[Cayo]

https://www.bbc.co.uk/news/articles/c8978vy2wnjo

[quote] A learner driver spent nearly £1,400 and 60 hours in trying to pass their driving theory test.

The candidate, from Redditch, notched up the highest number of unsuccessful attempts made in 2023 for one person, before finally passing the test.

New government figures also show a total of 93,204 UK practical driving tests taken in the year to the end of March, were at least a candidate’s sixth attempt to pass.[/quote]

Notwithstanding the fact this appears to be a rehash of an article from December 2023 – https://www.bbc.co.uk/news/uk-67610152 – surely there has to be a point someone with authority says “repetition doesn’t make you better, it just shows you’re not competent enough to be allowed behind the wheel”?

Now, I don’t drive, but I appreciate that nervousness can affect someone’s ability to take [i]any[/I] examination, however it worries me that people who need anywhere near 60 attempts to complete the part of a driving test that doesn’t even involve the potential stresses of actually [I]driving[/i] are allowed to be in charge of a motor vehicle.

[quote]New government figures also show a total of 93,204 UK practical driving tests taken in the year to the end of March, were at least a candidate’s sixth attempt to pass.

The success rate for those taking at least their sixth test was 41.4%, compared with an average pass rate across all tests of 47.9%.[/quote]

Imagine how small that percentage would be if those who get banned and those who show by their actions that they shouldn’t without getting banned were subtracted from the figure. How many would counted as ‘passes’ then?

Topic

[OnYerBike]

mdavidford wrote:

 

• What kind of error rate would give you a better than evens chance of passing in 60 attempts?

If you know the answer to 35 questions and then completely guess the remaining 15, the chance of getting at least 8 (in order to pass) is 1.7%, which would mean if you did that 60 times, you would have a 65% chance of passing at least once. 

Obviously the reality is it won’t be a dichotomy between “know” and “complete guess” – some you might know, others you might narrow down to two or three options etc. 

It won’t be compeltely accurate as the probability of getting a question right will vary from question-to-question and test-to-test, but we can still use the binomial distribution to consider statistics, but by considering the probability of answering each question correctly (p_success). If you’re perfectly knowledgeable, you will answer every question correctly (p_success = 1), and so are certain to pass. If you are perfectly ignorant (completely guessing i.e. p_success = 0.25), you will basically never pass (as per mark1a’s calculation). You need a p_success of 0.85 in order to have a >50% chance of passing first time.

Assuming p_success remains constant, in order to have a >50% chance of passing at least once in 60 test sitting, your p_success needs to be above 0.71. 

In other words, on an “average” test sitting, you could expect to score about 36/50 – a definite fail. But keep doing that enough times, without actually learning anything new, and just by chance you will eventually pass.

[mark1a]

mdavidford wrote:

Maths question for anyone who can be bothered:

Given that the theory test is multiple choice, with four choices per question, and a pass mark of 86%​

According to the binomial distribution calculator at WolframAlpha, the probability of getting 43 out of 50 in a single test with 1/4 chance on each question is 1.819e-19, I think I’d go with actually knowing the answers to the questions.

[Jetmans Dad]

mdavidford wrote:

Maths question for anyone who can be bothered:

Given that the theory test is multiple choice, with four choices per question, and a pass mark of 86%

• What are the odds of you passing in 60 attempts if you just pick answers at random?
• What kind of error rate would give you a better than evens chance of passing in 60 attempts?

Maybe if you fail, the pass mark for each subsequent attempt should be increased a little.

So, on your third attempt … 110% in order to pass. I confess that appeals. 

As for the rest of your question. Not high. There is a reason the infinite monkeys with infinite typewriters are given infinite time to come up with the complete works of Shakespeare. 

Given an infinite number of attempts, you would, at some point, reach the pass mark … possibly even on your first attempt. 

[chrisonabike]

According to the regulations,

According to the regulations, the answer can be “none whatsoever”.

Becoming a “licenced driver” is exactly a rite-of-passage: once per lifetime.  Unless you manage to get a conviction and driving ban (and points don’t need to mean prizes either).

No expectation we’ll see it but the whole “graduated licencing” thing not only sounds logical but (for young folks) has some research behind it.

Though I’d graduate it further…

[Mr Hoopdriver]

I wonder if they’ll continue

I wonder if they’ll continue to show as much dogged determination to maintaining their skills once (if) they manage to pass the practical test and actually get out on the roads with the rest of us.

[mdavidford]

Maths question for anyone who

Maths question for anyone who can be bothered:

Given that the theory test is multiple choice, with four choices per question, and a pass mark of 86%

• What are the odds of you passing in 60 attempts if you just pick answers at random?
• What kind of error rate would give you a better than evens chance of passing in 60 attempts?

Maybe if you fail, the pass mark for each subsequent attempt should be increased a little.

[mattw]

Good comment.

Good comment, and a good question to raise.

One thing we could consider is making this part of the needed Graduated Licensing scheme, as proposed by groups such as Brake. The idea seems to be one of those which may be able to get political traction at present.

“You received X marks”, so yo uare entitled to drive Y category vehicles in Z circumstances.

eg Quadricycles capable of up to 28mph in daylight hours only.

One of my suggestions for Graduated Licences for starting-out drivers to reduce the KSIs on themselves and others, is that it restricte for 3 years or up to age 21, which can be reduced to 2 years or age 20 by more advanced training.

I’d suggest a similar thing for less competent elderly drivers as a way of easing them out.

[Author]

Replies