Repeated measurement with different interval

[Disorn Su]

Hi everyone,

I have dependent variable "pressure" measured several time with different interval in multiple groups eg.

case 1 (group 1) : pressure 10 at month 1, pressure 5.5 at month 2.5, pressure 3 at month 6;
case 2 (group 2): pressure 13 at month 2, pressure 8 at month 6, pressure 3 at month 8.5
etc

there are several variables I need to test whether they affect "pressure" or not. I understand that I should use mixed model ?

And how can I get a graph of mean change of "pressure" overtime in each group?

Thank you so much !!

[Joseph Coveney]

You can include time interval as a continuous predictor in your regression model. If you believe that there is a substantial nonlinear trend of the response over time, then you can form polynomial interaction terms or splines of the time variable. You'd include interaction terms involving the grouping variable as needed. For graphing the predicted outcome over time, you'd use margins and marginsplot after fitting the regression model.

help factor variables
help mkspline
help margins
help marginsplot

It's not entirely clear from your post whether you have repeated measures on individuals within groups, but if so then you can fit a hierarchical / multilevel / mixed-effects regression model, and you can include the time variable as a random coefficient. (Is your question about accommodating autocorrelation?)