Pearson correlation with data sets that have values on different scales

3

I have two datasets with which I want to do a Pearson correlation analysis. I have carried out the analysis which makes sense, however I want to be sure it is valid given that both datasets have values on different scales.

The features in both datasets are exactly the same (the actual samples are of course different). The range of values are as follows:

dataset1 = 3-20
dataset2 = 10-30

Now my understanding is that pearson correlation coefficient is not dependent on the change or difference in scale of the different datasets, however for all examples I have seen, although the scales are different, the lower value is always 0. As you can see though, the lower value on one dataset is 3 and the lower value on the other is 10. (in either case though they mean they same level of 'low', it is just how they are measured.

Can anyone confirm this is still fine?

user9317212

Posted 2020-12-20T11:33:32.297

Reputation: 59

Question was closed 2021-01-25T05:08:30.967

what is the analysis you carried out and the results you computed ? you seem to have improper readings and interpretations about scaling ? – Subhash C. Davar – 2021-01-24T13:39:12.813

Answers

3

For quantitative-quantitative comparisons, scales generally do not matter in spearmean, pearson, and kendall correlation.

hH1sG0n3

Posted 2020-12-20T11:33:32.297

Reputation: 1 165

1thank you very much! I guess i will not change the data in any way in this case! – user9317212 – 2020-12-20T12:13:03.617

1

@user9317212 please see What should I do when someone answers my question?

– desertnaut – 2020-12-20T13:03:08.370

1Thank you @desertnaut, swstos. – hH1sG0n3 – 2020-12-22T10:44:02.710