I want to point something here - for everyone -.
When you want to use correlations, you can't just drop the number like that because a correlation of 0,2 on 10 points isn't the same as a correlation of 0,2 on 100 000 points.
We need significativity to analyse it properly. Or at least, the number of occurrences in the study.

Besides, jungian functions aren't a "real" things. It's a reading grid. There is no such things as "Fe likes everyone" or "Si is healthy behavior". Many SLI had poor health and are Si-leading for instance... It's a really simplistic and misleading approach to jungian functions. I can understand that people who discover jungian typology (MBTI, Socionics...) use those shortcuts but in the end it doesn't work like that.