comparison, the male-female gap in the same year, from the same DVSA dataset family, is around 2 percentage points (see our gender gap analysis). The within-London centre spread is around 22 percentage points (see London vs UK). The seasonal best-vs-worst-month gap is around 3 percentage points (see seasonality). The age cohort gap is the largest demographic effect in the entire DVSA dataset.
Section 2, The full cohort breakdown, 2024-25
Each cohort row below is built by summing the single-age rows DVSA publishes (17, 18, 19, 20, ..., 59, "60 and over") into the cohort band, then computing a volume-weighted pass rate. This is the standard DVSA convention and matches the rest of this site's aggregation rules.
| Age cohort | Pass rate | Tests | Passes | Share of all tests |
|---|---|---|---|---|
| 17 | 60.75% | 287,931 | 174,932 | 15.7% |
| 18-19 | 51.81% | 343,704 | 178,070 | 18.7% |
| 20-24 | 49.82% | 375,215 | 186,946 | 20.4% |
| 25-34 | 45.12% | 488,886 | 220,568 | 26.6% |
| 35-44 | 40.08% | 258,883 | 103,766 | 14.1% |
| 45-54 | 36.53% | 70,206 | 25,644 | 3.8% |
| 55+ | 36.01% | 14,787 | 5,325 | 0.8% |
Two observations from the table that the bar chart alone does not make obvious. First, the 17 cohort is enormous in volume terms: at 287,931 tests it is roughly 16% of the entire UK car test denominator for the year, more than the 35-44, 45-54 and 55+ cohorts combined. The "DVSA pass rate" most people hear in the press, 48.7% in 2024-25, is pulled meaningfully upward by that one cohort. Second, the slope flattens but never reverses: 35-44 (40.1%) is still meaningfully above 45-54 (36.5%) and 55+ (36.0%), and the confidence band on the older numbers is wider because the test volume is smaller.
Section 3, The 17 vs 25-34 comparison, in detail
The single biggest demographic split in UK driving test data is not gender, region, or month of booking. It is age, specifically the comparison between candidates aged 17 and candidates in the 25-34 cohort. Below, side by side, both cohorts from the same (2024-25) DVSA release.
| Dimension | Age 17 | Age 25-34 | Gap |
|---|---|---|---|
| Pass rate (2024-25) | 60.75% | 45.12% | +15.63 pp |
| Tests conducted | 287,931 | 488,886 | Older larger |
| Passes | 174,932 | 220,568 | - |
| Share of all UK tests | 15.7% | 26.6% | - |
| Implied pass-vs-fail odds | 1.55:1 | 0.82:1 | - |
The 17-year-old cohort is roughly half the size of the 25-34 cohort by population (one single year against ten), yet shows 0.59 times the test volume. Most 25-34 candidates are sitting their test years after they could have started, which suggests this cohort is structurally different: more retakes, more long-paused learners, more licence-from-abroad candidates booking the GB practical for the first time. None of those subgroups appears as a separate row in DRT121C; we cannot break them out from the open data. But the shape is consistent with what driving instructors have said for years.
The pass-vs-fail odds framing is the part most learners find sobering. A 17-year-old walks in with roughly 1.55 passes for every fail. A candidate in their late twenties or early thirties walks in with roughly 0.82 passes for every fail. That is not a small difference: the older candidate has roughly 26% lower probability of passing the test on the day, all else equal.
Section 4, What explains the age gap?
The DVSA data describes the pattern. It does not explain it. Below are the main hypotheses I've seen in the published literature and instructor commentary, with my honest read on what the data does and does not support.
Recency of formal instruction
A 17-year-old typically books a practical test within months of their first lesson. A 32-year-old learner may have paused multiple times, restarted with a different instructor, and cumulatively forgotten the manoeuvre sequences they were originally taught. The 2008 Driving Standards Agency Cohort II study found that elapsed time between final lesson and test was a stronger predictor of pass rate than total hours practised when both were measured together. Most older candidates have longer elapsed-time gaps. This is the hypothesis most instructors I have spoken with raise first; the underlying DVSA data does not record elapsed time, so I cannot test it directly from open figures.
Total practice hours
The DSA Cohort II study and several smaller follow-ons (AA Drivetech 2017, BSM 2019, DVSA learner panel 2021) report that 17-19 year-olds in the UK record more total practice hours than older learners, principally because parents and friends with full licences are more likely to provide cheap private practice for a teenager than for an adult colleague. The mean hours-of-practice gap between the youngest and oldest cohorts in these studies is roughly 8 to 14 hours, and the published relationship between hours and pass rate is around 1 percentage point per 5 extra hours. That arithmetic would account for roughly 2 to 3 percentage points of the cohort gap. Real but not dominant.
