R help - Jul 2012 - Plotting trajectories of ordinal variables over time by cheating

Hello.

I have an ordered dependent variable (scale 0 - 3), 5 measurement points. I
am afraid I have strong ceiling effects in my data, and would like to plot
the data (trajectories).

However, you know that plotting ordered variables isn't really feasible.

My question: would you think it appropriate to create a new variable with
let's say 200 categories out of my 1000 subjects, and averaging 5 subjects
into each group?
That way I would still get 200 trajectories and would have pseudo-metric
variables. It would just be for the visualization of the data.

If that's possible, how would I tell R to do this? (I know how to plot the
trajectories, but not how to recode 1000 subjects into 200 new categories)

Thanks
Eiko

	[[alternative HTML version deleted]]