STEP 2020 grade boundries

My opinion (I sat both):

So, I think it's easier to first talk about S3 boundaries since the difficulty seemed relatively standard and the performance of students seemed it too.

STEP 3:

Average from past 20 exams (2000-2019):
S = 84
1 = 62
2 = 50
3 = 32

Average from past 10 exams (2010-2019):
S = 85.4 = 85
1 = 62.2 = 62
2 = 51
3 = 28.9 = 29

Average from past 5 exams (2015-2019):

S = 87
1 = 62.8 = 63
2 = 52.6 = 53
3 = 28.6 = 29

I think the 2020 paper (in difficulty) is similar to some from 2nd and 3rd averages - some questions are similar to those from 2008 e.g. (the badly done complex number question). So I think, if the paper was sat normally, the boundaries would be about same as 3rd (87/63/53/29).

Effects of being online:
- messed up sleep schedules for a lot of people (from experience and people I know): due to no A levels/schools and freedom during quarantine etc, a lot of students probably had bad sleep schedules which affects performance (sometimes a little, sometimes a lot)
- distractions: self-explanatory e.g. from home, siblings, noises, pressure from family being in the house, etc.
- proctoring issues: being asked to do 360s, interrupting train of thought, wasting your time, questions not loading, proctors not answering to inquiries and questions e.g. "may I go to the toilet etc." which could have an affect on marks
- uploading issues: a lot of pupils had technical issues and lost time but also couldn't upload everything
- potential cheating (I suspect none, if not very little)

So, I personally think overall the fact that the exam was sat online (with all those complications) and non-standard circumstances, this would slightly decrease the boundaries. Also, a lot of students claim very high marks but of course, this happens yearly anyways so it's no reason to suspect insanely high boundaries and not to mention that a lot of people may have "fake-solved" (I did for a few questions lol), done typos and mistakes and issues e.g. did wrong base case for induction or w/e.

My boundary predictions:
S = 86
1 = 63
2 = 52
3 = 29

---------------------------------
Time for S2 (cue the flashbacks) - I think it's a lot harder to think about potential boundaries for this because of the raw difficulty of exam, poor (estimated) performance of most students and shift in style (to an extent) as well as other issues that arise from it being online.

For STEP 2, I don't think it's that useful to look at all the averages since a) the paper was definitely not standard and b) students just did a lot worse than normal. A lot of very talented people performed "poorly" or at least, a lot worse than usual. A lot of people think they got 1/S in S3 but 3/2 in S2 for example. Also, as it's the first exam with no proper warm-up, a lot of people weren't properly prepared for exam conditions e.g. and suffered from anxiety and stuff like that. Anyhow, I think the boundaries could be about record low (I think this may be the first time S3 boundaries > S2 boundaries or at least very, very comparable). I think S2 boundaries could be similar to S3 average boundaries.

I think if the paper were sat normally, S would be about 87 and a 1 around 62-65 (harder and students did less well) considering the 2019 paper (an overwhelming majority of students thought it was a lot easier) was 90/68. With the whole picture, I think S could be anywhere from high 70s to solid 80s and 1 could be anywhere from like 58 to 64 imo.

My predictions:
S = 84
1 = 61
2 = 48
3 = 32

What do others think?

I completely agree with your assessment of STEP II. However, eventhough I personally didn't think STEP III was that much easier than other years, based on what I read here many people did phenomenonaly well. Now people maybe overestimating or maybe they are just a really small group of test takes (hopefully ). But the general consensus seems to be that is was quite easy.

Highest ever STEP III boundries belong to STEP 2017:
S=95
1=69
2=57
3=28

My Predictions:
STEP III
S=97
1=70
2=57
3=32

STEP II
S=83
1=60
2=49
3=30

Step 2 was just a terribly hard paper so I think the grade boundaries will be similar to the recent S3 papers so basically mid 60s for a grade 1 and 80+ for an S. (I'm just praying that I get a 2 in this paper it was so bad)

Step 3 seemed like a standard paper to me so I think grade boundaries will be slightly lower than previous years just because of the online aspects and challenges that came with it so I'm predicting maybe 70 to be a grade 1 and 95+ to be an S.

