OCR MEI FP1 Thread - AM 20th May 2016

Knowing MEI 57/72 or something. Not even kidding. Lots of people have said it was hard and I thought so too, but lots disagreed. I'm starting to worry I was just one of the stupid people who think it is hard each year because they aren't very good not because it was hard. Although TBF, last night I got 69/72 in a paper but I would be surprised if I even got 55/72 int that! I know I definitely can't get anymore thats 63/72 anyway, and that would only be if everything else was correct, which it probably won't be

Loads of people in my school were CRYING. I can confirm most people found it hard. I'm thinking c.55 for an A. Even though Jan13 was 48 I noticed January exams often seem to have lower grade boundaries.

These are all the boundaries from january 2005 and all the january boundaries are bold:

57 54 57 58 59 59 55 60 57 53 55 59 57 59 56 63 48 63 63 I don't think you can really say January boundaries tend to be lower. I think you are just seeing a pattern that doesn't really exist. I used to think the same thing at GCSE that somehow January exams are slightly different to June exams or boundaries are different but I don't think that has any real merit to it.

looking at ur numbers i can clearly see that there is a pattern, so in fact i think it is you not seeing a pattern which is actually there.

(from your numbers)
Average Jan boundary = 55.6
Average June boundary = 59.1

That's a significant enough difference to suggest that Jan boundaries are generally lower.

And now seeing that the average boundary is around 59, I think that the A boundary may actually be around 50-55 for this year

Maybe we should carry out a hypothesis test to see if this is true! I'm assuming your doing stats 1 next week right? Might be some good practice

Also I think the 48 has really skewed the data quite a lot for January as well though. Like if we work out the standard deviation I'm sure it will be much higher for January than June.

I worked out average for January without January 2013. Its 56.6. I guess its still a bit lower, but I'm not sure how significant that difference really is. I still think its just a coincidence...

And I guess I could just about pull of a 52/72 A grade boundary (I hope) but I reckon it'd be close...

My prediction is that the grade boundary for an A will be quite low this year. Possibly at an all-time low, since this was definitely the hardest paper in terms of MEI FP1. I discussed it with friends who also did FP1 today and we all agreed on that pretty much.

To sum it up in one sentence: the paper basically had the hardest question types from each topic in FP1.

Unlike other papers, where you have some really lovely questions that require zero brainpower (a bit of an exaggeration, if you pardon), these questions had a bit more of a challenging aspect to them. I would say that, for most previous FP1 papers, there's about 1-2 'tougher' questions and the rest are really predictable. Today, they just went full-on with every topic.

But I'm torn between having a tough exam and lenient grade boundaries - or vice versa (i.e., an easy exam, but really high grade boundaries). Which do you guys think is better?

It has to be an easier exam with higher grade boundaries, I like to come out of the exam confident with my answers, not a wreck who loses confidence in their own maths ability lol

Yeah I prefer an easier exam with less room for error, a combination of this exam and the new Aqa physics spec has knocked my confidence entirely

Lol yeah I am doing S1 next week. I don't know how I'd do a hypothesis test on this. I'm not sure exactly what I'm even testing. Do you do S2? it think this would me a normal sample mean test on whether μ=59.1 when from a sample of 10 the x̄ = 55.6 if X~N(59.1,σ²). You know I am actually gonna do this test.

EDIT: well the test said that the probability that the two sets of scores are drawn from the same population is for all purposes 0. So there is actually a difference between the jan and june that is unlikely to be due to coincidence. though my test is flawed by the fact that i used the sample deviation as population.

We should create a poll or something and send it to MEI. From that we can see how many people actually thought it was hard. The poll should be about the paper being just as hard as the Jan 13 so the boundaries should be around the same. That sounds very reasonable to me.

Think I've been stupid and left the answer to the last question as 0.00816 or something? Didn't realise 0's at the front classed as sig figs lol

what's the answers for the limitation question?

Should you not do a continuity correction since exam data is discrete (the mean is okay as Continuous as the expected value is afterall just a repersentation)

I got 0.00813 3.S.F