ANNEXE
Les estimations, les différents tests faites sur le
modèle et les commandes sur stata. - Test d'Independence entre la
pauvreté et le milieu de résidence
tabulate p0 milieu, chi2
Incidence |
de la | milieu de résidence
|
pauvret~ |
+
|
urbain
|
|
rural |
+
|
Total
|
|
Non pauvre |
|
2,053
|
|
1,423 |
|
3,476
|
|
Pauvre |
|
1,010
|
|
1,467 |
|
2,477
|
|
+
|
|
|
+
|
|
|
Total |
|
3,063
|
|
2,890 |
|
5,953
|
|
Pearson
|
chi2(1)
|
=
|
193.6343
|
Pr = 0.000
|
-Test d'indépendance entre la pauvreté et la
région de résidence du chef de ménage
tabulate p0 a01 ,chi2
|
Incidence |
de la |
pauvret~ | Dakar
|
Ziguincho
|
Diourbel
|
Saint-Lou
|
A1_R~gion Tambacoun
|
Kaolack
|
Thi~s
|
Louga
|
|
Fatick | Total
|
|
|
|
|
|
|
|
|
+
|
|
|
|
|
|
|
|
|
+
|
|
|
|
|
|
|
|
|
Non pauvre | 491
|
151
|
278
|
288
|
283
|
222
|
317
|
344
|
|
195 | 3,476
Pauvre | 108
|
235
|
160
|
125
|
231
|
225
|
132
|
80
|
|
207 | 2,477
|
|
|
|
|
|
|
|
|
+
|
|
|
|
|
|
|
|
|
+
Total | 599
|
386
|
438
|
413
|
514
|
447
|
449
|
424
|
|
402 | 5,953
|
|
|
|
|
|
|
|
Incidence |
de la | A1_R~gion
pauvret~ | Kolda Matam Kaffrine K~dougou S~dhiou | Total
|
+
|
|
|
|
|
+
|
|
|
Non pauvre |
|
167
|
217
|
233
|
115
|
175 |
|
3,476
|
|
Pauvre |
|
221
|
142
|
202
|
171
|
238 |
|
2,477
|
|
+
|
|
|
|
|
+
|
|
|
Total |
|
388
|
359
|
435
|
286
|
413 |
|
5,953
|
Pearson chi2(13) = 497.8753 Pr = 0.000
-Test d'indépendance entre la pauvreté et le sexe
du chef de ménage
. tabulate p0 sexe_cm ,chi2
Incidence |
de la | B2_Sexe
|
pauvret~ |
+
|
Masculin
|
F~minin |
+
|
Total
|
|
Non pauvre |
|
2,437
|
1,039 |
|
3,476
|
|
Pauvre |
|
2,024
|
453 |
|
2,477
|
|
+
|
|
+
|
|
|
Total |
|
4,461
|
1,492 |
|
5,953
|
Pearson chi2(1) = 103.6665 Pr = 0.000
-Test d'indépendance entre la pauvreté et la
situation matrimoniale du chef de ménage
tabulate p0 etatmat_cm ,chi2
59
Incidence |
de la | Etat matrimonial
|
pauvrete |
+
|
Marie (e)
|
Marie (e)
|
Celibatai
|
Veuf (ve)
|
Divorce
|
( Concubina |
+
|
Total
|
|
Non pauvre |
|
1,832
|
1,028
|
121
|
386
|
95
|
0 |
|
3,462
|
|
Pauvre |
|
1,276
|
843
|
52
|
248
|
42
|
2 |
|
2,463
|
|
+
|
|
|
|
|
|
+
|
|
|
Total |
|
3,108
|
1,871
|
173
|
634
|
137
|
2 |
|
5,925
|
Pearson chi2(5) = 30.2394 Pr = 0.000
-Test d'indépendance entre la pauvreté et le groupe
d'âge du chef de ménage.
tabulate p0 groupage1_cm ,chi2
|
Incidence |
de la | Groupe d'ege
pauvrete | Moins de 35 - 60 a Plus de
+
|
6 |
+
|
Total
|
|
Non pauvre | 462
|
|
2,175
|
822
|
|
|
3,459
|
|
Pauvre | 246
|
|
1,497
|
719
|
|
|
2,462
|
|
+
|
|
|
|
+
|
|
|
Total | 708
|
|
3,672
|
1,541
|
|
|
5,921
|
|
Pearson chi2(2)
|
=
|
30.9686
|
Pr =
|
0.000
|
|
-Test d'indépendance entre la pauvreté et
l'occupation du chef de ménage.
