longitudinal data analysis - mixed model regression
Posted: Fri Jul 22, 2011 11:57 am
Hi,
I am trying to explain why a football player gets released (0) or retained (1) by a youth football programme after each season(dichotomous outcome) on the basis of fitness variables (predictors), across time.
My data set started in 2007 where I measured several continuous predictors such as height, mass, agility in players from the following age groups (U9, U10, U11, U12, U13, U14, U15, U16, U18). I re-tested this group of players (the players retained anyway, and new players came in) about 3 times a season for the next four seasons.
However, at the end of each season the U18s ‘graduate’ and there is a new intake at the U9 age group as the players move to the next age group (as in a school system). I therefore have varying repeated measures on these participants.
I thought a mixed model logistic regression would be suitable with repeated measures at level 1 and the player at level 2.
Any thoughts / advice would be appreciated.
Chris
I am trying to explain why a football player gets released (0) or retained (1) by a youth football programme after each season(dichotomous outcome) on the basis of fitness variables (predictors), across time.
My data set started in 2007 where I measured several continuous predictors such as height, mass, agility in players from the following age groups (U9, U10, U11, U12, U13, U14, U15, U16, U18). I re-tested this group of players (the players retained anyway, and new players came in) about 3 times a season for the next four seasons.
However, at the end of each season the U18s ‘graduate’ and there is a new intake at the U9 age group as the players move to the next age group (as in a school system). I therefore have varying repeated measures on these participants.
I thought a mixed model logistic regression would be suitable with repeated measures at level 1 and the player at level 2.
Any thoughts / advice would be appreciated.
Chris