2
$\begingroup$

Bartels, 2012 found a correlation of 0.21 in what I understand to be within-person correlations (when one trait increased throughout life, the other did as well), however the between-person correlations don't seem significant.

I would expect that if the factors truly influenced each other positively, they would also be correlated between people. Is Bartels' finding statistically confounded? (e.g. via a change in response style) Or is there a way to reconcile these two findings?

Improve this question
$\endgroup$
5
  • $\begingroup$ Maybe I'm missing something, but why do you think Bartels is measuring within-person correlations? $\endgroup$
    – Bryan Krause ♦
    Commented May 19 at 22:27
  • $\begingroup$ Table 3 shows phenotypic correlations which are always accompanied by "(within person)" in the text $\endgroup$
    – Probably
    Commented May 20 at 13:06
  • 1
    $\begingroup$ Ah, I see; they're using that to mean that the IQ and agreeableness are measured in the same person, in contrast to measures between twins where they correlate agreeableness in one twin with IQ in the other twin, not the way you are interpreting it. $\endgroup$
    – Bryan Krause ♦
    Commented May 20 at 14:34
  • $\begingroup$ @BryanKrause Oh, right, thanks! In that case, I'm just curious about the cause of this discrepancy $\endgroup$
    – Probably
    Commented May 20 at 15:17
  • 2
    $\begingroup$ en.wikipedia.org/wiki/Replication_crisis is probably relevant. I would say that the statistical description in both papers is too sparse to interpret clearly what they've done. Unfortunately not atypical. $\endgroup$
    – Bryan Krause ♦
    Commented May 20 at 15:30

0

Reset to default

Browse other questions tagged or ask your own question.