[Gretl-users] Re: cross-sectional panel data
On Wed, 28 Sep 2022, Alison Loddick wrote:
Thank you.
I wonder if changing the data format will make it easier - as I
will have to teach students with limited ability.
If teaching advanced data manipulation is not on your agenda, my
recommendation would be to create a proper gretl panel dataset (gdt
file) using the method I showed, and share that with your students.
I did try the process in the manual, but I kept getting only one
year of data when I put it into the panel structure
The method discussed in the manual is designed to handle a somewhat
simpler case, where the organization of the original source data is
closer to, though not equal to, what gretl wants. Frankly, the
format you're talking about is pretty weird and I'd be surprised if
any econometric software could make sense of it without a good deal
of work.
In case there's any doubt, what gretl wants from a panel dataset is
a column for each variable and a block of T rows, in temporal order,
for each of the N units or "individuals". Considered as a matrix it
will be NT x v, where v is the number of variables. This is the
"natural" generalization of cross-sectional data (N x v) and
time-series data (T x v).
Allin