0
$\begingroup$

Trying to interpret and write down my SEM results, but not sure if this is 1. correct and 2. well formulated. This is what my models and parameters look like (first unstandardized, then standardized):

enter image description here enter image description here

Unstandardized. “If trait X at time 2 increases by 1 unit, trait X at time 3 increases by 0.32 units on average.” So since those are two personality traits measured by a 6-point likert scale: “If trait X at time 2 increases by 1 point, trait X at time 3 increases by 0.32 points on average.”

Standardized: “If trait X at time 2 increases by 1 standard deviaton, trait X at time 3 increases by 0.34 standard deviations on average.” (SD X t2: .347; SD X t3: .324) “If trait X at time 2 increases by .35 points, trait X at time 3 increases by .11 points on average.”

Are those interpretations correct and formulated as they should be?

Additional questions: Why is the adjusted R^2 for time 1 displayed in the unstandardized model, but not in the standardized model? And why is every textbook warning me from standardized coefficients >1? It is entirely possible that something causes a change of more than one standard deviation…

$\endgroup$
1
$\begingroup$

Yes, I think your interpretation is correct.

It's not adjusted R2 (I think), it's the (residual) variance. It's not shown because it's 1.00.

$\endgroup$
  • 1
    $\begingroup$ how would you explain negative residual variance from trait A (t=1) to trait X (t=1)? $\endgroup$ – AdamO Mar 14 '19 at 17:38
  • 1
    $\begingroup$ That's not residual variance, that's residual covariance. It means that the residuals are negatively correlated (although that's so low, I would guess it's essentially zero). $\endgroup$ – Jeremy Miles Mar 14 '19 at 17:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.