Edexcel AS Further Maths (New Spec) - Statistics Discussion

So this is tomorrow...
I'm doing FS1, feeling pretty confident right now, probably should check up on exam answers for why a Poisson distribution works for an estimate of the binomial distribution and those types of questions.
I think the numerical side should be fine.

Pretty much same! Which other unit are you taking? I'm finding mechanics a complete pain, so if stats is bad then it's all gonna go to pot as I need the extra time from stats to do mechanics

I'm taking FM1 too! I find it harder as well so I reckon it might be a good idea to start with mechanics and move onto statistics at the end. I always get adrenaline at the end so I think that it'll increase my writing speed for the long Chi-squared testing procedure (and typing speed in my calculator). Whereas the adrenaline will muddle up my thinking bigtime in mechanics.

Which topics do you find hardest in stats?

Cool! Ahh right. That wouldn't work for me as I'm tight for time usually anyway, so I need to do the stuff I can pick up the marks on easier (eg. stats) first to maximise my marks.
The hardest bit for me is probably using one probability distribution to represent another...like sometimes it works and it's great and sometimes it's like what?! Also I always seem to get the degrees of freedom wrong on those type questions as well which is weird. How about you?

I mean if I'm on a question and know that I can't do it I'll skip it too--but I usually have a clearer head at the beginning of an exam than the end if I've prepared myself properly.

Chi-squared testing was probably the hardest for me to get my head around too, but I'd just say I need to make sure I have all the wording right. I figure that I'm liable to mix up which statement is the null hypothesis and which one is the alternative hypothesis, so I need to remember that contingency tables have H0 as no association and H1 as association, while general Chi-squared tests have H0 as the model works and H1 as the model doesn't work.

I think that degrees of freedom should be OK...it's ((number of cells after combining) - 1) if you haven't calculated the probability or lambda and it's ((number of cells after combining) - 2) if you have calculated it, right?

What practice have you done?

Yeah me too, at the end I have to really try and stop myself from panicking I'm running out of time
Definitely...wording is such a big thing, and getting it the right way round as well. I always forget the proper conclusion as I'm lazy in class and don't write it, so I need to remember that as well...
Yeah I think it's that, I have a tendency to forget when I've just calculated lambda though (literally just did it in a question now!) And also on contingency tables it's (rows-1)*(columns-1)
We haven't been finished long so we only did the specimen paper just over a week ago, so I've done that, most of the questions Edexcel provided the school on certain topics and I've just done the sample assessment material paper tonight (am currently marking, and stats has gone relatively well, we'll just see about mechanics lol)

Haha, I'm going to try and blitz through the sample paper tomorrow morning before the exam, most of my previous revision was rinse and repeat from the textbook (and the specimen)! Now I think it's probably best to get some rest for tomorrow, no point in running out of time because I fell asleep midway through ..gn!

It was a relatively nice paper, I got 67 which is definitely my best score (usually I much up at least 2 of the mechanics questions and today it was only one which helped), although no impulses and very little work, power and energy might have had something to do with that
Yeah I need to sleep too, was meant to go up ages ago
Good luck tomorrow!

Posted from TSR Mobile

Yeah good luck to you too!

Thank you!
Quick question... Do we need to know type 1 and type 2 errors? In a stats lesson right now and despite our teacher saying we didn't need to know it she just started teaching it

Posted from TSR Mobile

Nope don't need to know that...your teacher's probably starting next year's course early.

How did you find it? What do you think the grade boundaries for A and B over the two papers will be?

I'm really annoyed at that question 1e where the answer was 0.506 as I kept getting 0.507...

Someone told me about this after the exam; for 1e) you had to use the X~Po(1.5) to find P(X = 6), use that as the p value in the binomial distribution Y~B(200, p), and then use the poisson approximation (yes, you're converting it back to Poisson again) Y~Po(200p) in order to get an answer for 1 - P(Y = 0) that rounds to 0.506.

Personally I think it's a little unnecessary considering how similar the two answers are, but that's what the question implied by referencing the approximation.

It was okay... How about you?
I'm really annoyed at that question 1e where the answer was 0.506 as I kept getting 0.507...
And too much poisson for my liking.
It's really difficult to tell tbh. Maybe around 120-130 for an A and like 20 less for a B?

Posted from TSR Mobile

Oh my goodness really? I did think it was odd that it said using poisson approximation but I'd have never thought of that Oh well

Posted from TSR Mobile

I mean there was always going to be a lot of Poisson in this paper because it's the only new distribution we've been taught at AS level, and I didn't find that type of question that bad. True, the show that question took me a little while to get, but thats more because they made us use a convoluted method to get there.
What did you get for the testing? I got that the critical region was greater than or equal to 13 so reject null hypothesis for heaters, that you reject null hypothesis for the contingency tables as you get chi squared value and test statistic which are both close to 11, and for the first one the Poisson distribution is a good fit.

Also, what did you get for variance? I got something like 25.

Your marks seem to foreshadow a nightmare for next year, because the gap between A and A* would also be 20 marks so you need a 90% to get it--hope it's lower!

I did the poisson approximation and got 0.506