The Student Room Group

Cambridge University to introduce written admissions tests

Scroll to see replies

Reply 100
Original post by Chief Wiggum
http://www.cambridgeassessment.org.uk/Images/109717-medical-and-veterinary-admissions-test-validation-study.pdf

Ah yes, look at this truly outstanding correlation of 0.078 between medicine interview score and Tripos performance. Note especially the slight negative correlation between male students' interview scores and Tripos performance. :p:

(Compare that to the 0.38 correlation between AS UMS and Medicine Tripos performance.)


And this ...

"The low association between Interview Scores and achievement is also very evident here, especially for males, for whom negative correlations are the norm rather than the exception."
Reply 101
Original post by jneill
The point about this news is not the interview. The interview is well established.

The point is about the use of Oxford-style pre-interview tests for many courses...

This gives rise to the possible implication that Cambridge will interview fewer applicants as a result (like Oxford).

We will have to wait for Cambridge to clarify that in due course.

Posted from TSR Mobile


Sorry, I may have drifted off topic. I think Vincrows made the point that the interviews were good for weeding out the obviously coached specimens. It suddenly occured to me that I have seen lots of correlations between UMS or GCSE's and final degree performance but nothing on interviews.

The problem is that the interviews might be weeding out much more than that. Why should females have a better correlation than males?

So I thought I'd ask.

At any rate, now that Cambridge can no longer use the UMS system it would probably be an appropriate time to consider whether the interview system is fit for purpose.
Original post by Pars12
Sorry, I may have drifted off topic. I think Vincrows made the point that the interviews were good for weeding out the obviously coached specimens. It suddenly occured to me that I have seen lots of correlations between UMS or GCSE's and final degree performance but nothing on interviews.

The problem is that the interviews might be weeding out much more than that. Why should females have a better correlation than males?

So I thought I'd ask.

At any rate, now that Cambridge can no longer use the UMS system it would probably be an appropriate time to consider whether the interview system is fit for purpose.


Why does the unavailabity of UMS mean interviews are not fit for purpose? If anything they become more important, not less. If Cambridge isn't receiving public exam info at the point of application (no AS-levels) then the interview provides an additional check point to evaluate the candidate's academic progress and suitability for the course.
Reply 103
Original post by jneill
Why does the unavailabity of UMS mean interviews are not fit for purpose? If anything they become more important, not less. If Cambridge isn't receiving public exam info at the point of application (no AS-levels) then the interview provides an additional check point to evaluate the candidate's academic progress and suitability for the course.


No, no, not causal. Coincidental.

But if there is no correlation between interview score and degree outcome, why do them?

If there is no correlation then it is not doing the things you say even if it appears obvious that it does.

And if there is a negative correlation ...
Original post by Pars12
No, no, not causal. Coincidental.

But if there is no correlation between interview score and degree outcome, why do them?

If there is no correlation then it is not doing the things you say even if it appears obvious that it does.

And if there is a negative correlation ...


What is coincidental?

Anyway, the data above was only for medics. Show me data for the other 26+ courses.

And, as I said, Cambridge must see a value in doing them otherwise they wouldn't invest all the time and expense in holding interview rounds.

Edit to add: you can ask Cambridge themselves here
http://www.thestudentroom.co.uk/showthread.php?t=3830287
or in the upcoming Admissions Tutor thread.
(edited 8 years ago)
Reply 105
Original post by jneill
What is coincidental?


Occurring at the same time but not causally related.

Anyway, the data above was only for medics. Show me data for the other 26+ courses.

I can't. But that just proves my point.

And, as I said, Cambridge must see a value in doing them otherwise they wouldn't invest all the time and expense in holding interview rounds.

That's why I assumed they must have evidence. So far, all we have seen is medicine (and veterinary science) and as Chief Wiggums has pointed out it is not exactly inspiring. Do you think the other 26+ might be even worse?


