[Siluleko Mkhize]

Or rather compare the three proportions within the male category and then the female category

Maybe I didn't communicate my objective clearly. I wish to do a within-group comparison of food security, such that I can generate a P-value comparing food security within males and another P-value within females.

[Mike Lacy]

For a bivariate analysis here, I would prefer -somersd-, available at SSC, as it respects the ordered and categorical character of your outcome, and is asymmetric, distinguishing outcomes from predictors. Its syntax presumes that the first variable listed is the outcome, unlike some other Stata commands.

somersd Sex Hunger_cat if (AGE_VQ_P >=20)

[Joseph Coveney]

I wish to do a within-group comparison of food security, such that I can generate a P-value comparing food security within males and another P-value within females.

Following up on Carlo's suggestion in #5, perhaps you can use the following to construct two omnibus tests, one for men and another for women. Each two-degree-of-freedom test would be across all three categories at once within a given sex.

.ÿ
.ÿversionÿ16.0

.ÿ
.ÿclearÿ*

.ÿ
.ÿinputÿbyteÿsexÿint(count0ÿcount1ÿcount2)

ÿÿÿÿÿÿÿÿÿÿsexÿÿÿÿcount0ÿÿÿÿcount1ÿÿÿÿcount2
ÿÿ1.ÿ0ÿ1930ÿ1457ÿ2682
ÿÿ2.ÿ1ÿ1119ÿÿ960ÿ1840
ÿÿ3.ÿend

.ÿ
.ÿquietlyÿreshapeÿlongÿcount,ÿi(sex)ÿj(grp)

.ÿ
.ÿmlogitÿgrpÿi.sexÿ[fweight=count],ÿnolog

MultinomialÿlogisticÿregressionÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿÿÿÿÿ=ÿÿÿÿÿÿ9,988
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿLRÿchi2(2)ÿÿÿÿÿÿÿÿ=ÿÿÿÿÿÿ12.52
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿProbÿ>ÿchi2ÿÿÿÿÿÿÿ=ÿÿÿÿÿ0.0019
Logÿlikelihoodÿ=ÿ-10624.342ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿPseudoÿR2ÿÿÿÿÿÿÿÿÿ=ÿÿÿÿÿ0.0006

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿgrpÿ|ÿÿÿÿÿÿCoef.ÿÿÿStd.ÿErr.ÿÿÿÿÿÿzÿÿÿÿP>|z|ÿÿÿÿÿ[95%ÿConf.ÿInterval]
-------------+----------------------------------------------------------------
0ÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿ1.sexÿ|ÿÿ-.1682874ÿÿÿ.0482506ÿÿÿÿ-3.49ÿÿÿ0.000ÿÿÿÿ-.2628569ÿÿÿ-.0737178
ÿÿÿÿÿÿÿ_consÿ|ÿÿ-.3290428ÿÿÿ.0298495ÿÿÿ-11.02ÿÿÿ0.000ÿÿÿÿ-.3875467ÿÿÿ-.2705389
-------------+----------------------------------------------------------------
1ÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿ1.sexÿ|ÿÿ-.0404043ÿÿÿ.0514232ÿÿÿÿ-0.79ÿÿÿ0.432ÿÿÿÿ-.1411919ÿÿÿÿ.0603833
ÿÿÿÿÿÿÿ_consÿ|ÿÿ-.6101833ÿÿÿ.0325453ÿÿÿ-18.75ÿÿÿ0.000ÿÿÿÿ-.6739709ÿÿÿ-.5463956
-------------+----------------------------------------------------------------
2ÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿ(baseÿoutcome)
------------------------------------------------------------------------------

.ÿ
.ÿmarginsÿsex,ÿpost

AdjustedÿpredictionsÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿÿÿÿÿ=ÿÿÿÿÿÿ9,988
ModelÿVCEÿÿÿÿ:ÿOIM

1._predictÿÿÿ:ÿPr(grp==0),ÿpredict(prÿoutcome(0))
2._predictÿÿÿ:ÿPr(grp==1),ÿpredict(prÿoutcome(1))
3._predictÿÿÿ:ÿPr(grp==2),ÿpredict(prÿoutcome(2))

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿDelta-method
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿMarginÿÿÿStd.ÿErr.ÿÿÿÿÿÿzÿÿÿÿP>|z|ÿÿÿÿÿ[95%ÿConf.ÿInterval]
-------------+----------------------------------------------------------------
_predict#sexÿ|
ÿÿÿÿÿÿÿÿ1ÿ0ÿÿ|ÿÿÿ.3180096ÿÿÿ.0059779ÿÿÿÿ53.20ÿÿÿ0.000ÿÿÿÿÿÿ.306293ÿÿÿÿ.3297261
ÿÿÿÿÿÿÿÿ1ÿ1ÿÿ|ÿÿÿÿ.285532ÿÿÿ.0072149ÿÿÿÿ39.58ÿÿÿ0.000ÿÿÿÿÿ.2713911ÿÿÿÿÿ.299673
ÿÿÿÿÿÿÿÿ2ÿ0ÿÿ|ÿÿÿ.2400725ÿÿÿ.0054828ÿÿÿÿ43.79ÿÿÿ0.000ÿÿÿÿÿ.2293265ÿÿÿÿ.2508185
ÿÿÿÿÿÿÿÿ2ÿ1ÿÿ|ÿÿÿ.2449604ÿÿÿ.0068698ÿÿÿÿ35.66ÿÿÿ0.000ÿÿÿÿÿ.2314959ÿÿÿÿÿ.258425
ÿÿÿÿÿÿÿÿ3ÿ0ÿÿ|ÿÿÿ.4419179ÿÿÿ.0063747ÿÿÿÿ69.32ÿÿÿ0.000ÿÿÿÿÿ.4294237ÿÿÿÿ.4544122
ÿÿÿÿÿÿÿÿ3ÿ1ÿÿ|ÿÿÿ.4695075ÿÿÿ.0079721ÿÿÿÿ58.89ÿÿÿ0.000ÿÿÿÿÿ.4538825ÿÿÿÿ.4851326
------------------------------------------------------------------------------

