Dear all,

it is not clear to me how to compute a variance partition coefficient for a

dichotomous dependent variable using GENLINMIXED in SPSS 24.

For example, assume I receive the score 1.13 for the intercept variance of a

null model (1.13 is contained in the table "Random effect", first column

labelled "Random effect covariance", second column labelled "estimate", row

labelled "Var [Intercept]").

Would it be correct to consider this score to (a) indicate the random

variance of the intercepts var(u0j)?

Then I would calculate the VPC as var(u0j) / [var(u0j) + (π2 / 3)] = 1.13 /

(1.13 + 3.29) = 25.5%? (e.g. Snijders & Bosker 1999)?

Best,

Nina

--

Sent from:

http://spssx-discussion.1045642.n5.nabble.com/=====================

To manage your subscription to SPSSX-L, send a message to

[hidden email] (not to SPSSX-L), with no body text except the

command. To leave the list, send the command

SIGNOFF SPSSX-L

For a list of commands to manage subscriptions, send the command

INFO REFCARD