scipy.stats._result_classes.PearsonRResult.confidence_interval#

PearsonRResult.confidence_interval(confidence_level=0.95)[source]#

The confidence interval for the correlation coefficient.

Compute the confidence interval for the correlation coefficient `statistic` with the given confidence level.

The confidence interval is computed using the Fisher transformation F(r) = arctanh(r) [1]. When the sample pairs are drawn from a bivariate normal distribution, F(r) approximately follows a normal distribution with standard error `1/sqrt(n - 3)`, where `n` is the length of the original samples along the calculation axis. When `n <= 3`, this approximation does not yield a finite, real standard error, so we define the confidence interval to be -1 to 1.

Parameters:
confidence_levelfloat

The confidence level for the calculation of the correlation coefficient confidence interval. Default is 0.95.

Returns:
cinamedtuple

The confidence interval is returned in a `namedtuple` with fields low and high.

References

[1]

“Pearson correlation coefficient”, Wikipedia, https://en.wikipedia.org/wiki/Pearson_correlation_coefficient