Wait so you think this years S3 will have the highest EVER boundaries? That seems really unlikely given it wasn't that easy and the online issues etc. Also a lot of people are probably overestimating (eg I came out thinking I got 90s-100s but after reviewing my script, I realised I got a lot lower and a lot of people similarly). Also in S3 more people are awarded the top grades, I think over 1 in 3 people get a 1 so people doing well (which is mostly overestimating etc I think anyways) is fairly standard for S3. Also, I know quite a few people who don't seem to have done amazingly so it seems unlikely that about 36% of students achieved at least 70 given that it's the highest EVER and it wasn't a particularly easy paper and other online complications.

I'm confused sorry. You said you expect S3 boundaries to be slightly lower than previous years but 70 for 1 and 95+ for S would be the highest its EVER been (which is 95/69). I've written above the S3 boundaries averages from past 20/10/5 years so if you're saying it's slightly less than previous years, it'd line up with my prediction of 86/63 ish.

Even though there were online issues, I understand they were very lenient with people who had problems (turned over their answers after the time, mismatched file names, etc.). Also, this year only offer holders participated in the exam so if they maintain the 1 out of 3 ratio, grade boundaries have to be higher than previous years.

..
What do others think?

After having done a sample of the questions (once the Qs were up here) a while back, I basically agree with the difficulty assessment and (although I can only go off what people have said) the added complication of being "disrupted" by the different format.

I think for S2 specifically, there's a good chance the boundaries are slightly lower. This is just due to my conjecture that the difficulty of Q1 & Q2 has quite a big relative impact on boundaries because most students do mostly pure questions and most students will choose questions roughly in order (while correcting for having preferred topics), and the momentum in a 3 hour exam does really matter. I think S2 2020 Q1-2 were tougher to "see" than in previous years. The effect isn't going to be 5+ marks though, perhaps in the realm of 2-3 marks.


Also want to just weight in and say; I'm not sure what the effect of schools and exams being disrupted by COVID will be in terms of the effect on preparation. There are 2 competing schools of thought:
1. Because schools (around the world but importantly in the UK) were essentially cancelled and from a relatively early date we knew that A-Levels were cancelled, everyone preparing for STEP had more time, less (other) academic disruptions/ distractions. So people should have had more time and more flexibility to get in preparation time. (So grade boundaries should be higher).
2. Because everything was disrupted and for a long time there was some serious uncertainty about if (and if not that, at least when) STEP was going to happen. On top of this any kind of school STEP help for people struggling to get a foot in the difficulty door wasn't open (I have no idea how widespread help from schools is from STEP so not sure if this is relevant). On top of this the lockdown was tough and probably had a impact on emotional wellbeing, motivation, etc. I also think it's easier to cultivate a good work ethic if you are already working. So boundaries should be lower.

Overall, I think #2 would outweigh #1 slightly but this is a baseless assumption so I'm really not sure. I think because of the uncertainty around this (as well as exam disruption), we could easily do +/- 5 to each boundary guess

Yeah I agree with this but from mine and others' experience, I don't think the lockdown resulted in more preparation than otherwise eg I think people still did a few hours a day which they would do anyways (at home and some during lessons). What numbers do you specifically predict for step 2 and 3 in your opinion (hard to judge but a rough estimate)?

82/57/45/28 and 85/60/50/30

I'm just reminded a bit of the old gym adage about the hardest part about going to the gym is going to the gym. Same with the hardest part about getting around to completing work is just getting around to starting it. Having very little/ nothing to do can make procrastination easier and knowing that you have a lot of free time over the next few months might encourage mindsets like "i have a lot of time, i can start STEP prep tomorrow". Then the longer you leave it, the more daunting it becomes and the more you want to avoid it.

I don't want to call all smart people lazy, but a very common personality type at Cambridge was the "I will do exactly what you make me to do and nothing more" which was often heavily mixed with "it's due on tuesday night? Ok I'm starting it on tuesday at 8pm then". In fairness to a lot of these, they work extremely hard when they eventually get around to it and never actually failed to hand things in on time. So it's impossible to know but I'd wager that a good chunk of these might have actually done less work than if covid hadn't happened. In contrast "steadier"/ more sensible students probably did more.