How to create a categorical variable from a list?

[Mohamud Hussein]

Hello members,

I would like to seek your suggestion please on how best I can create a categorical variable from the list of the list below, so that I can have a breakdown of a surveyed population into 4 distinct categories: IDP, nomads, urban (u) and rural (r)?

The goal is to use the categorical variable to create a dummy variable which I can use to compare some estimates for the different categories.

Thanks,
Mohamud

tab astrata

                  Analytical strata |      Freq.     Percent        Cum.
------------------------------------+-----------------------------------
                                IDP |     48,632        9.02        9.02
                            Nomadic |     53,847        9.99       19.01
                        (u):Banadir |     94,493       17.53       36.55
                         (u):Nugaal |      4,554        0.84       37.39
                     (u):Bari+Mudug |     40,523        7.52       44.91
                (u):Woqooyi_Galbeed |     20,783        3.86       48.77
     (u):Awdal+Sanaag+Sool+Togdheer |     29,108        5.40       54.17
(u):Hiraan+MiddleShabelle+Galgaduud |     62,693       11.63       65.80
       (u):Bay+Bakool+LowerShabelle |     74,231       13.77       79.58
         (r):Awdal+Togdheer+Woqooyi |      5,292        0.98       80.56
(r):Hiraan+MiddleShabelle+Galgaduud |     61,334       11.38       91.94
       (r):Bay+Bakool+LowerShabelle |     43,450        8.06      100.00
------------------------------------+-----------------------------------
                              Total |    538,940      100.00

[Carlo Lazzaro]

Muhamud:
you may want to try:

tab astrata, gen(astrata_cat)

[David Benson]

dataex astrata // data shared via -dataex-. To install: ssc install dataex
clear
input str36 astrata
"IDP"                                
"Nomadic"                            
"(u):Banadir"                        
"(u):Nugaal"                        
"(u):Bari+Mudug"                    
"(u):Woqooyi_Galbeed"                
"(u):Awdal+Sanaag+Sool+Togdheer"    
"(u):Hiraan+MiddleShabelle+Galgaduud"
"(u):Bay+Bakool+LowerShabelle"      
"(r):Awdal+Togdheer+Woqooyi"        
"(r):Hiraan+MiddleShabelle+Galgaduud"
"(r):Bay+Bakool+LowerShabelle"      
end

gen astrata_cat = "IDP" if astrata=="IDP"
replace astrata_cat = "nomads" if astrata=="Nomadic"
replace astrata_cat = "urban"  if strpos(astrata, "(u)") ==1
replace astrata_cat = "rural"  if strpos(astrata, "(r)") ==1

. list, noobs abbrev(12)

  +---------------------------------------------------+
  |                             astrata   astrata_cat |
  |---------------------------------------------------|
  |                                 IDP           IDP |
  |                             Nomadic        nomads |
  |                         (u):Banadir         urban |
  |                          (u):Nugaal         urban |
  |                      (u):Bari+Mudug         urban |
  |---------------------------------------------------|
  |                 (u):Woqooyi_Galbeed         urban |
  |      (u):Awdal+Sanaag+Sool+Togdheer         urban |
  | (u):Hiraan+MiddleShabelle+Galgaduud         urban |
  |        (u):Bay+Bakool+LowerShabelle         urban |
  |          (r):Awdal+Togdheer+Woqooyi         rural |
  |---------------------------------------------------|
  | (r):Hiraan+MiddleShabelle+Galgaduud         rural |
  |        (r):Bay+Bakool+LowerShabelle         rural |
  +---------------------------------------------------+


tabulate astrata_cat, gen(d)

astrata_cat |      Freq.     Percent        Cum.
------------+-----------------------------------
        IDP |          1        8.33        8.33
     nomads |          1        8.33       16.67
      rural |          3       25.00       41.67
      urban |          7       58.33      100.00
------------+-----------------------------------
      Total |         12      100.00



sort astrata_cat astrata
list, noobs abbrev(12) sepby( astrata_cat)

  +-----------------------------------------------------------------------+
  |                             astrata   astrata_cat   d1   d2   d3   d4 |
  |-----------------------------------------------------------------------|
  |                                 IDP           IDP    1    0    0    0 |
  |-----------------------------------------------------------------------|
  |                             Nomadic        nomads    0    1    0    0 |
  |-----------------------------------------------------------------------|
  |          (r):Awdal+Togdheer+Woqooyi         rural    0    0    1    0 |
  |        (r):Bay+Bakool+LowerShabelle         rural    0    0    1    0 |
  | (r):Hiraan+MiddleShabelle+Galgaduud         rural    0    0    1    0 |
  |-----------------------------------------------------------------------|
  |      (u):Awdal+Sanaag+Sool+Togdheer         urban    0    0    0    1 |
  |                         (u):Banadir         urban    0    0    0    1 |
  |                      (u):Bari+Mudug         urban    0    0    0    1 |
  |        (u):Bay+Bakool+LowerShabelle         urban    0    0    0    1 |
  | (u):Hiraan+MiddleShabelle+Galgaduud         urban    0    0    0    1 |
  |                          (u):Nugaal         urban    0    0    0    1 |
  |                 (u):Woqooyi_Galbeed         urban    0    0    0    1 |
  +-----------------------------------------------------------------------+

[Joseph Coveney]

If you want to collapse the rural and urban subcategories, then you can try something like the following.

generate byte category = .
local index 0
foreach item in ID No "(u" "(r" {
    replace category = `index' if substr(strtrim(astrata), 1, 2) == "`item'"
    local ++index
}

Recently, Nick Cox showed an elegant approach using the : word extended macro function (or something similar—I can't remember the details). You might also want to search for that on the forum.

[Joseph Coveney]

Oh, and one final touch: add value labels, if desired.

label define Categories 0 IDP 1 Nomadic 2 Urban 3 Rural
label values category Categories
<Stata estimation command> quantitative_attribute i.category

[Mohamud Hussein]

Thanks Carlo, David and Joseph.