combining*three different dataset

[Paris Rira]

Hi dear statalists,


I am going to estimate 5 different regressions with three different datasets. In order to avoid using several messy files, and make the Do-file more beautiful, I decided to combine (append) three datasets together. The question is, does it hurt the panel data this way? Because these three data have different variables, some are missing once appending horizontally.

clear
input double(year NPC_FIC) str2 nacio float(region mig_jump imm_TFP_below S_tfp_lp_sh S_tfp_lp_iv) byte per double impu_sh_origin
2010 500008310 "PT" 2 0 . . . 0 .
2010 500027353 "PT" 5 0 . . . 0 .
2010 500049799 "UA" 4 0 . . . 0 .
2010 500050215 "PT" 5 0 . . . 0 .
2010 500135017 "PT" 1 0 . . . 0 .
2010 500139105 "PT" 3 0 . . . 0 .
2010 500205042 "PT" 3 0 . . . 0 .
2010 500206942 "PT" 1 0 . . . 0 .
2010 500251226 "PT" 1 0 . . . 0 .
2010 500265754 "PT" 2 0 . . . 0 .
2010 500321035 "PT" 1 0 . . . 0 .
2010 500369375 "PT" 4 0 . . . 0 .
2010 500392001 "PT" 5 0 . . . 0 .
2010 500422489 "PT" 2 0 . . . 0 .
2010 500458395 "PT" 1 0 . . . 0 .

any ideas appreciated.
Cheers,
Paris

[Jared Greathouse]

Why not have each dataset in a frame? That way you don't need to have 3 unrelated datasets in one sheet.

[Paris Rira]

Jared,
Mainly because of cosmetic treatment and lowering the probability of making mistakes. If I had them all together I could be able to make a loop for all of them easily.