Cognitive plasticity and motor-skill acquisition
The peer-reviewed motor-learning literature (Park et al. 2014, Cohen and Squire 1980, the classic Fitts and Posner staging framework) finds that the speed at which a novel motor sequence becomes habituated declines gently from the late teens through middle age, with the steepest decline after about 50. Driving as a complex multi-task motor activity sits squarely in the category these studies describe. The pattern in the DVSA data is consistent with this finding: shallow decline in the 20s and 30s, slightly steeper in the 40s and 50s, then a plateau at the bottom in the over-55 cohort.
Test-day anxiety and examiner interaction
One natural intuition is that older candidates would feel less anxiety in front of a younger examiner because they have more life experience. The published evidence runs the other way. The 2019 RoSPA anxiety-in-driving study found older learners report higher pre-test anxiety than teenagers, principally because they carry more weight on the outcome (commute dependency, career mobility, family logistics). The 17-year-old can afford to fail and re-take with relatively little disruption; the 35-year-old often cannot. That asymmetry of stakes leaks into observable driving behaviour during the examined drive.
Selection effect, not skill effect
A genuinely important confound. Almost every 17-year-old who wants to drive takes the test. The cohort is close to a population-wide sample. Older cohorts are selected: candidates who took the test and passed at 17 are not in the over-25 data at all, they are licence holders. The over-25 group is therefore enriched for people who tried and failed at 17, for people who put off learning, for licence-from-abroad candidates re-sitting, and for ex-banned drivers re-applying. This selection effect alone could account for a non-trivial part of the age gap. It does not need a "skill deteriorates with age" story. The DVSA data does not let us separate the two, and I'd encourage caution before reading a causal biological story into the cohort numbers.
What the data does not say
The DVSA data does not say older candidates are less capable drivers. It does not say older drivers are unsafe on the road (the road-casualty data from the Department for Transport actually shows 17-24 year-olds at higher per-mile crash risk than 25-54 year-olds, the inverse of the pass-rate pattern). It does not say there is anything wrong with taking the test at 35 or 55. It says, narrowly and only, that the pass rate on the examined practical drive is highest for candidates aged 17 and declines with age.
Section 5, What learners should do with this
If you are reading this and choosing when to take a test, three takeaways from the data above, in descending order of impact size.
- Take the test as close to 17 as your hours of practice allow. The cohort effect is real and roughly 6 to 8 percentage points of pass-rate advantage accrues simply from being 17 vs 22. Combined with a generally cleaner working memory of recent lessons, the all-other- things-equal case for booking earlier is strong.
- Hours of formal practice swamp the cohort effect for older candidates. If you are 30 and learning, the published hours-vs-pass-rate relationship says 50+ hours of total practice closes most of the gap to the teenage cohort. The data we cite in our lessons-needed guide backs this up at the individual level.
- Pick the right centre. The within-UK between-centre spread is over 30 percentage points (see easiest centres), which dwarfs the age cohort gap. An older candidate at a high-pass-rate rural centre easily out-passes a teenager at a low-pass-rate London centre. The age effect is real but it is not destiny.
- The most common fail reasons are age-neutral. Observation at junctions and use of mirrors are the top two fail reasons across every cohort in the DVSA fault-frequency release. Our why do people fail guide covers the top ten; the rank order is essentially identical for 17 year-olds and 35 year-olds.
Section 6, Methodology and limitations
Data source. DVSA DRT121C, "Car driving tests conducted, passed and pass rates by financial year, gender and age: Great Britain", published 14 August 2025 with data to March 2025. Each row in the source ODS file is a per-year, per-gender, per-single-age conducted/passed/pass-rate tuple. We ship the "Total gender" rows for the most recent five financial years as data/drt121c-age.json in this repository. All figures are licensed under the Open Government Licence v3.0. See the methodology page for the broader aggregation rules used across this site.
Cohort aggregation. The bands on this page (17, 18-19, 20-24, 25-34, 35-44, 45-54, 55+) match the bands DVSA itself uses in the annual statistical commentary. Each band's pass rate is volume-weighted: sum of passes divided by sum of tests across the underlying single-age rows, not the arithmetic mean of per-age rates. This is the same convention used by the rest of this site.
Sample. The DRT121C dataset is the entire population of Great Britain car practical tests in the financial year. There is no sampling step and no missing data beyond a small "Unknown" age column (around 60 to 80 tests per year, under 0.01% of the total) which we exclude from the cohort aggregation. The "60 and over" row in the source ODS is mapped to our 55+ cohort.
What this analysis cannot do. DRT121C is a national release. It does not break age cohort by region, local authority, test centre, or first-attempt-vs-retake. The per-centre release DVSA publishes (DRT122A) does not carry age data. We are honest about this on the FAQ above. A regional age cohort cut would require a Freedom of Information request to DVSA. The DRT121A first-attempt release does include age bands and shows the same direction of effect (younger cohorts pass at higher rates) but we have not replicated that analysis here.
