I am trying to figure out correct command for fixed and random effects

for longitudinal censored continuous data.

The data is proportional data, lasting from 0 to 1, without either 0

or 1. And the data is longitudinal across 30 years. In the model, I

want to address the fixed and random effects of years OR congresses

(every two years).

TIPS:

1. the D.V. is proportional data, but not a usual proportional. It is the sum of three proportional scores divided by three, though the result is never bigger than 0.8...

2. also, my data is not exactly panel data people usually think. Each case is unique. But there are a lot of cases in each year, and the whole dataset covers continuous 30 years. I am just considering that this year's average level may influence the next year's. So, this is still a longitudinal data.

I have found: xttobit can address random effects for longitudinal

censored continuous data, but I do not know how to address fixed

effects.

My questions:

1, what are the correct command of STATA to address fixed effects for

longitudinal censored continuous data. Any good examples?

2, xttobit can address random effects, but can you recommend any good

examples? Thanks.

Jwang