. tabulate p0 occupation_cm ,chi2
|
Incidence |
de la | Occupation
pauvrete | Occupe (1 Chomeur (
+
|
Inactif ( |
+
|
Total
|
|
Non pauvre |
|
2,360
|
|
110
|
992 |
|
3,462
|
|
Pauvre |
|
1,681
|
|
74
|
708 |
|
2,463
|
|
+
|
|
|
|
+
|
|
|
Total |
|
4,041
|
|
184
|
1,700 |
|
5,925
|
|
Pearson
|
chi2(2)
|
=
|
0.1441
|
Pr = 0.930
|
|
-Test d'indépendance entre la pauvreté et le niveau
d'instruction du chef de ménage.
tabulate p0 nivinst_cm ,chi2
Incidence |
de la | Niveau d'instruction
|
pauvrete |
|
Sans inst
|
Primaire
|
Moyen
|
Secondair
|
Superieur |
|
Total
|
|
+
|
|
|
|
|
+
|
|
|
Non pauvre |
|
2,147
|
563
|
325
|
216
|
190 |
|
3,441
|
|
Pauvre |
|
1,938
|
329
|
115
|
59
|
16 |
|
2,457
|
|
+
|
|
|
|
|
+
|
|
|
Total |
|
4,085
|
892
|
440
|
275
|
206 |
|
5,898
|
Pearson chi2(4) = 251.7500 Pr = 0.000
- Estimation du modèle logit (coefficients).
|
logit pauvre taille taille2 feminin rural sansinstrucyion
primaire moyen secondaire ziginchor diourbel saintlouis tambacounda
k
> aolack thies louga fatick kolda matam kafrine
kedougou sedhiou polygame C~libatair veuf divorcer moinde35ans plusde603
chomeur
> inactif
|
|
Iteration 0:
|
log
|
likelihood
|
=
|
-4003.2212
|
|
|
|
|
Iteration 1:
|
log
|
likelihood
|
=
|
-3178.0567
|
|
|
|
|
Iteration 2:
|
log
|
likelihood
|
=
|
-3163.6587
|
|
|
|
|
Iteration 3:
|
log
|
likelihood
|
=
|
-3163.5119
|
|
|
|
|
Iteration 4:
|
log
|
likelihood
|
=
|
-3163.5116
|
|
|
|
|
Logistic regression
|
|
|
|
Number of obs
|
=
|
5,894
|
|
|
|
|
|
LR chi2(29)
|
=
|
1679.42
|
|
|
|
|
|
Prob > chi2
|
=
|
0.0000
|
60
Log likelihood = -3163.5116
|
|
Pseudo R2
|
=
|
0.2098
|
|
pauvre | Coef.
|
Std. Err.
|
z
|
P>|z|
|
[95% Conf.
