dicseg segregation

[Chiara Tasselli]

Good Morning to everyone,
I'm having trouble with the dicseg command, and I can't understand why it isn't working. It keeps telling me that the variables of interest must have two groups, but I believe they already do (dummy 0-1). Where am I going wrong?


Below is what I see:

Many thanks in advance for your time

[Maarten Buis]

what happens when you type tab AREA_A female, missing ?

[Chiara Tasselli]

Maarten many thanks for your advice,
actually I checked also for the missingness of AREA_A, but the real problem was the fact thatAREA_A was a string. Now the command works and I have a further question I just plotted the graph and I would like to know to which index(es) it belongs.


Many thanks in advance for your time

dicseg AREA_Aencoded female if  AREA_A != "". , sc

--------------------------------------------------------------------------------------------------
> --
Computing Segregation of groups (female) across units (AREA_Aencoded) based on individual data

Number of units (AREA_Aencoded) =  17
Proportion group 1 (female = 0) =    0.5237
Proportion group 2 (female = 1) =    0.4763


-------------------------------------------
Segregation Measures |               female
---------------------+---------------------
       Dissimilarity |               0.0934
   Karmel-MacLachlan |               0.0466
              GE(-2) |               0.0598
              GE(-1) |               0.0460
               GE(0) |               0.0374
             GE(.10) |               0.0367
             GE(.25) |               0.0358
             GE(.50) |               0.0344
             GE(.75) |               0.0331
             GE(.90) |               0.0324
               GE(1) |               0.0319
               GE(2) |               0.0286
        Squared root |               0.0086
   Cof. of Variation |               0.2393
              A(.10) |               0.0032
              A(.25) |               0.0083
              A(.50) |               0.0171
              A(.75) |               0.0266
              A(.90) |               0.0326
                A(1) |               0.0367
                A(2) |               0.0843
                A(4) |               0.2045
  Mutual Information |               0.0122
                Gini |               0.1115
-------------------------------------------

Notes:
 - GE(alpha) = Generalized Entropy Family
 - A(epsilon) = Atkinson family
 - Mutual Information index (log base 2)

--------------------------------------------------------------------------------------------------
> --

.

[Maarten Buis]

As far as I know, there is no direct link between a segregation curve and a segregation measure like there is between the Gini coefficient and the Lorenz curve. Stephen Jenkins may know more.

[Stephen Jenkins]

I asked him, Maarten, and he said he'd forgotten!