I would like to know if there is a free software programme that allows me to calculate Spearman's correlation coefficient easily and quickly.
You can do so in R. Without installing, you can run this code at https://rdrr.io/snippets/ .
You can also install R. Or, for Windows, install it as a portable file on an external device.
There are other ways to input the data if this method is cumbersome with the format you have.
A = c(1,2,3,4,5,6,7,8,9)
B = c(2,4,3,5,4,6,7,7,9)
cor.test(A, B, method="spearman")
plot(A,B)
Dear Antonio,
If you have Excel, go for it. If not, SAS university edition will be applicable for you.
Another idea is to download and install software like Jamovi. It is available for Windows, iOS, Linux, and ChromeOS. It can import e.g. .csv files. It's easy to use, is respected, and produces attractive output.
Analysis > Regression > add variables to the right > check Spearman.
If this is a one-time (or rare) thing, you might find this online calculator convenient:
https://www.socscistatistics.com/tests/spearman/default2.aspx
https://www.socscistatistics.com/tests/
In testing it, entered some numbers in two columns in Excel, and then copied them (one at a time) and pasted into the appropriate boxes for the calculator.
Hi Antonio,
If R feels uncomfortable, you may remember that the idea of Spearman correlation is that, in the special case of rank-ordered variables, Spearman correlation is just a simplification of Pearson correlation - that was the essesnce of "Spearman rank-order correlation". Then, going backwards: Spearman correlation is the same as Pearson correlation if your variables are rank-ordered. So, if you (1) recode your variables as rank-order variables (that is, e.g., a Likert-scaled variable with 1 - 5 is transformed to scale 1 to ~N) and (2) you use Excel (or some other spreadsheer software) and calculate the normal Pearson (product-moment correlation) coefficient between these variables, (3) you get Spearman correlation. Note: using the original variables (without rank-ordered) would give you Pearson correlation (which is usually higher than Spearman correlation)
JMe
Looking at free software, the procedure in the link supplied by Jochen Wilhelm will work also in LibreOffice Calc (which is free software similar to Excel).
Hello J. Antonio,
Regardless of what statistical package you use, you can always: (a) convert each set of scores to ranks (high to low, or low to high...it doesn't matter as long as you are consistent); and (b) compute the Pearson correlation on the two sets of ranks.
That's the Spearman correlation, correctly adjusted for any tied scores!
Good luck with your work.
Edit thanks to @Salvatore
Or take David Morse 's advice a step further so that it is more powerful than Spearman's.
1. Rank
2. Use inverse normal so these are more normally distributed
3. Run Pearson's.
Directions how to in freeware (R) plus simulations showing it is more powerful than Spearman's and often more powerful than Pearson's alone in: Preprint A Robust Alternative to Pearson's Correlation for Testing As...
https://www.xlstat.com/en/solutions/features/correlation-tests
Daniel Wright , it looks like that paper uses an inverse normal scores transformation, which is different from a z-score transformation. I'm pretty sure if you do a z-score transformation it doesn't change the correlation at all.
Sal Mangiafico , yep, writing late at night. not a z score made with a linear transformation.
J. Antonio Cortiñas Rodríguez Ph.D.
https://www.stata.com/manuals/rspearman.pdf
this link is for Stata software. you can use the code and example of that.
also The Spearman correlation can be found in SPSS software under Analyze > Correlate > Bivariate.
Daniel Wright , Oh ! You wrote the article. I probably don't need to point out the methods used ! ....
Sal Mangiafico , it is good you did. My shorthand when writing on paper for the "normrank" transform is often z, but I do know most people think of that as standardizing, so it would make my comment confusing (and I have changed it!).
Spearman coefficent?Hope that the variables have natural ordering(ordinal).Spss has the option but its not free.R is free.Definitly there should be the coding.just google.
Thank you all for the information.
Best regards.
EPI INFO AND SPSS !!: They are the best; simpler EPI INFO, but less complete, even for Spearman's Correlation it is more than enough!
try JASP, in the "Regression" Tab you find Spearman's rho under Correlations
You can use SPSS, Excel or R programs
https://www.youtube.com/watch?v=81GifK9vKxc
- Badges
- Science topic
- Similar topics
- Psychology