Spearman Correlation - SPSS (part 2)

I demonstrate how to perform and interpret a Spearman rank correlation in SPSS. I also demonstrate how the Spearman rank correlation can be useful when dealing with non-normally distributed data.

Category:

Education

Tags:

• Spearman
• rank
• correlation
• SPSS

[how2stats]

The main arguments against transformations are: (1) all too often, the parametric statistical analysis is relatively impervious to quite substantial deviations from normality, so the transformation is not necessary from the perspective of having faith in the accuracy of the p value; (2) once you transform your data, the inferences from your results, in my opinion, are limited to the transformed data, which are not "real world" data. Thus, your results will suffer from a lack of applicability.

[jdithchrn]

interesting. could you please explain your opinion on why doing log-base transformations is not a good idea and especially if you think this is also the case of TIME SERIES? i think for cross-sectional data you are right, the log-base is an unnecessary approximation which can be surmounted by means of rank ordering, but i don't think this is of any use when dealing with series that are not stationary. thank you!