The Student Room Group

How do SJT scores work?

So I've just sat the UKFPO online practice paper and got a score of 988/1216 or 81%, it doesn't tell me (anywhere that I can see) what this would equate to out of 50 points.

Obviously in the real thing it is all weighted against other candidates scores - but does anyone know roughly how many points 81% would get? Surely it can't change too much year on year, a SJT score of 81% doesn't really mean anything to me - and it would be helpful to know how I'm getting on with it, any help would be much appreciated :smile:
I'm not exactly sure how they work it out but they scale it somehow.

Last year 83% was average, which equated to around 38 points out of 50 (I think, may be 39!)
Every single person I know got 80-83 on the practice tests, so you are within that.
I would think that 0.81 x 50 wouldn't be far off?

They don't force the Gaussian distribution onto how people score (the marking is set before the exam and not adjusted based on what other candidates get, so in theory everyone could get 50 points which should equal 100%). I believe the reason for this, is that it is only 1 part of your total score out of 100, and the total score is then used in the score ranking VS. preference system when you apply for jobs which makes the adjustment in a way (and thus it would be unfair to do this twice).

These documents may be of interest if you've not seen them already: http://www.foundationprogramme.nhs.uk/download.asp?file=Stats_and_facts_FP2014_interim_report_4_April_2014_FINAL.pdf

http://www.foundationprogramme.nhs.uk/download.asp?file=Understanding_your_SJT_score_FP_2014.pdf
(edited 9 years ago)
Reply 4
Original post by Differential
I would think that 0.81 x 50 wouldn't be far off?

They don't force the Gaussian distribution onto how people score (the marking is set before the exam and not adjusted based on what other candidates get, so in theory everyone could get 50 points which should equal 100%). I believe the reason for this, is that it is only 1 part of your total score out of 100, and the total score is then used in the score ranking VS. preference system when you apply for jobs which makes the adjustment in a way (and thus it would be unfair to do this twice).

These documents may be of interest if you've not seen them already: http://www.foundationprogramme.nhs.uk/download.asp?file=Stats_and_facts_FP2014_interim_report_4_April_2014_FINAL.pdf

http://www.foundationprogramme.nhs.uk/download.asp?file=Understanding_your_SJT_score_FP_2014.pdf


Thanks very much everyone!

Thanks in particular for those links they've really helped me, cheers! :smile:
Original post by Differential
I would think that 0.81 x 50 wouldn't be far off?

They don't force the Gaussian distribution onto how people score (the marking is set before the exam and not adjusted based on what other candidates get, so in theory everyone could get 50 points which should equal 100%). I believe the reason for this, is that it is only 1 part of your total score out of 100, and the total score is then used in the score ranking VS. preference system when you apply for jobs which makes the adjustment in a way (and thus it would be unfair to do this twice).

These documents may be of interest if you've not seen them already: http://www.foundationprogramme.nhs.uk/download.asp?file=Stats_and_facts_FP2014_interim_report_4_April_2014_FINAL.pdf

http://www.foundationprogramme.nhs.uk/download.asp?file=Understanding_your_SJT_score_FP_2014.pdf


That's not entirely true, while yes they do not force the distribution on the scores, however, they do give the lowest score 0 points and the highest, 50 points.

So say the exam was out of 1000 marks, and the range of scores were miraculously 990.0-999.9/1000, The person scoring 990 would be given 0 points, and the person scoring 999.9 would would be given 50.

There would then be a distribution of scores between 990.0 and 999.9, who will be allocated scores between 0 and 50 points. So it makes no difference to the range of scores unless everyone gets the exact same score.


Posted from TSR Mobile
(edited 9 years ago)
Where did you come across that? From reading around it (and looking at previous year's score distributions), I'm not quite sure that's the case. However, I'm totally willing to admit I may be wrong!

If that is so, then it's a bit unfair given that you have already been subjected to that adjustment for your decile score.

Also, when I suggested multiplying the score above, it was only for a ball park approximate score.
Original post by Differential
Where did you come across that? From reading around it (and looking at previous year's score distributions), I'm not quite sure that's the case. However, I'm totally willing to admit I may be wrong!

If that is so, then it's a bit unfair given that you have already been subjected to that adjustment for your decile score.

Also, when I suggested multiplying the score above, it was only for a ball park approximate score.