Edit to add: you can ask Cambridge themselves here
http://www.thestudentroom.co.uk/showthread.php?t=3830287
or in the upcoming Admissions Tutor thread.

Thank you for the link. I will probably have to. Still wondering why they haven't put it up already. There's a lot of stuff on UMS.
Original post by Pars12
Occurring at the same time but not causally related.


I can't. But that just proves my point.


That's why I assumed they must have evidence. So far, all we have seen is medicine (and veterinary science) and as Chief Wiggums has pointed out it is not exactly inspiring. Do you think the other 26+ might be even worse?


Thank you for the link. I will probably have to. Still wondering why they haven't put it up already. There's a lot of stuff on UMS.


Interview scores are much on personal judgement and relative. One colleges 7 is another 8 etc. It will vary alot hence why actual written notes can't be analysed and hence why other colleges read the notes of an interview etc. So if data does cone back it will have a general trend but will not be accurate.
Remember interviews also test surpervision criteria and how they work under the Cambridge system not just purely academic criteria. Although this would count as one of them, being teachable.


Posted from TSR Mobile
Reply 107
Original post by physicsmaths
Interview scores are much on personal judgement and relative. One colleges 7 is another 8 etc. It will vary alot hence why actual written notes can't be analysed and hence why other colleges read the notes of an interview etc. So if data does cone back it will have a general trend but will not be accurate.
Remember interviews also test surpervision criteria and how they work under the Cambridge system not just purely academic criteria. Although this would count as one of them, being teachable.


Posted from TSR Mobile


If interview score has a negative correlation against degree outcome I think it means that the higher the interview score the lower the degree outcome.

It doesn't really matter about supervision criteria. No point being well suited to the Cambridge system and getting a third.
Original post by Pars12
If interview score has a negative correlation against degree outcome I think it means that the higher the interview score the lower the degree outcome.

It doesn't really matter about supervision criteria. No point being well suited to the Cambridge system and getting a third.


Stats can lie very easily funnily enough.


Posted from TSR Mobile
Reply 109
Original post by physicsmaths
Stats can lie very easily funnily enough.


Posted from TSR Mobile


That's what they said to Galileo when they arrested him.
I wonder how they'll do it considering every specification is slightly different.

Will they have an exam dependent on which specification you are doing or will they create their own specification that people will have to learn separately over summer.
(edited 8 years ago)
Original post by Excuse Me!
I wonder how they'll do it considering every specification is slightly different.

Will they have an exam dependent on which assessment you are doing or will they create their own specification that people will have to learn separately over summer.


Maybe find where all the major soecifications overlap! Or like MAT C1-2 basic knowledge but they make it very intersting or like PAT stuff.


Posted from TSR Mobile
Original post by physicsmaths
Stats can lie very easily funnily enough.


Posted from TSR Mobile


But what's a better way of investigating the effectiveness of something?
Original post by Chief Wiggum
But what's a better way of investigating the effectiveness of something?


Exactly, something will always be wrong. Although medicine is different with strict guidlines etc.
For Other subjects I would strongly expect a positive correlation although who knows for sure.


Posted from TSR Mobile
Original post by physicsmaths
Exactly, something will always be wrong. Although medicine is different with strict guidlines etc.
For Other subjects I would strongly expect a positive correlation although who knows for sure.


Posted from TSR Mobile


But surely statistics are a more reliable method than essentially anything else?

Indeed, I wouldn't know what other subjects would show, but I think it would be useful to know. To be honest, I suspect Cambridge colleges may investigate this at a college level.

Interesting quick article about the issue here:

http://www.telegraph.co.uk/education/educationnews/9782370/Drop-Oxbridge-interviews-don-suggests.html

Unless that guy is a serious maverick trying to kick up controversy, his comments would suggest that at that time, no serious efforts at researching the predictive effectiveness of interviews had been carried out by the university?
(edited 8 years ago)
Original post by Pars12
Occurring at the same time but not causally related.

