Skip to main content

Table 1 Spearman’s correlation matrices for sepal length, sepal width, petal length and petal width

From: Rank correlation under categorical confounding

\( \hat {\boldsymbol {\rho }}=\left [\begin {array}{rrrr} 1&&&\\ -0.16&1&&\\ 0.88&-0.30&1&\\ 0.83&-0.28&0.94&1\\ \end {array}\right ]\)

\(\hat {\boldsymbol {\rho }}_{1}=\left [\begin {array}{rrrr} 1&&&\\ 0.77&1&&\\ 0.27&0.18&1&\\ 0.30&0.37&0.23&1\\ \end {array}\right ]\)

\(\hat {\boldsymbol {\rho }}_{2}=\left [\begin {array}{rrrr} 1&&&\\ 0.52&1&&\\ 0.74&0.57&1&\\ 0.55&0.66&0.79&1\\ \end {array}\right ]\)

\(\hat {\boldsymbol {\rho }}_{3}=\left [\begin {array}{rrrr} 1&&&\\ 0.43&1&&\\ 0.82&0.39&1&\\ 0.32&0.54&0.36&1\\ \end {array}\right ]\)

  1. While \(\hat {\boldsymbol {\rho }}_{i}\) contain Spearman’s coefficients for each each of the three species of iris, namely Setosa (\(\hat {\boldsymbol {\rho }}_{1}\)), Versicolor (\(\hat {\boldsymbol {\rho }}_{2}\)) and Virginica (\(\hat {\boldsymbol {\rho }}_{3}\)), the matrix \(\hat {\boldsymbol {\rho }}\) contains Spearman’s correlation for the 150 iris taken as a single dataset, hence ignoring marginal discrepancies