Chapter 4: Analysing the Data |

## Correlation, slope and z-scoresIf the X and Y variables are first standardised (to Z-scores) before the regression, the regression weight for X will be equal to the correlation between X and Y. This illustrates another interpretation of the correlation coefficient: it is the number of standard deviations that two cases are predicted to differ on Y if they differ by one standard deviation on X. So, for example, if applicant A scored one standard deviation higher than applicant B on the logical reasoning test, weĠd predict that applicant A would be about 0.736 standard deviations above applicant B on the creativity test (because the correlation is about 0.736). |

© Copyright 2000 University of New England, Armidale, NSW, 2351. All rights reserved Maintained by Dr Ian Price |