"SJT scores are reported on a 0 to 50 scale with the lowest mark receiving 0.00 points and the highest mark receiving 50.00 points. The distribution of the scale is set to match the mean and distribution of the EPM scores."

Page 2 of the second document you posted.


Posted from TSR Mobile
Original post by carcinoma
That's not entirely true, while yes they do not force the distribution on the scores, however, they do give the lowest score 0 points and the highest, 50 points.

So say the exam was out of 1000 marks, and the range of scores were miraculously 990.0-999.9/1000, The person scoring 990 would be given 0 points, and the person scoring 999.9 would would be given 50.

There would then be a distribution of scores between 990.0 and 999.9, who will be allocated scores between 0 and 50 points. So it makes no difference to the range of scores unless everyone gets the exact same score.


Posted from TSR Mobile


This is true.

And at an SJT course I booked on, they mentioned that most people tend to score 800ish out of 1040, so Carcinoma is right in that it really is your relative position when the scores are ranked that affects your scaled SJT score and not necessarily the actual score you get. The lowest score, no matter how "high" it is, scores a 0.00 and the maximum score, no matter how close it was to the minimum score, scores a 50.00. Its a shame, but that's my understanding of it.
Cool, that's not too great news though!
Original post by Differential
Cool, that's not too great news though!


It is the opposite of great news, it does not bode well for any of us.

I think I will just end up pretty familiar with all the GMC guidelines and the SJT monograph by the time I sit it.
Original post by carcinoma
That's not entirely true, while yes they do not force the distribution on the scores, however, they do give the lowest score 0 points and the highest, 50 points.

So say the exam was out of 1000 marks, and the range of scores were miraculously 990.0-999.9/1000, The person scoring 990 would be given 0 points, and the person scoring 999.9 would would be given 50.

There would then be a distribution of scores between 990.0 and 999.9, who will be allocated scores between 0 and 50 points. So it makes no difference to the range of scores unless everyone gets the exact same score.


Posted from TSR Mobile


Does this mean some people are bound to "fail" irrespective of their score?
Original post by The Only Rivo
Does this mean some people are bound to "fail" irrespective of their score?


Yes, but only 2-3% of candidates will score less than 30/50.

ImageUploadedByStudent Room1414974256.566199.jpg


Posted from TSR Mobile
Original post by carcinoma
Yes, but only 2-3% of candidates will score less than 30/50.

ImageUploadedByStudent Room1414974256.566199.jpg


Posted from TSR Mobile


Yeah I know that, but I still have a big problem with the system. The reason is it is not suitable in the hypothetical scenario where all candidates get a 'good' score, in other words a score which requires answering a big chunk of the questions correctly. I know there will always be a few people who bomb it but in that scenario no one should 'fail'.
Original post by The Only Rivo
Yeah I know that, but I still have a big problem with the system. The reason is it is not suitable in the hypothetical scenario where all candidates get a 'good' score, in other words a score which requires answering a big chunk of the questions correctly. I know there will always be a few people who bomb it but in that scenario no one should 'fail'.


In any case it is better than the highly subjective white space questions.

Which are subjective at the answering level, where people had two weeks to answer them, get help from doctors in post and more senior. Subjective at the marking level where your answers would be marked by your first choice foundation school, which could in cases give preference to the local students.

So yes the SJt is not perfect, it's not weighted correctly and the questions are not ideal. However, it's an improvement at least everyone is in the same boat for this.


Posted from TSR Mobile
Original post by carcinoma
In any case it is better than the highly subjective white space questions.

Which are subjective at the answering level, where people had two weeks to answer them, get help from doctors in post and more senior. Subjective at the marking level where your answers would be marked by your first choice foundation school, which could in cases give preference to the local students.

So yes the SJt is not perfect, it's not weighted correctly and the questions are not ideal. However, it's an improvement at least everyone is in the same boat for this.


Posted from TSR Mobile

Agreed.
I wish they could just use the EPM part and scrap the SJT/white space stuff. Given that the EPM represents 5 or 6 years of medical school and the other test is just a randomiser!
Can anyone tell me how I know if I passed or failed my STJ? I received the feedback report immediately after I had finished the test, but it doesn't say anywhere on it how many points I scored or if I passed or failed.
:-/
This is an old discussion but hopefully someone will be able to help.
Thanks

Latest

Trending

Trending