I have run an analysis where I am examining whether the percentage of females in the sample moderates the relationship between experiences of mistreatment and withdrawal behaviours. The b weight for the coefficient (percentage of females in the sample) is -.75. How do I interpret this correctly? For every one percentage increase in females in the sample, the correlation between experienced mistreatment and withdrawal decreases by .75? The number .75 seems way to high. How can a correlation decrease by .75?