Spearman vs Pearson correlations

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Spearman vs Pearson correlations
Message posted by Karen Putnam on September 7, 2000 at 12:00 AM (ET)

When is it appropriate to use the Spearman vs the Pearson, when doing correlations

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Spearman vs Pearson correlations
Message posted by nancy diehl on September 7, 2000 at 12:00 AM (ET)

If the data collected is qualitative or there's an ordering to the data, then most likely the data will
be analyzed using nonparametric statistical procedures. If you are testing the hypothesis for association
(correlation) between two variables nonparametrically, then the Spearman rank correlation coefficient would
be used. However, if the data is measured on a quantitative scale or can be studied in a
parametric manner, then the Pearson correlation coefficient is used. The nonparametric Spearman
method uses ranks and Pearson uses the raw data values.

Re: Spearman vs Pearson correlations
Message posted by JG on September 8, 2000 at 12:00 AM (ET)

As explained above, the two correlations are used for different data, such as using pounds for weight and quarts for volume.

Re: Spearman vs Pearson correlations
Message posted by Phil on September 8, 2000 at 12:00 AM (ET)

I believe the first response is more helpful. If you have sets of rankings for a group, (for example, two judges who have ranked items as first, second, third, fourth, etc.), then use Spearman's to see how much correlation there is between the two judges' rankings. If you have quantitative data (pairs of measures expressed as numbers x,y), then use Pearson's to see if their is any correlation between the two variables.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.