Different R squared from fitlm and corrcoeff?

Hello,

I've noticed that I get different R-squared values from the fitlm() and corrcoeff() functions. Here's a simple example:

x = [1,2,3,4,5,6,7];
y = [3,6,4,8,10,10,13];
scatter(x,y)
fitlm(x,y)
corrcoef(x,y)

fitlm() says the ordinary R-squared value is 0.893. corrcoef(x,y) says the R-squared value is 0.945. Why is there a discrepancy between these two functions?

Thanks in advance, Diana

Answer 1

[Peter Perkins]

Diana, you're comparing two different things, though they are related. See, e.g. this wikipedia entry.

Answer 2

[Sebastiano Piccolroaz]

If I'm not wrong, corrcoef() provides the Pearson coefficient R.

In this case, R = 0.945, and R^2 = 0.945^2 = 0.893, that is what is provided by fitlm().