repeated time values within the same panel of repeted measures

[Mario Saugo]

Hallo

have multiple observations of Systolic Blood Pressure and Diastolic Blood Pressure for the same ID and the same monthly lag time since the enrollment (lag==0).

Example data:

----------------------- copy starting from the next line -----------------------

* Example generated by -dataex-. For more info, type help dataex
clear
input double(SBP_V DBP_V) float(Date_V VISIT enrolled_new) long ID float(Date_enrollment lag)
170 80 21780 0 0 1 21780 0
120 80 22910 0 0 2 22910 0
130 80 23098 0 0 3 23098 0
120 70 23245 1 0 3 23098 4
175 95 23161 0 0 4 23161 0
170 95 23209 1 0 4 23161 1
140 85 23245 2 0 4 23161 2
140 80 23274 3 0 4 23161 3
150 90 22342 0 0 5 22342 0
135 80 23121 0 0 6 23121 0
end
format %tdDD/NN/CCYY Date_V
label values ID ID_encoded
label def ID_encoded 1 "AdeMdof1943Kip", modify
label def ID_encoded 2 "AdeMgof1961Kip", modify
label def ID_encoded 3 "AgnValf1975Kip", modify
label def ID_encoded 4 "AjeMarf1966Mwa", modify
label def ID_encoded 5 "AleMtem1976Mga", modify
label def ID_encoded 6 "AnaMdef1954Isi", modify

------------------ copy up to and including the previous line ------------------

In order to evaluate the change in SBP measurement over subsequent time lags I would choose do xtset and anlalyzing the data this way:

. xtset ID

. xi: xtreg SBP_V i.lag
. xi: xtreg DBP_V i.lag

I kindly ask if it is appropriate to omit "lag" in the xtset command in this case.

Mario Saugo

[Carlo Lazzaro]

Mario:
if you omit the time variable in -xtset-, the time-series operator will not work.
In addition, the -xi:- prefix is redundant with the -fvvarlist- notation.

[Joseph Coveney]

I kindly ask if it is appropriate to omit "lag" in the xtset command in this case.

It wouldn't do anything in that case even if you could include it.

Including a time variable in xtset is useful only if you're going to be using time series operators (lags, leads etc.) or xtreg , pa corr() with a working correlation structure that requires it. The default options for xtreg as you're using it are for the random-effects estimator, which makes no use of the time variable in xtset, and so you might as well omit it from xtset's specification.

Why are you using the xi: prefix?

[Mario Saugo]

Dr. Lazzari and Dr. Coveney thanks for your correction. You are right, the xi: prefix is redundant here

[Carlo Lazzaro]

Mario:
Carlo is enough. Thanks.