|
Interval]
|
|
taille
+
|
| .2436793
|
.0147072
|
16.57
|
0.000
|
.2148537
|
.2725049
|
|
taille2
|
| -.0032127
|
.0004796
|
-6.70
|
0.000
|
-.0041526
|
-.0022728
|
|
feminin
|
| -.487462
|
.094513
|
-5.16
|
0.000
|
-.672704
|
-.30222
|
|
rural
|
| .1511685
|
.06952
|
2.17
|
0.030
|
.0149118
|
.2874251
|
|
sansinstrucyion
|
| 2.301416
|
.2825635
|
8.14
|
0.000
|
1.747602
|
2.85523
|
|
primaire
|
| 1.870197
|
.2885087
|
6.48
|
0.000
|
1.30473
|
2.435663
|
|
moyen
|
| 1.417742
|
.3032585
|
4.68
|
0.000
|
.8233661
|
2.012118
|
|
secondaire
|
| 1.051266
|
.3214845
|
3.27
|
0.001
|
.4211678
|
1.681364
|
|
ziginchor
|
| 2.207772
|
.1697219
|
13.01
|
0.000
|
1.875123
|
2.540421
|
|
diourbel
|
| .5327335
|
.1640757
|
3.25
|
0.001
|
.211151
|
.8543161
|
|
saintlouis
|
| .2373773
|
.1698213
|
1.40
|
0.162
|
-.0954663
|
.570221
|
|
tambacounda
|
| .7149973
|
.160335
|
4.46
|
0.000
|
.4007466
|
1.029248
|
|
kaolack
|
| 1.197388
|
.1594455
|
7.51
|
0.000
|
.8848802
|
1.509895
|
|
thies
|
| .1029906
|
.1670141
|
0.62
|
0.537
|
-.224351
|
.4303321
|
|
louga
|
| -.391457
|
.1818801
|
-2.15
|
0.031
|
-.7479354
|
-.0349785
|
|
fatick
|
| 1.393147
|
.1658009
|
8.40
|
0.000
|
1.068183
|
1.718111
|
|
kolda
|
| 1.512746
|
.1703309
|
8.88
|
0.000
|
1.178904
|
1.846588
|
|
matam
|
| .6441411
|
.171304
|
3.76
|
0.000
|
.3083915
|
.9798908
|
|
kafrine
|
| .9448465
|
.1652939
|
5.72
|
0.000
|
.6208763
|
1.268817
|
|
kedougou
|
| 1.762357
|
.1819696
|
9.68
|
0.000
|
1.405703
|
2.119011
|
|
sedhiou
|
| 1.464034
|
.1679942
|
8.71
|
0.000
|
1.134772
|
1.793297
|
|
polygame
|
| -.3408496
|
.074214
|
-4.59
|
0.000
|
-.4863064
|
-.1953928
|
|
C~libatair
|
| .5988161
|
.2099442
|
2.85
|
0.004
|
.187333
|
1.010299
|
|
veuf
|
| .2868266
|
.1258383
|
2.28
|
0.023
|
.0401882
|
.5334651
|
|
divorcer
|
| .1978437
|
.2245417
|
0.88
|
0.378
|
-.2422499
|
.6379373
|
|
moinde35ans
|
| .1168014
|
.1039288
|
1.12
|
0.261
|
-.0868954
|
.3204981
|
|
plusde603
|
| -.0388947
|
.0797217
|
-0.49
|
0.626
|
-.1951463
|
.1173569
|
|
chomeur
|
| .5128739
|
.1799739
|
2.85
|
0.004
|
.1601316
|
.8656162
|
|
inactif
|
| .0231097
|
.0784995
|
0.29
|
0.768
|
-.1307464
|
.1769658
|
|
_cons | -5.11088
|
.3154629
|
-16.20
|
0.000
|
-5.729176
|
-4.492584
|
-Estimation du modèle logit (odds ratio)
|
logistic pauvre taille taille2 feminin rural
sansinstrucyion primaire moyen secondaire ziginchor diourbel saintlouis
tambacound
> a kaolack thies louga fatick kolda matam kafrine
kedougou sedhiou polygame C~libatair veuf divorcer moinde35ans plusde603
chome
> ur inactif
|
|
Logistic regression
|
|
|
Number of obs
|
=
|
5,894
|
|
|
|
LR chi2(29)
|
=
|
1679.42
|
|
|
|
Prob > chi2
|
=
|
0.0000
|
|
Log likelihood = -3163.5116
|
|
Pseudo R2
|
=
|
0.2098
|
|
pauvre | Odds Ratio
|
Std. Err.
|
z
|
P>|z|
|
[95% Conf.
