Regressing specific years using panel data transformed from wide to long

[Lalo Salamanca]

Hi All,

I just transformed some wide panel data into long form panel data so that I could carry out fixed effect analysis easily.


Before I had columns representing variables by year so e.g. profit21, profit20, profit19,......, employees21, employees20,..... and so on

now it looks more like:

yr	profit	employees	debt
17	value for 2017	data	....
18	value for 2018	data	....
19	value for 2019 ....	data	...
20	.....	data	....

and so on. Sorry if this was obvious i am still fairly new to stata


now my question is

before I could regress:

reg profit21 employees19 debt20 ......

but now how can i regress combination of different years like I did when it was in wide form?

I've been making new variables but its starting to look like wide form again which seems like a waste of time and generally wrong.

Any help is appreciated

[William Lisowski]

Do you mean, for example, that you only want to regress with the data for 2017-2019?

regress y x if inrange(yr,2017,2019)

Or, for example, you only want to use the data for 2015, 2017, and 2019?

regress y x if inlist(yr,2015,2017,2019)

For more details

help inrange()
help inlist()

its starting to look like wide form again which seems like a waste of time and generally wrong

Agreed, wide layout is wrong for panel data. You may want to look at the Stata Longitudinal-Data/Panel-Data Reference Manual PDF included in your Stata installation and accessible from Stata's Help menu.

[Lalo Salamanca]

Sort of,

I want to regress y for year 2021 only and then x, say, for year 2019 only and another variable say z for year 2018

can i set specific year ranges for each individual component of the regression?

so if it was wide form it would look like this:

reg y_21 x_19 z_18

[William Lisowski]

reg profit21 employees19 debt20 ......

translated literally becomes

xtset firm yr
reg profit L2.employees L1.debt ...... if yr==21

where firm is the identifier for your firm and yr is your date variable.

You should definitely review the documentation I suggested in post #2. Also

help tsvarlist

[Lalo Salamanca]

I see thank you so much William, I will look at the pdf too!