Hi,
I have posted previously but got no response - so I am trying to narrow my search field down and articulate my questions better.
If I am doing a longitudinal analysis looking at the trajectory of a dependent variable (y) over time (x) - but I have baseline measures of y at x=0-1 to x=0-7, what is the best way to incorporate these measures into my model?
I am trying to decide between the following options:
(1) Use x as a continuous variable, and make a new "covariate" variable using the mean value of x baseline (replicated to be the same size as x).
or (2) Recode x as a categorical variable - giving all baseline x the value 0 - so that each positive increment in x(y) is being compared to the average baseline (x)y.
Any ideas which is best?
Thanks for your time
Dave
I have posted previously but got no response - so I am trying to narrow my search field down and articulate my questions better.
If I am doing a longitudinal analysis looking at the trajectory of a dependent variable (y) over time (x) - but I have baseline measures of y at x=0-1 to x=0-7, what is the best way to incorporate these measures into my model?
I am trying to decide between the following options:
(1) Use x as a continuous variable, and make a new "covariate" variable using the mean value of x baseline (replicated to be the same size as x).
or (2) Recode x as a categorical variable - giving all baseline x the value 0 - so that each positive increment in x(y) is being compared to the average baseline (x)y.
Any ideas which is best?
Thanks for your time
Dave