How to read CSDD statistics on driving schools and instructors
CSDD (the Road Traffic Safety Directorate) has something rare for this market — official exam statistics by driving school and by instructor. It's a useful tool, but percentages make it easy to fool yourself. Here's where the data lives, how I read it, and where people most often trip up.
Where the official statistics live
On csdd.lv, under "Autoskolas", there's a page called "Statistika par autoskolām" — it leads to the driver-training statistics page with downloadable files: theory exams by school, practical exams by school, theory teacher statistics, and instructor statistics — over 6 months and over a year. What's counted is the first-attempt result. The exact tables and columns you'll see by opening the files themselves — the format changes over time.
What matters is how this data is built: it's not a "best schools" ranking, just raw pass percentages. You need to read them with a clear head.
The baseline: roughly a third pass first time
To judge whether "school N has a 40 % pass rate" is a lot or a little, you need a baseline. According to CSDD data for May 2024 – May 2025 (reported by Latvia's public broadcaster LSM), around 50,000 people took the category B driving exam, and roughly 33 % passed on the first attempt. So nationally, about one in three passes first time.
Against that, 40 % is above average — and the "impressive 33 %" in an ad is simply the average result. Without a baseline, any number sounds convincing.
Trap one: small samples
Picture a school where 8 students reached the exam in half a year and 5 of them passed — 63 %! Sounds brilliant. But with eight students, a single pass or fail moves the percentage by 12 points. It's like flipping a coin eight times and drawing conclusions about the coin.
A school with hundreds of students has a stable percentage; a small one jumps from period to period. So next to any percentage, always look for the number of candidates: a figure built on a large sample weighs more than a pretty percentage built on a tiny one. For the same reason, for instructors I'd look at the one-year table rather than the six-month one — the sample is bigger.
Trap two: a school's average is not your instructor
Within one school, instructors differ — and the spread between them can be wider than the spread between schools. A school's average is made of its strong and weak instructors, but you'll be taught by one specific person. It's no accident CSDD publishes separate instructor statistics: if you're going to compare anything, compare your future instructor's row, not the school's signboard.
One more distortion worth keeping in mind: schools decide themselves who gets to the exam. Where the internal exam is strict, only well-prepared students reach CSDD — and the pass rate looks higher. That doesn't necessarily mean the teaching is better: it may mean the filtering is harsher. That's my reasoning, not a fact from the tables — but it's exactly how percentages work.
What the statistics won't tell you
The numbers are silent on what matters most to a learner: whether you feel calm with this person in the car, whether they go over mistakes or shout, what language lessons are in, whether they're honest about timelines, whether they pad the hours. All of that you find out differently — from reviews, a conversation, and a first lesson. I've covered how to check it in how to choose a driving instructor and what to look for in a driving instructor, and what training actually costs is here.
How to use the numbers wisely
- Compare against the baseline: nationally, about 33 % pass on the first attempt.
- Look at the number of candidates, not just the percentage — a small sample devalues the percentage.
- For instructors, take the one-year file, not the six-month one.
- Remember the filtering: a high percentage can mean strict selection, not better teaching.
- Percentages are background. Decide based on a conversation, reviews, and a first lesson.
I keep track of my own results too: in recent years, 44 % of my students pass the CSDD driving exam on the first attempt — against roughly a third nationally. You're welcome to ask me about those figures directly.
The CSDD data and how the statistics are organised were verified in July 2026; check the current files and procedure at csdd.lv.
The best way to check any instructor isn't a table — it's one lesson with them. You can book a first one here.