How many people resit in the first year?

Original post by Imshiningsolo

I was just wondering how many people in percentages actually resit in the first year. I keep on receiving different things and it will help me to understand how hard first year actually is

Its highly dependent on the course and university, some places its a fraction of 1%, other places it might be much higher.

You can often get course specific data on this kind of thing from Unistats. Also students might be resitting just one or two sub par exams, its pretty rare to fail everything. Remember also that dropout rates will include students who left for medical / family / financial / visa isses etc as well as exam failure, thats about 50/50 in my experience.

(edited 4 years ago)

Reply 2

gjd800

Resit rate in my dept for first years is tiny. We get maybe 2 a semester

Reply 3

Depends quite a lot on the subject and of course variance year-to-year.

I'd say STEM gets a lot more resits & fails, just in the nature of if you cant answer the questions you get zero, even if you attempt some questions if you dont know what your doing you might as well not answer, hence theirs always some students who have poorly prepared for exams and hence its easier to flop. That said generally marks are fairly linear to paper marks so its much easier to score 80+%.

Im guessing like 10-25% resit in STEM but i think less than 5% have to re-do the whole year.

Reply 4

Original post by mnot

Much less than 10-25% resit rates on a good STEM course in my experience, more like 2-3%. Yes quite a few bits of a STEM exam will have a very well defined numerical answer, but a sensible examiner will always set questions that require you to tackles a number of sub-elements of increasing difficulty, e.g. write down obvious equation from lectures, then apply to simple case, now derive special case, next throw tricky maths tool X at the answer to generate new result, use that to explain weird thing mentioned in passing in problem sheet 5. For the last few sections, students can get them completely wrong, but still get useful marks by tackling them in a sensible way, a lot of the skill of marking well goes into picking through that mess.

To see how bad things can get on a non STEM degree with >60% dropout rates take a look at the following link.

https://www.standard.co.uk/news/two-thirds-of-students-dropping-out-of-some-courses-at-uks-biggest-universities-a3640286.html

Ouch !

Reply 5

Original post by Mr Wednesday

Yes sorry, im not to sure tbh,
It was very much guessing 10-25% for having to resit any given exam during the course and 5% for resitting a whole year.

TBH it was very much a guess I never turned up to the resit exams (humble brag) so never saw just haw many or few people there were, I only knew 1 person who resat a year from my undergrad and but knew a couple who resat exams and tried to very loosely estimate.

Reply 6

Original post by mnot

Yes sorry, im not to sure tbh,
It was very much guessing 10-25% for having to resit any given exam during the course and 5% for resitting a whole year.

If you see 25% re-sitting your exam (for class sizes bigger than 4) then its usuaully you as an examiner that have probably done something wrong :afraid:

. What you do tend to see in a big class is three very distinct populations of students rather than a nice single Gaussian. There is a bump at the high end (fully engaged with course, did all the problem sheets + all past papers), bump in the middle (solid students, came to most of the lectures, missed a few problem sheets, did a few old exam questions ) and a little bump at the low end (that student on the class list who never showed up and then asked you to re-explain the _entire_ course in the final office hour session).

For "accidentally tough" exams (which can happen 1st time you teach a new course) you can of course tweak the marking scheme once you have done 20-30 papers to skew the average where you want it. We also have an algorithmic rescaling system that cuts in for "unexpectedly high failure rates" based on past performance of individual students, but that requires sign off by the entire staff in the examiners meeting before we accept the outcome so is considered a bit of a nuclear option.

(edited 4 years ago)

Reply 7

Original post by Mr Wednesday

If you see 25% re-sitting your exam (for class sizes bigger than 4) then its usuaully you as an examiner have probably done something wrong :afraid:

I meant 10-25% of the cohort on any exam so if in year 1 there are say 7 exams, then in the cohort 10-25% will fail one of the 7 exams not that high on any 1 exam it will be that failure rate (although on my masters I did have an exam with 20% failure rate, although this module only had 25 people so it was put down to low sample size), tbh it was a guess hence why i left such large boundaries.

Reply 8

Original post by mnot

So that's a nice analysis but misses the key point that students who fail one exam are much more likely to be the same student that also fail another course unit or two. You see strong correlations in the tail end of a class rather than a nice even distribution of fails over the entire cohort. This means that you have quite a small number of students accounting for most of the re-takes, and one or two "one off" outliers who had a bad day and just have one thing to redo.

(edited 4 years ago)

Reply 9

Original post by Mr Wednesday

So that's a nice analysis but misses the key point that students who fail one exam are much more likely to be the same student that also fail another course unit or two. You see strong correlations in the tail end of a class rather than a nice even distribution of fails over the entire cohort. this means that you have quite a small number of students accounting for most of the re-takes, and one or two "one off" outliers who had a bad day and just have one thing to redo.

Fair, noted. As a say it was a guess, probably the numbers were a bit over zealous.

Reply 10