2
$\begingroup$

I calculated the bias and variance of sample mean $\hat{\mu}$ and sample variance $\hat{\sigma}^2$ but I could not calculate the bias and variance for sample correlation $\hat{\rho}$. How can I do the calculation? Can you give me some references, textbooks, articles about it?

Edit: by sample correlation I mean $$ \hat{\rho}_{xy} = \frac{\sum_{i} x_i y_i}{\sqrt{\sum_{i}x_i^2 \sum_{i}y_i^2}} $$

$\endgroup$
2
  • $\begingroup$ Can you explain what you mean by bias, and/or how you calculated it? $\endgroup$ Commented May 8, 2014 at 22:55
  • $\begingroup$ By bias I mean the difference between the estimator's expected value and the true value of the parameter being estimated $\endgroup$
    – neticin
    Commented May 9, 2014 at 9:15

1 Answer 1

1
$\begingroup$

A basic into text should give the variance (is this what you mean?): $\sigma^{2}_{\rho}=\frac{1-\rho^{2}}{n-2} \approx \frac{1-\hat{\rho}^{2}}{n-2}$

Fisher addresses the bias: Fisher, R.A. (1915). Frequency distribution of the values of the correlation coefficient in samples from an indefinitely large population. Biometrika, 10, 507-521.

For a fuller treatment see: Zumbo, B. D., Williams, R. H., and Zimmerman, D. W. (2003). Bias in estimation and hypothesis testing of correlation. Psicológica: Revista de metodología y psicología experimental, 24(1):133– 158.

$\endgroup$
5
  • $\begingroup$ In fact, I mean sample correlation between two variables x, y $\endgroup$
    – neticin
    Commented May 9, 2014 at 9:06
  • $\begingroup$ @Alexis good answer. Is there any bias in rho ρ ? If it is there, please explain to me. $\endgroup$
    – user10619
    Commented May 9, 2014 at 14:24
  • $\begingroup$ @neticin look at the last term of my equation on the first line, that's the sample correlation. $\endgroup$
    – Alexis
    Commented May 9, 2014 at 14:47
  • $\begingroup$ @Alexis can you point me a textbook where I can find this one because I didn't understand why the first equality is true $\endgroup$
    – neticin
    Commented May 9, 2014 at 15:41
  • $\begingroup$ Sure: Glantz, S. A. (2005). primer of biostatistics. McGraw-Hill Medical, New York, NY, 6th edition. $\endgroup$
    – Alexis
    Commented May 9, 2014 at 16:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.