Im doing some analysis and am a bit rusty on my stats. I need to find out

(a) Which test do I need to use to examine the correlation between an interval-level variable and a nominal-level variable ? (preferably one that is available in SPSS)

(b) I have looked at the correlation between an inteval level variable and an ordinal level variable. I used the Spearman rho test to do this and it turned out to be significant at 0.051.

However I have read in sporadic places on the web that if the ordinal variable has 5 or more choice (which my`n does) then it may be treated as interval level. Running pearsons test, gives a co-efficient which is significant at 0.026 and thus I would much prefer to report this result and test than the above one.

If this is allowed, can anyone suggest any references I may be able to check/quote that show this is ok as I only have (non-authoratative web refs at the moment which would not be appropriate for the paper writing.

Many thanks for any help anyone can give me with either these queries


(mods- sorry about repeat posting, accidentally posted in homework help section when actually this section is more approriate)