I know what coincidental means...
I was asking what factors you think were coincidental.
Original post by Excuse Me!


Will they have an exam dependent on which specification you are doing or will they create their own specification that people will have to learn separately over summer.


No.

The whole point is they won't require any extra learning.
Original post by Chief Wiggum
But surely statistics are a more reliable method than essentially anything else?

Indeed, I wouldn't know what other subjects would show, but I think it would be useful to know. To be honest, I suspect Cambridge colleges may investigate this at a college level.

Interesting quick article about the issue here:

http://www.telegraph.co.uk/education/educationnews/9782370/Drop-Oxbridge-interviews-don-suggests.html

Unless that guy is a serious maverick trying to kick up controversy, his comments would suggest that at that time, no serious efforts at researching the predictive effectiveness of interviews had been carried out by the university?


Contrast that to Gavin Lowe (Oxford CompSci admissions tutor) who repeatedly said that he considers interviews to be more predictive than MAT or academic history. (This was in the TSR's most recent MAT prep thread.)

Don't think he had done a proper study though.

Original post by jneill
No.

The whole point is they won't require any extra learning.


Is this actually confirmed? I suppose it doesn't make too much of a difference as the new A-levels are quite tightly specified so if there is off syllabus material, everyone will have to learn it
Original post by Pars12
Occurring at the same time but not causally related.


I can't. But that just proves my point.


That's why I assumed they must have evidence. So far, all we have seen is medicine (and veterinary science) and as Chief Wiggums has pointed out it is not exactly inspiring. Do you think the other 26+ might be even worse?


Thank you for the link. I will probably have to. Still wondering why they haven't put it up already. There's a lot of stuff on UMS.


One important point.
They are NOT doing interview to use it as an indicator for future performance (=Tripos) but BECAUSE it can shed a different light on an applicant in a way other aspects of application (like exam grades, PS/SAQ, reference, etc.) cannot do.

By meeting applicants face-to-face and probing them by interview, they are trying to find if they have ability to 'think' (rather than just being good at remembering what they were taught and at exam techniques), if they have flexibility of mind to accept and absorb new/different ideas on which they can build their own understandings/ideas, if they can engage and benefit from their unique style of teaching = supervisions.

http://www.undergraduate.study.cam.ac.uk/applying/interviews/what-do-interviews-involve

http://www.schools-online.co.uk/linlithgowacadem/applying_for_university/cambridge%20interviews.pdf

AFAIK A-level grades is the only piece of application aspect Cambridge has been saying they have found quite good correlation to future Tripos performance. But just using it will only tell them that they are 'good at exams.'
Interview score is another piece of jigsaw puzzle, along with other aspects of application, which helps the admission people to build a more complete picture of each candidate's genuine ability to 'think' and to see if they can benefit from studying at Cambridge.

If you Google, there're lots of articles with quotations by Oxbridge dons on their interview and what they're trying to find out from it....and you can understand it is NOT a correlation to future Tripos performance they're trying to deduct from interview performance. In a very simple term, they are trying to see how your brain works.

http://www.telegraph.co.uk/education/universityeducation/10998008/Perfect-performance-unnecessary-in-Cambridge-interview.html

http://www.telegraph.co.uk/education/universityeducation/student-life/10380653/Oxford-interviews-Difficult-questions-explained.html

http://www.telegraph.co.uk/education/9550464/A-survivors-guide-to-the-Oxbridge-interview.html
(edited 8 years ago)
Original post by shamika




Is this actually confirmed? I suppose it doesn't make too much of a difference as the new A-levels are quite tightly specified so if there is off syllabus material, everyone will have to learn it


I think so: http://www.undergraduate.study.cam.ac.uk/applying/admissions-assessments

"* No advance preparation will be needed, other than revision of relevant recent subject knowledge where appropriate."

Posted from TSR Mobile

Quick Reply

Latest