creating a new var: Type mismatch

[Paris Rira]

Hi

I am going to classify firms into regions. I have a variable that specifies the firm's locations: "nut2_emp", When I tried to generate the region variable it keeps saying type mismatch. Is there any of specialist to assist me what is the possible problem? Thank you so much.

Cheers,

* Example generated by -dataex-. For more info, type help dataex
clear
input double nut2_emp
20
16
11
16
16
16
16
16
16
16
17
17
15
15
11
11
11
11
16
16
16
15
11
15
17
15
18
15
17
16
16
16
17
16
17
17
17
17
16
15
17
16
17
17
11
17
16
11
17
17
11
17
11
16
17
17
17
17
15
11
16
18
17
17
16
16
11
11
17
17
11
17
17
17
17
17
16
17
11
18
11
30
11
11
30
16
17
17
17
15
18
16
16
15
11
16
16
17
17
17
end
label values nut2_emp nut2_emp
label def nut2_emp 11 " Norte", modify
label def nut2_emp 15 " Algarve", modify
label def nut2_emp 16 " Centro", modify
label def nut2_emp 17 " Lisboa", modify
label def nut2_emp 18 " Alentejo", modify
label def nut2_emp 20 " Azores", modify
label def nut2_emp 30 " Madeira", modify



gen region=0
replace region=1 if nut2_emp=="Norte"
type mismatch

replace region=2 if nut2_emp=="Centro"
replace region=3 if nut2_emp=="Lisboa"
replace region=4 if nut2_emp=="Algarve"
replace region=5 if nut2_emp=="Alentejo"
replace region=6 if nut2_emp=="Azores"
replace region=7 if nut2_emp=="Madeira"
replace region=8 if nut2_emp=="Estrangeiro"

[Clyde Schechter]

There are a number of problems here.

The first is the fact that all of the labels in value label nut2_emp begin with a leading blank. That's legal, but it's also not compatible with then referring to those values without the leading blank. So you need to be consistent. It's either " Norte" or "Norte" all the time, but not mixing and matching.

Next, since nut2_emp is a numeric variable, you cannot ask Stata whether or not it is equal to a string like "Norte". You have to ask whether it corresponds to the numeric value paired with "Norte" in the value label. Finally, I notice that for regions 1 and 2 you refer to nut2_emp, but then for regions 3 and higher you are referring to a different variable nut2_est. There is no nut2_est variable in your example data, and I don't know if such a variable exists in your full data set or whether that is just typographical error. I'm assuming typographical error. In that case your code gets fixed with:

* Example generated by -dataex-. For more info, type help dataex
clear
input double nut2_emp
20
16
11
16
16
16
16
16
16
16
17
17
15
15
11
11
11
11
16
16
16
15
11
15
17
15
18
15
17
16
16
16
17
16
17
17
17
17
16
15
17
16
17
17
11
17
16
11
17
17
11
17
11
16
17
17
17
17
15
11
16
18
17
17
16
16
11
11
17
17
11
17
17
17
17
17
16
17
11
18
11
30
11
11
30
16
17
17
17
15
18
16
16
15
11
16
16
17
17
17
end
label values nut2_emp nut2_emp
label def nut2_emp 11 "Norte", modify
label def nut2_emp 15 "Algarve", modify
label def nut2_emp 16 "Centro", modify
label def nut2_emp 17 "Lisboa", modify
label def nut2_emp 18 "Alentejo", modify
label def nut2_emp 20 "Azores", modify
label def nut2_emp 30 "Madeira", modify



gen region=0
replace region=1 if nut2_emp=="Norte":nut2_emp
replace region=2 if nut2_emp=="Centro":nut2_emp
replace region=3 if nut2_emp=="Lisboa":nut2_emp
replace region=4 if nut2_emp=="Algarve":nut2_emp
replace region=5 if nut2_emp=="Alentejo":nut2_emp
replace region=6 if nut2_emp=="Azores":nut2_emp
replace region=7 if nut2_emp=="Madeira":nut2_emp
replace region=8 if nut2_emp=="Estrangeiro":nut2_emp

If you really do wish to refer to a variable nut2_est, then post back with a new -dataex- that includes both nut2_emp and nut2_est.

