Correction: In “Analysis of Variance” (Inside Science 91, 8 June) the
quantity s12 was incorrectly defined (page 2, column 3, line 13). It
is the sum of squared differences divided by the number of degrees of freedom.
The calculated t-value (page 3, column 1) should therefore be 2.68. This remains
insignificant at the 99% level but would be significant at the 95% level.
