I am trying to fit conditional fixed effect model for my panel data set (Unbalanced).
My data set consists of 15 different groups (GroupID) each with 96 unique time variables(TimeID) (with gaps) .

Rough visualization of my data set is as shown below:-
Group ID TimeID V1 V2 V3 V4 V5 V6 DependentV
002 0 2 1 0 1 1 3 4
002 46 0 1 0 3 1 0 3
002 95 1 1 3 3 2 1 1
019 0 1 0 3 1 3 2 4
019 56 1 2 1 0 1 2 3
019 95 0 1 1 2 1 4 5

When I run fixed effect model, I don't get the value for Number of groups as 15. Instead, I get different values for number of groups depending on the dependent variable I use for analysis (8 for some, 3 for some and 4 for some). Also, Number of observations do not match with the total observations of that particular dependent variable obtained by using summarize function.

I was wondering if anybody could help me understand where my mistake could be?

Thank you!


Hopefully, the codes I used and outputs I got as provided below will help readers to understand the issues.

-xtset GroupID TimeID

panel variable: GroupID (unbalanced)
time variable: TimeID, 0 to 95, but with gaps
delta: 1 unit


- xtdescribe, patterns(15)

GroupID: 2, 3, ..., 19 n = 15
TimeID: 0, 1, ..., 95 T = 96
Delta(TimeID) = 1 unit
Span(TimeID) = 96 periods
(GroupID*TimeID uniquely identifies each observation)

Distribution of T_i: min 5% 25% 50% 75% 95% max
22 22 94 96 96 96 96

Freq. Percent Cum. | Pattern
---------------------------+--------------------------------------------------------------------------------------------------
9 60.00 60.00 | 11111111111111111111111111111111111111111111111111 1111111111111111111111111111111111111111111111
1 6.67 66.67 | ............................1111111111111111111111 ..............................................
1 6.67 73.33 | 111111111111111111111111..111111111111111111111111 1111111111111111111111111111111111111111111111
1 6.67 80.00 | 1111111111111111111111111.111111111111111111111111 1111111111111111111111111111111111111111111111
1 6.67 86.67 | 1111111111111111111111111111...1111111111111111111 1111111111111111111111111111111111111111111111
1 6.67 93.33 | 1111111111111111111111111111111111111111.111111111 1111111111111111111111111111111111111111111111
1 6.67 100.00 | 111111111111111111111111111111111111111111111111.. .............111111111111111111111111111111111
---------------------------+--------------------------------------------------------------------------------------------------
15 100.00 | XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX


- xtnbreg DependentV V1 V2 V3 V4 V5 V6, fe

 Conditional FE negative binomial regression Number of obs = 669
Group variable: GroupID Number of groups = 8

Obs per group:
min = 7
avg = 83.6
max = 96

Wald chi2(6) = 694.85
Log likelihood = -1445.7082 Prob > chi2 = 0.0000

------------------------------------------------------------------------------
DependentV | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
V1 | .2890287 .1794249 1.61 0.107 -.0626377 .640695
V2 | .0279617 .0335241 0.83 0.404 -.0377443 .0936678
V3 | .0341106 .0287747 1.19 0.236 -.0222867 .090508
V4 | .0018275 .0001124 16.26 0.000 .0016072 .0020478
V5 | -.014157 .0020148 7.03 0.000 .010208 .0181059
V6 | .0253777 .002351 10.79 0.000 .0207698 .0299857
_cons | -1.308499 .11084 -11.81 0.000 -1.525742 -1.091257
------------------------------------------------------------------------------