[Paris Rira]

Prof, thank you for getting back to me.
nut2_est was a typo, although there exist such a variable in the data which is locations at the establishment level. I already edited

. gen region=0

. replace region=1 if nut2_emp=="Norte":nut2_emp
(value label dereference "Norte":nut2_emp not found)
(0 real changes made)

. replace region=2 if nut2_emp=="Centro":nut2_emp
(value label dereference "Centro":nut2_emp not found)
(0 real changes made)

. replace region=3 if nut2_emp=="Lisboa":nut2_emp
(value label dereference "Lisboa":nut2_emp not found)
(0 real changes made)

. replace region=4 if nut2_emp=="Algarve":nut2_emp
(value label dereference "Algarve":nut2_emp not found)
(0 real changes made)

. replace region=5 if nut2_emp=="Alentejo":nut2_emp
(value label dereference "Alentejo":nut2_emp not found)
(0 real changes made)

. replace region=6 if nut2_emp=="Azores":nut2_emp
(value label dereference "Azores":nut2_emp not found)
(0 real changes made)

. replace region=7 if nut2_emp=="Madeira":nut2_emp
(value label dereference "Madeira":nut2_emp not found)
(0 real changes made)

. replace region=8 if nut2_emp=="Estrangeiro":nut2_emp
(value label dereference "Estrangeiro":nut2_emp not found)
(0 real changes made)

[Clyde Schechter]

Yes, but you also have to edit the value label nut2_emp change. Notice from your original code in #1:

label def nut2_emp 11 " Norte", modify
label def nut2_emp 15 " Algarve", modify
label def nut2_emp 16 " Centro", modify
label def nut2_emp 17 " Lisboa", modify
label def nut2_emp 18 " Alentejo", modify
label def nut2_emp 20 " Azores", modify
label def nut2_emp 30 " Madeira", modify

But from my code in #2:

label def nut2_emp 11 "Norte", modify
label def nut2_emp 15 "Algarve", modify
label def nut2_emp 16 "Centro", modify
label def nut2_emp 17 "Lisboa", modify
label def nut2_emp 18 "Alentejo", modify
label def nut2_emp 20 "Azores", modify
label def nut2_emp 30 "Madeira", modify

Notice that I have removed the leading blanks from the labels. You have to do that, too.

[Paris Rira]

You are right. First, I have to remove the leading blanks.
It solved. Really appreciated.

label def nut2_emp 11 "Norte", modify
label def nut2_emp 15 "Algarve", modify
 label def nut2_emp 16 "Centro", modify
label def nut2_emp 17 "Lisboa", modify
label def nut2_emp 18 "Alentejo", modify
label def nut2_emp 20 "Açores", modify
label def nut2_emp 30 "Madeira", modify
label def nut2_emp 90 "Estrangeiro", modify


. gen region=0

. replace region=1 if nut2_emp=="Norte":nut2_emp
(107,724 real changes made)

. replace region=2 if nut2_emp=="Centro":nut2_emp
(66,286 real changes made)

. replace region=3 if nut2_emp=="Lisboa":nut2_emp
(72,013 real changes made)

. replace region=4 if nut2_emp=="Algarve":nut2_emp
(16,278 real changes made)

. replace region=5 if nut2_emp=="Alentejo":nut2_emp
(21,010 real changes made)

. replace region=6 if nut2_emp=="Açores":nut2_emp
(5,492 real changes made)

. replace region=7 if nut2_emp=="Madeira":nut2_emp
(6,229 real changes made)

. replace region=8 if nut2_emp=="Estrangeiro":nut2_emp
(8 real changes made)

. 
end of do-file

. tab region

     region |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |    107,724       36.51       36.51
          2 |     66,286       22.47       58.98
          3 |     72,013       24.41       83.39
          4 |     16,278        5.52       88.90
          5 |     21,010        7.12       96.02
          6 |      5,492        1.86       97.89
          7 |      6,229        2.11      100.00
          8 |          8        0.00      100.00
------------+-----------------------------------
      Total |    295,040      100.00