|
Interval]
|
|
taille
+
|
|
|
1.275935
|
.0187654
|
16.57
|
0.000
|
1.239681
|
1.31325
|
|
taille2
|
|
|
.9967925
|
.000478
|
-6.70
|
0.000
|
.995856
|
.9977298
|
|
feminin
|
|
|
.6141832
|
.0580483
|
-5.16
|
0.000
|
.5103268
|
.7391755
|
|
rural
|
|
|
1.163193
|
.0808651
|
2.17
|
0.030
|
1.015024
|
1.332991
|
|
sansinstrucyion
|
|
|
9.988317
|
2.822334
|
8.14
|
0.000
|
5.740819
|
17.37844
|
|
primaire
|
|
|
6.489573
|
1.872298
|
6.48
|
0.000
|
3.686694
|
11.42339
|
|
moyen
|
|
|
4.127789
|
1.251787
|
4.68
|
0.000
|
2.278155
|
7.479139
|
|
secondaire
|
|
|
2.861271
|
.9198542
|
3.27
|
0.001
|
1.52374
|
5.372879
|
|
ziginchor
|
|
|
9.09543
|
1.543694
|
13.01
|
0.000
|
6.521623
|
12.68501
|
|
diourbel
|
|
|
1.703583
|
.2795166
|
3.25
|
0.001
|
1.235099
|
2.349767
|
|
saintlouis
|
|
|
1.267919
|
.2153197
|
1.40
|
0.162
|
.908949
|
1.768658
|
|
tambacounda
|
|
|
2.044181
|
.3277538
|
4.46
|
0.000
|
1.492939
|
2.798961
|
|
kaolack
|
|
|
3.311455
|
.5279965
|
7.51
|
0.000
|
2.422694
|
4.526255
|
|
thies
|
|
|
1.108481
|
.1851319
|
0.62
|
0.537
|
.7990346
|
1.537768
|
|
louga
|
|
|
.6760711
|
.1229639
|
-2.15
|
0.031
|
.4733428
|
.9656261
|
|
fatick
|
|
|
4.027504
|
.6677637
|
8.40
|
0.000
|
2.910087
|
5.573987
|
|
kolda
|
|
|
4.539179
|
.7731623
|
8.88
|
0.000
|
3.250809
|
6.33816
|
|
matam
|
|
|
1.904351
|
.3262229
|
3.76
|
0.000
|
1.361234
|
2.664165
|
|
kafrine
|
|
|
2.572418
|
.4252052
|
5.72
|
0.000
|
1.860558
|
3.556641
|
|
kedougou
|
|
|
5.826152
|
1.060182
|
9.68
|
0.000
|
4.078393
|
8.322899
|
|
sedhiou
|
|
|
4.323367
|
.7263007
|
8.71
|
0.000
|
3.110464
|
6.009233
|
|
polygame
|
|
|
.7111659
|
.0527785
|
-4.59
|
0.000
|
.6148934
|
.8225115
|
|
C~libatair
|
|
|
1.819963
|
.3820906
|
2.85
|
0.004
|
1.206029
|
2.746422
|
|
veuf
|
|
|
1.332193
|
.1676409
|
2.28
|
0.023
|
1.041007
|
1.70483
|
61
divorcer | 1.218772 .2736651 0.88 0.378 .78486 1.892573
moinde35ans | 1.123896 .1168052 1.12 0.261 .916773 1.377814
plusde603 | .961852 .0766804 -0.49 0.626 .8227143 1.124521
chomeur | 1.670084 .3005715 2.85 0.004 1.173665 2.37647
inactif | 1.023379 .0803347 0.29 0.768 .8774403 1.19359
_cons | .0060308 .0019025 -16.20 0.000 .0032498 .0111917
Note: _cons estimates baseline odds.
-Test d'ajustement du modèle de hosmer-lomeshov.
estat gof
Logistic model for pauvre, goodness-of-fit test
number of observations = 5894
number of covariate patterns = 4430
Pearson chi2(4400) = 4466.68
Prob > chi2 = 0.2375
-Test de prédiction correcte du modèle
lstat
Logistic model for pauvre
True
Classified | D ~D | Total
+ +
+ | 1538 702 | 2240
- | 918 2736 | 3654
+ +
Total | 2456 3438 | 5894
Classified + if predicted Pr(D) >= .5 True D defined as pauvre
!= 0
Sensitivity Pr( +| D) 62.62%
Specificity Pr( -|~D) 79.58%
Positive predictive value Pr( D| +) 68.66%
Negative predictive value Pr(~D| -) 74.88%
False + rate for true ~D Pr( +|~D) 20.42%
False - rate for true D Pr( -| D) 37.38%
False + rate for classified + Pr(~D| +) 31.34%
False - rate for classified - Pr( D| -) 25.12%
Correctly classified 72.51%
- Test du pouvoir discriminant du modèle(ROC)
lroc
Logistic model for pauvre
number of observations = 5894
area under ROC curve = 0.7958
0.00 0.25 0.50 0.75 1.00
1 - Specificity
Area under ROC curve = 0.7958
62
| 
