Official Thread: OCR MEI S2/M1

If I'm honest I can't remember how much I used, but as long as you picked the biggest possible scale that you saw as sensible you'll be fine.

Worst case scenario (I think, please remember I'm not an examiner xD) is you'd lose 1 for lack of specificity if your graph is too small.

You can't of expected to get all the marks mate head up

Conclude is a very assertive word, can you remember the whole sentence? If you can I can give you a better opinion there.

Got pretty much all the same but got 6513 for sigma, 58350 for mean and 40094 for h

For why is Poisson a good approximation distribution? I put:

number of possible genes mutating is large and probability of a gene mutating is small.
Hence normal is good approximation to binomial (I meant to say Poisson will I lose both marks)?

For why is Poisson a good approximation distribution? I put:

number of possible genes mutating is large and probability of a gene mutating is small.
Hence normal is good approximation to binomial (I meant to say Poisson will I lose both marks)?

For why is Poisson a good approximation distribution? I put:

number of possible genes mutating is large and probability of a gene mutating is small.
Hence normal is good approximation to binomial (I meant to say Poisson will I lose both marks)?

Phew, all this talk of S.D. = 1.74 had me worried until I just remembered using sqrt(3.03/60) as the standard error (the denominator of the Z equation), and sqrt(3.03) is 1.74. Noice.
Damn variance trying to rustle my jimmies.

same, there wasnt any reason to work out the SD was there?

Posted from TSR Mobile

for all the hypothesis throughout the paper i got reject H1. for the very last question i got 0.3692<1.645

for 4(ii) i got 5.324<5.991

Was any taught the modelling assumptions for spearmans rank???? I was never told it had to be monotonic

All I can see is that a "monotonic" set of data is any data with a correlation.

And assuming a correlation so that you can test for a correlation seems a bit weird to me. Plus on page 5 of this I saw there being no assumptions:
http://www.mei.org.uk/files/pdf/spearmanrcc.pdf

Monotonic does not mean correlated. It means that the function is either always increasing or always decreasing. So an exponential curve for example is monotonic, but a quadratic curve is not. So if the plotted points exactly fell perfectly on a quadratic curve (I.e a perfect association) SR would give an association of or near zero.


On page 5 in the section 'Codnitions underlying the Spearman's test' it says "For the Spearman test to work, the underlying relationship must be monotonic"

That's exactly what correlated means lol. If they're correlated, one increasing means the other increasing, or visa versa.

And exactly. It says both.