Spearman Rank Correlation Coefficient

Watch this thread
This discussion is closed.
sah_davies
Badges: 0
? You'll earn badges for being active around the site. Rep gems come when your posts are rated by other community members.
#1
Report Thread starter 18 years ago
#1
I sat an exam today that asked you to state the strengths and weaknesses of the Spearmans Rank Correlation Coefficient. Other than saying it wouldnt be much use if the data you were correlating had errors in, I couldnt think of what to write! It's definately cost me 1/5 of the exam paper. Does anyone know the answer??
0
lou p lou
Badges: 2
Rep:
? You'll earn badges for being active around the site. Rep gems come when your posts are rated by other community members.
#2
Report 18 years ago
#2
(Original post by sah_davies)
I sat an exam today that asked you to state the strengths and weaknesses of the Spearmans Rank Correlation Coefficient. Other than saying it wouldnt be much use if the data you were correlating had errors in, I couldnt think of what to write! It's definately cost me 1/5 of the exam paper. Does anyone know the answer??
all i can suggest is that because it takes every piece of data into account, one anomolous result can affect the whole result, particularly if n is small.
oh, and don't you have to have a normal distribution for it to work properly? *is sure she had to prove this in coursework*, so most of the time you end up assuming normality, when it might not be the case.

and the strengths are that the result is easy to work with ie. the closer to 1 the stronger the correlation etc. and it can be easily checked using a graph ie. right shape.

hope this helps. *useful post*

lou xxx
0
]{ingnik
Badges: 1
Rep:
? You'll earn badges for being active around the site. Rep gems come when your posts are rated by other community members.
#3
Report 18 years ago
#3
(Original post by lou p lou)
all i can suggest is that because it takes every piece of data into account, one anomolous result can affect the whole result, particularly if n is small.
oh, and don't you have to have a normal distribution for it to work properly? *is sure she had to prove this in coursework*, so most of the time you end up assuming normality, when it might not be the case.

and the strengths are that the result is easy to work with ie. the closer to 1 the stronger the correlation etc. and it can be easily checked using a graph ie. right shape.

hope this helps. *useful post*

lou xxx
on a less technical basis spearmans rank takes bloody ages. it is an extremely precise method of measuring correclation, which may not b suitable (you can knock up a nice scatter diagram in seconds with a line of best fit, which will give you a rough idea of correlation) and, as has been mentioned is likely to be affected severly by dirty data etc.
0
X
new posts
Back
to top
Latest

Do you know what you'll do if you don't get the grades you're hoping for?

Find something else in clearing (24)
30.77%
Take a gap year (14)
17.95%
Resit my exams (17)
21.79%
Look for alternate pathways to the career I want (11)
14.1%
I don't know yet (9)
11.54%
Something else (tell us in the thread) (3)
3.85%

Watched Threads

View All