.ÿ
.ÿ//ÿWomen
.ÿtestÿ1._predict#0.sexÿ=ÿ2._predict#0.sex,ÿnotest

ÿ(ÿ1)ÿÿ1bn._predict#0bn.sexÿ-ÿ2._predict#0bn.sexÿ=ÿ0

.ÿtestÿ1._predict#0.sexÿ=ÿ3._predict#0.sex,ÿaccumulate

ÿ(ÿ1)ÿÿ1bn._predict#0bn.sexÿ-ÿ2._predict#0bn.sexÿ=ÿ0
ÿ(ÿ2)ÿÿ1bn._predict#0bn.sexÿ-ÿ3._predict#0bn.sexÿ=ÿ0

ÿÿÿÿÿÿÿÿÿÿÿchi2(ÿÿ2)ÿ=ÿÿ386.28
ÿÿÿÿÿÿÿÿÿProbÿ>ÿchi2ÿ=ÿÿÿÿ0.0000

.ÿ
.ÿ//ÿMen
.ÿtestÿ1._predict#1.sexÿ=ÿ2._predict#1.sex,ÿnotest

ÿ(ÿ1)ÿÿ1bn._predict#1.sexÿ-ÿ2._predict#1.sexÿ=ÿ0

.ÿtestÿ1._predict#1.sexÿ=ÿ3._predict#1.sex,ÿaccumulate

ÿ(ÿ1)ÿÿ1bn._predict#1.sexÿ-ÿ2._predict#1.sexÿ=ÿ0
ÿ(ÿ2)ÿÿ1bn._predict#1.sexÿ-ÿ3._predict#1.sexÿ=ÿ0

ÿÿÿÿÿÿÿÿÿÿÿchi2(ÿÿ2)ÿ=ÿÿ311.09
ÿÿÿÿÿÿÿÿÿProbÿ>ÿchi2ÿ=ÿÿÿÿ0.0000

.ÿ
.ÿexit

endÿofÿdo-file


.

[Siluleko Mkhize]

Or rather compare the three proportions within the male category and then the female category

Maybe I didn't communicate my objective clearly. I wish to do a within-group comparison of food security, such that I can generate a P-value comparing food security within males and another P-value within females.

[Carlo Lazzaro]

Siluleko:
you may want something along the following lines:

g Food_security=0 in 1/47
replace Food_security=1 in 48/72
replace Food_security=2 in 73/100
replace Food_security=0 in 101/144
replace Food_security=1 in 145/168
replace Food_security=2 in 169/200
label define Food_security 0 "Food secure" 1 "At risk of hunger" 2 "Starvation"
label val Food_security Food_security
mlogit Food_security i.Gender
Iteration 0:   log likelihood = -212.81539 
Iteration 1:   log likelihood = -212.62236 
Iteration 2:   log likelihood = -212.62229 
Iteration 3:   log likelihood = -212.62229 

Multinomial logistic regression                 Number of obs     =        200
                                                LR chi2(2)        =       0.39
                                                Prob > chi2       =     0.8244
Log likelihood = -212.62229                     Pseudo R2         =     0.0009

-----------------------------------------------------------------------------------
    Food_security |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------+----------------------------------------------------------------
Food_secure       |  (base outcome)
------------------+----------------------------------------------------------------
At_risk_of_hunger |
           Gender |
          Female  |    .025136   .3545004     0.07   0.943     -.669672     .719944
            _cons |  -.6312718   .2475411    -2.55   0.011    -1.116443   -.1461001
------------------+----------------------------------------------------------------

*Between strata comparison - Female*
. test [Starvation]1.Gender=[At_risk_of_hunger]1.Gender=[Food_secure]1.Gender

 ( 1)  - [At_risk_of_hunger]1.Gender + [Starvation]1.Gender = 0
 ( 2)  - [Food_secure]1o.Gender + [Starvation]1.Gender = 0

           chi2(  2) =    0.39
         Prob > chi2 =    0.8247

 *Between strata comparison - Male*
. test [Starvation]_cons=[At_risk_of_hunger]_cons=[Food_secure]0.Gender

 ( 1)  - [At_risk_of_hunger]_cons + [Starvation]_cons = 0
 ( 2)  - [Food_secure]0b.Gender + [Starvation]_cons = 0

           chi2(  2) =    8.30
         Prob > chi2 =    0.0157