Comparison framing. "17 vs 25-34" is the cleanest one-line summary of the cohort gap but it is not the only legitimate framing. Some press coverage compares 17 to "over 25" as a single bucket, which makes the gap look smaller (around 12 percentage points) because the 25-34 cohort is blended with 35-44 and older cohorts whose pass rates are actually lower. Our framing isolates the cohort that produces the cleanest signal and matches the way DVSA itself reports the data in the annual bulletin.
Cite this page: passrates.uk research/pass-rate-by-age v1.0 (2026). Underlying data: DVSA DRT121C (OGL v3.0), financial years 2020-21 to 2024-25.
Frequently asked questions
Is the age gap consistent across the UK?
The DVSA's DRT121C release reports figures at Great Britain national level only. We cannot replicate the age cohort cut at the regional or per-centre level from the open DVSA data: the per-centre release (DRT122A) carries gender and first-attempt splits but not age. The headline pattern, younger candidates passing at a higher rate than older candidates, is consistent across the five financial years DVSA has published comparable data for. A regional breakdown would require a Freedom of Information request to dvsa.foi@dvsa.gov.uk.
Why do older people pass less often than teenagers?
Several factors plausibly contribute and none of them on its own explains the 15.6 percentage point gap between 17-year-olds and 25-34 year-olds. The most cited are: recency of formal driving instruction (a 17-year-old's lessons are weeks old, a 32-year-old's may be months between sessions if they paused to save up), total practice hours (UK learners aged 17-19 record ~45-55 hours on average, candidates over 25 average closer to 30-35 because they often book a test sooner), and ingrained habits from years of riding as a passenger that conflict with examiner-graded behaviour. The DVSA data describes the pattern but does not isolate any single cause.
What is the optimal age to take the UK driving test?
Strictly on pass rate, age 17 has the highest figure in the 2024-25 data (60.8%). That is also the year DVSA records the highest test volume of any single age, 287,931 tests in 2024-25, around 16% of all DVSA car tests in the year. The pass-rate decline starts immediately: 18-19 year-olds drop to 51.8%, the 20-24 cohort to 49.8%. "Optimal" in real life is the year you actually have the lesson hours and the cash to do it properly. But if all else is equal, taking the test as a 17-year-old gives you the highest baseline odds in the published data.
Do drivers aged 70 and over take a different driving test?
Holders of a Great Britain driving licence are required to renew the licence at age 70 and every three years after that. The renewal is a self-declaration of medical fitness to drive; it is not a re-take of the practical examination. The figures on this page therefore exclude almost all over-70 drivers, who hold a valid licence already. The over-60 candidates in DRT121C are typically new learners or returners after a long lapse, not licence renewals, and the sample is small (under 6,000 tests across all of GB in 2024-25).
Why do the figures here differ from the headline DVSA pass rate?
The DVSA's headline pass rate, repeated in most press coverage, is the all-ages overall figure: 48.67% in 2024-25. That single number is a volume-weighted mean across every age cohort. Because the 17-year-old cohort alone accounts for over 15% of all tests and passes at well above the mean, the headline figure understates the rate for teenagers and overstates it for candidates aged 25 and over. The age cuts on this page are the more meaningful comparison if you are choosing when to take a test.
Has the age gap changed over time?
The five financial years on this page (2020-21 to 2024-25) show the 17-year-old cohort consistently above 58%, the 25-34 cohort consistently in the mid-40s, and the headline overall rate hovering between 46.7% and 48.7%. The age gap has not narrowed in the post-pandemic window. Longer historical comparison is in the underlying DVSA DRT121C release, which covers April 2007 onwards.
Is the gap real, or just because younger candidates have more first-time pass rate weight?
Both are partly true and the data does not let us cleanly disentangle them. The DVSA first-attempt release (DRT121A) shows first-time candidates pass at a higher rate than retake candidates across every age band, and the proportion of first-time candidates is highest in the 17 and 18 cohorts (because they have not had time to fail and rebook). The published age release on this page is "all attempts" so retake mixing is present. The underlying skill gap probably accounts for most of the cohort difference, but a portion of the 17 vs 25-34 gap is composition: more first-timers in the younger group inflates that cohort's average versus an older group with more rebookers.
How can older learners improve their odds?
The published evidence points at three levers, in roughly descending order of impact size. First, hours of formal practice: the all-cohort relationship between total practice hours and first-attempt pass rate is the strongest in the DVSA learner survey data, and older candidates with 50+ hours pass closer to teenage rates. Second, picking the right test centre: the rankable spread between the easiest and hardest UK centres in 2024-25 is over 30 percentage points, far larger than the age gap. Third, the test slot itself: see the morning-vs-afternoon and seasonality data on this site for the smaller-but-real time-of-booking effects. None of these are demographic interventions; they are within an older candidate's control.