[derbOac]

Their Figure 1 is why we keep having these discussions in society. It's grossly misleading and not what a scatterplot for a 0.46 correlation looks like. I know what the figure is, it's just done in a way to overstate a case and ignore variability within bin.

If that figure were about anything else, people would be screaming bloody murder about misleading figures and overly generalized interpretation.

I'm in favor of allowing for the use of test scores but they get abused and the language in this report is a good example of how this happens. Scores have these real but modest correlations with real world situations, but then get used as rulers of atomic precision without any context or recognition of their massive limitations.

It makes the authors of this report look either deceiving or ignorant of statistics or both.