The reasons for discrepancies may include
boundary conditions: there is no observation
before the first, and none after the last
gaps
and
missing values
In particular, you need to understand the difference
between the last two categories. If I have an obs
at t = 1 and the next is at t = 3, then I have a
gap at t = 2. This is not the same as having an
obs at t = 2, but an explicit missing value.
Nick
n.j.cox@durham.ac.uk
Carol Bao
> I have a panel data and using XTTRANS command I get a
> matrix as follows:
>
> | stat
> stat | 0 1 2
> 3 | Total
> -----------+--------------------------------------------+----------
> 0 | 843 1,989 139
> 126 | 3,097
> | 27.22 64.22 4.49
> 4.07 | 100.00
> -----------+--------------------------------------------+----------
> 1 | 1,294 13,259 1,987
> 867 | 17,407
> | 7.43 76.17 11.41
> 4.98 | 100.00
> -----------+--------------------------------------------+----------
> 2 | 99 1,547 1,271
> 834 | 3,751
> | 2.64 41.24 33.88
> 22.23 | 100.00
> -----------+--------------------------------------------+----------
> 3 | 127 837 693
> 2,155 | 3,812
> | 3.33 21.96 18.18
> 56.53 | 100.00
> -----------+--------------------------------------------+----------
> Total | 2,363 17,632 4,090
> 3,982 | 28,067
> | 8.42 62.82 14.57
> 14.19 | 100.00
>
> First of all, the total number of observations here,
> 28067 is not the same as (fewer than) the number of
> observations for the sample if I do .sum. Also, if I
> do:
>
> . tab stat if lag==0;
>
> stat | Freq. Percent Cum.
> ------------+-----------------------------------
> 0 | 946 27.59 27.59
> 1 | 2,174 63.40 90.99
> 2 | 166 4.84 95.83
> 3 | 143 4.17 100.00
> ------------+-----------------------------------
> Total | 3,429 100.00
>
> The number of observations don't equal to the
> corresponding ones in the first row of the matrix
> either.
>
> I really appreciate it if anyone could kindly give me
> some information or tips as to what I am missing here.
> Each person has only one spell of observations on STAT
> in the sample, though the numbers of observations per
> spell diff.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/