pmcc, but i doubt they'll ask us, i quoted the syllabus a few posts back saying that they won't ask us, but might just ask what to do if it is tied rank (which they have once in a prev paper)

Reply 345

DCMed96

16

Original post by tggfootymad

To clarify, if there's a tied rank, do you input the average rank into spearman's correlation or person's correlation?

I want to ask the same question as well,.. do you carry out the pmcc test instead and use the original values or the ranks??

Reply 346

tggfootymad

2

Original post by mmms95

pmcc, but i doubt they'll ask us, i quoted the syllabus a few posts back saying that they won't ask us, but might just ask what to do if it is tied rank (which they have once in a prev paper)

Exactly, they might ask us what to do. What do we say then?

Reply 347

Mina Makin

can someone pls help me with june 2014 q7b, why is the critical region -2.3263 and not 2.3263?
thnx

Reply 348

math42

20

Original post by tggfootymad

Exactly, they might ask us what to do. What do we say then?

Think the 2012 one went something like

They are now tied ranks
Average the ranks and use *formula for PMCC*

Don't think it mentions whether or not you use the ranks or the data; I don't see what the point of first averaging the ranks would be if you were going to use the actual data though

Reply 349

DCMed96

16

Original post by jsoso

Does arsey do s3?

no but i can try to make one

Original post by 1 8 13 20 42

Think the 2012 one went something like

They are now tied ranks
Average the ranks and use *formula for PMCC*

Don't think it mentions whether or not you use the ranks or the data; I don't see what the point of first averaging the ranks would be if you were going to use the actual data though

indeed, there is no point of using tied ranks if we are not going to use the spearman rank. IN the textbook, they say it is fairly accurate to use spearman with one or two tied ranks because it is not going to make any difference. But since the mark scheme mentioned pmcc, I was slightly confused because is it implying that we should just use pmcc on the orginal set of data (Sxx, Sxy, Syy stuff) instead?

i hope it won't come out tomorrow, but i guess everyone is so stressed now..

Reply 350

math42

20

Original post by Mina Makin

can someone pls help me with june 2014 q7b, why is the critical region -2.3263 and not 2.3263?
thnx

You are looking for a less than probability rather than a more than probability: we want to know how likely it is that x could be less than or equal to 999.54 when the mean is assumed to be 1000

Usually we deal with the latter, P(z > 2.3263) = 0.01, but P(z < -2.3263) = 0.01 also
For a one-tailed test you can usually discern the sign of the critical region by the sign of your test statistic

Reply 351

tggfootymad

2

Does anyone know which year the paper was which had this tied rank question?

Reply 352

Mina Makin

Original post by 1 8 13 20 42

You are looking for a less than probability rather than a more than probability: we want to know how likely it is that x could be less than or equal to 999.54 when the mean is assumed to be 1000

Usually we deal with the latter, P(z > 2.3263) = 0.01, but P(z < -2.3263) = 0.01 also
For a one-tailed test you can usually discern the sign of the critical region by the sign of your test statistic

Thanks a lot! that makes sense

Reply 353

math42

20

Original post by DCMed96

indeed, there is no point of using tied ranks if we are not going to use the spearman rank. IN the textbook, they say it is fairly accurate to use spearman with one or two tied ranks because it is not going to make any difference. But since the mark scheme mentioned pmcc, I was slightly confused because is it implying that we should just use pmcc on the orginal set of data (Sxx, Sxy, Syy stuff) instead?

i hope it won't come out tomorrow, but i guess everyone is so stressed now..

I'm just planning on regurgitating that 2012 answer; if it's there, considering numerical questions won't occur, it can only really be 2 or 3 marks. Also the same answer, however senseless it seems, as they gave back then should surely still suffice