Question about interpreting correlations When reporting a positive correlation, I typically see things worded in the positive sense of the term... For example, for a positive correlation between eating vegetables and health outcomes you typically read something like: eating more vegetables was linked to better health outcomes. But would it be just as fine to report a correlation in the negative sense of the term? So for the same example of a positive correlation between eating vegetables and health outcomes would it be just as ok to say that eating less vegetables was linked to worse health outcomes? Or is it standard practice to always report in the positive sense of the term?
Thanks.
 A: Yes, you can always reverse any variable and call it the opposite.
It usually helps the reader if you keep things positive, using the name that's usually used for the quantity. So you could say less vegetables is negatively associated with  health (or less vegetables is positively associated with more poor health), but you put some work onto the reader when you do that.
It's easier to talk about having stuff (intelligence, money, vegetables) than not having it.
Some variables have a clear description at each end, and you can choose the end that makes more sense:

*

*happiness - sadness

*extraversion - introversion

*mobility - disability

*Illiteracy - literacy

It doesn't make a difference (from a statistical point of view) which end you choose to be positive, but it can help with interpretation if you pick one or the other.
People often have problems with 'larger/smaller' vs 'higher/lower' correlations. 0.3 is higher and larger than 0.2. -0.3 is lower and larger than -0.2, so it can help to keep things positive (if you can).
