I have performed a linear mixed model with the aim to determine the variability in a sports metric called Champion Data Rank across seasons and matches. My analysis was performed in SAS with player and team as random effects (interaction). The entire analysis was done "by season".

I am now trying to interpret the covariance parameter estimates (no log transformation due to potential for negative dependent variable values). I have a fairly basic question about the random effects - what is the interaction between player and team telling me? Is it telling me that an individual players champion data rank score for a match will vary depending on the team that they belong to?

  • $\begingroup$ Have you considered to nest player within team? $\endgroup$ – ocram Nov 15 '16 at 7:22
  • $\begingroup$ No I haven't considered this. How would this change the analysis? $\endgroup$ – Courtney Nov 15 '16 at 8:00
  • $\begingroup$ Well, instead of "random team player team*player", you would have "random team player(team)". That is, a random effect of the team and a random effect of the player within his team. $\endgroup$ – ocram Nov 15 '16 at 8:39

Your Answer

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

Browse other questions tagged or ask your own question.