The Official 2017 Cambridge Applicants Thread [part 3]

I wonder who it could be...

Cheers, do you have a version with the Courses and Colleges?

Can I ask you to spoiler the key sections as the post is rather, er, large

Posted from TSR Mobile

Thanks!
(Just tried and can't get anything decent out of it haha)

Well done! It would be interesting to look at percentages of the total (once that's known) to see if TSRites are more skewed towards a college or subject compared to the overall population.

It's just a pile of mess now hahaha

Nice one. Although comparing with 2016 admissions stats would probably be close enough for now. @NeverLucky can you do that?

Haha, I tried making a graph like Churchill did for their interview scores, but there's some terrible gaps in the frequency distribution hahaha

Sure, I don't really have any work to do at work this week

Yeah it's hard to get the outcome data hahaha

No rush - tomorrow at 9am will be fine
(jks)

And yeah, tell me about it... I think you suffered from external linkages though...

I just ended up scrolling through 100 pages of this thread to identify outcomes

Each candidate with a confirmed offer/rejection was given a generic, anonymised I.D. number and each offer/rejection has been plotted with that candidate’s respective interview rating.

I wonder who it could be...

Finally (for curiosity’s sake), a confirmed successful/rejected applicant was compared with their respective number of posts in the main TSR thread for Cambridge applicants. Data was only taken from the first and second thread.

Note: The data was only gathered from the TSR post-applicant survey. This data is not representative of the entire applicant pool and should not be taken as such. It is also important to remind people that most the data used here was gathered from an applicant’s own assessment of their application.

Starting with course and college choices, out of 461 interview applicants:

Natural Sciences was by far the most popular course at 18% of the TSR cohort. This was followed by Mathematics at 9.67% and Engineering at 8.24%. Anglo-Saxon, Norse and Celtic was the least popular course with only 1 applicant out of 461 applicants (0.22%).

The most popular colleges appear to be St John’s and Downing closely followed by King’s and St Catharine’s. Perhaps unsurprisingly, the least popular colleges appear to be the mature colleges collectively attracting 5 applicants out of 461 (1.08%).

Out of the 461 interview applicants, 396 of them sat a pre-interview assessment (including BMAT etc.). Applicants were asked to rate their pre-interview assessment based off their personal assessment and thoughts. Applicants were asked to give a rating between 1-10 with 1 being the worst and 10 being the best.


A sample size of 73 applicants out of the original 461 was confirmed to be given offers. 66 of who sat a pre-interview assessment. 29 applicants were confirmed to be given rejections; 22 of who sat a pre-interview assessment. Interestingly, on average, applicants that received rejections generally gave themselves a lower rating on their interview performance than those who received offers. However, there is also the consideration that the low sample size of rejected applicants can cause the data to be somewhat skewed.


The general conclusion that can be drawn is that applicants are poor at assessing their own performance at interview.



Note that the data points are set at 50% transparancy so darker data points indicates a greater concentration of data at that point. The wide spread of data seems to support our previous conclusion that applicants are poor at assessing their own performance. It is interesting to note that there appears to be a somewhat positive correlation between an appliant’s rating of their admissions assessment and their interview. In other words, a candidate is more likely to rate their interview performance higher if they rate their admissions assessment higher.

Finally (for curiosity’s sake), a confirmed successful/rejected applicant was compared with their respective number of posts in the main TSR thread for Cambridge applicants. Data was only taken from the first and second thread.

So yeah. This is absolutely stupendous work.

Do you have the coef correl for the interview outcomes? Mine was 0.08 ...

And similarly for the AA outcomes? 0.45

Hopefully it will be enough to help people stop over-analysing their interviews next year...

If you chose 100 numbers at random coming out with a correlation coefficient of 0.08 wouldn't be surprising.