OCR MEI Statistics 1 (S1)

and 7 (vi) on http://www.mei.org.uk/files/papers/s107ju_sw8f3.pdf

Thanks

Can anyone explain to me how go about doing a hyposthesis test including the bits about CR's?

Help appreciated!

http://www.mei.org.uk/files/papers/2011_June_s1.pdf

Q7

Hey, unfortunately probability really isn't my strong point but I can try help with 7(vi)

You're working out the probability that the person is tested negative originally OR tests doubtful and is then cleared in the exam.

P(Negative Originally) = 0.91

That's easy. Now the hard(er) bit...

P(Doubtful THEN Cleared) you would just do as:
0.06 * 0.9

BUT with the new examination, if they HAVE the disease, it is still only 98% correct at identifying it. So, it's 2% or 0.02 likely to CLEAR the person incorrectly.

So as well as 0.06 * 0.9 you must include the outcome that they are incorrectly tested as clear, despite having the disease.

Which is 0.06 * (0.1 * 0.02)

So overall, it's:

0.91 + [(0.06 * 0.9) + (0.06 * 0.1 * 0.02)] = 0.96412

Hope this helps

Thanks! makes more sense

I just don't understand why you need to add P(doubtful then cleared) since the question just said "the probability that this person either tests negative originally or tests doubtful and is then cleared in the detailed medical examination..."
Why do you need to include doubtful and cleared without the med exam??

There is no doubtful and cleared without the exam.

EVERYBODY who is doubtful is given the exam... 0.9 of whom are cleared with a 100% reliably. So 0.9 x 1.

Then out of those that are tested that HAVE the disease, only 98% are successfully identified, the other 2% have the disease but are cleared. In this event it is 0.06 * (0.1 * 0.02) the 0.1 * 0.02 being the probability they have the disease but get cleared anyway.

Hope this makes it clearer

Can someone please help me with what you label histogram axis with?? I was fine till I looked at markschemes...

like http://www.mei.org.uk/files/papers/2011_Jan_s1.pdf Q7
they said Accept any suitable unit for fd, eg freq per £10K, BUT
NOT FD per £1000

why can't it be FD per £1000 ???

Because frequency density IS frequency/£1000. So it's Frequency per £1000.

Saying frequency density is frequency density per £1000 is like saying: frequency density = frequency density/1000

thanks again

for the rest of that question, do we not need to multiply mean and sd by 1000? since income is given in 1000s?
for this markscheme they don't do it, but they did in another one :/

On this paper http://www.mei.org.uk/files/papers/s110ju_wtfd.pdf
in the examiners report they say Candidates should be reminded that total accuracy in intermediate working is important when dealing with probabilities eg 0.857375 should not be rounded to 0.86 so do they mean we should give the whole thing?? I know they mean in the working we shouldnt round off

but for the answer to Q7 ii C they gave their answer to 6dp... and it said CAO (correct answer only) so i don't really know what we should do

also, would it be ok to leave answers as fractions??

It says on the extra points "Also accept £40760, £40800 etc"

I honestly can't imagine them severely penalizing you in any case like this, you've done the maths correctly... the only case I remember seeing things like this mattering are when you're dealing with £s and pennies.