# 28 de octubre 2013
load("~/Dropbox/odontologia/maestria_anunziatta/BancoUnidoBrasUru/BRASURU(28102013).RData")
library(survey)
## Attaching package: 'survey'
##
## The following object(s) are masked from 'package:graphics':
##
## dotchart
levels(disenio_urubra$variables$prevoms)
## [1] "1-SinCavi" "2-conCavi"
table(disenio_urubra$variables$cpodoms, disenio_urubra$variables$prevoms)
##
## 1-SinCavi 2-conCavi
## 0 1105 0
## 1 0 495
## 2 0 404
## 3 0 259
## 4 0 217
## 5 0 108
## 6 0 38
## 7 0 22
## 8 0 14
## 9 0 12
## 10 0 3
## 11 0 3
## 14 0 1
## 15 0 1
REGRESIÓN LOGÍSTICA: 1.País 2.Escomaerecat23 3. Freqescavacat
round(svymean(~prevoms, disenio_urubra, deff = TRUE) * 100, 2)
## mean SE DEff
## prevoms1-SinCavi 41.130 0.018 3.8
## prevoms2-conCavi 58.870 0.018 3.8
round(confint(svymean(~prevoms, disenio_urubra, deff = TRUE) * 100), 2)
## 2.5 % 97.5 %
## prevoms1-SinCavi 41.10 41.17
## prevoms2-conCavi 58.83 58.90
round(svymean(~cpodoms, disenio_urubra, deff = TRUE), 2) * 1
## mean SE DEff
## cpodoms 1.5300 0.0758 4.74
round(confint(svymean(~cpodoms, disenio_urubra, deff = TRUE)), 2)
## 2.5 % 97.5 %
## cpodoms 1.39 1.68
round(svyby(~prevoms, ~pais, disenio_urubra, svymean, deff = TRUE)[, 2:4] *
100, 2)
## prevoms1-SinCavi prevoms2-conCavi se.prevoms1-SinCavi
## Bra 44.56 55.44 2.35
## Uru 38.54 61.46 2.50
round(confint(svyby(~prevoms, ~pais, disenio_urubra, svymean, deff = TRUE)) *
100, 2)
## 2.5 % 97.5 %
## Bra:prevoms1-SinCavi 39.95 49.17
## Uru:prevoms1-SinCavi 33.64 43.43
## Bra:prevoms2-conCavi 50.83 60.05
## Uru:prevoms2-conCavi 56.57 66.36
round(svyby(~cpodoms, ~pais, disenio_urubra, svymean, deff = TRUE)[, 2:4] *
1, 2)
## cpodoms se DEff.cpodoms
## Bra 1.39 0.09 4.10
## Uru 1.64 0.11 4.05
round(confint(svyby(~cpodoms, ~pais, disenio_urubra, svymean, deff = TRUE)) *
1, 2)
## 2.5 % 97.5 %
## Bra 1.22 1.57
## Uru 1.42 1.86
modelo2multi.bin <- svyglm(prevoms ~ pais + escolmaerecat23cat + freqescovacat,
disenio_urubra, family = quasibinomial())
summary(modelo2multi.bin)
##
## Call:
## svyglm(formula = prevoms ~ pais + escolmaerecat23cat + freqescovacat,
## disenio_urubra, family = quasibinomial())
##
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec,
## 1), data = brasuru, nest = TRUE)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.184 0.196 -0.94 0.35010
## paisUru 0.493 0.125 3.95 0.00017
## escolmaerecat23cat2-high school 0.521 0.158 3.29 0.00153
## escolmaerecat23cat3-elementary school 1.049 0.179 5.87 1.1e-07
## freqescovacat2-veces al d\xeda -0.284 0.134 -2.12 0.03749
## freqescovacat3-veces al d\xeda -0.457 0.151 -3.03 0.00332
##
## (Intercept)
## paisUru ***
## escolmaerecat23cat2-high school **
## escolmaerecat23cat3-elementary school ***
## freqescovacat2-veces al d\xeda *
## freqescovacat3-veces al d\xeda **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 1.002)
##
## Number of Fisher Scoring iterations: 4
anova(modelo2multi.bin)
## Anova table: (Rao-Scott LRT)
## svyglm(formula = prevoms ~ pais, disenio_urubra, family = quasibinomial())
## stats DEff df ddf p
## pais 9.85 3.22 1.00 80 0.086 .
## escolmaerecat23cat 175.44 1.69 2.00 78 1.4e-12 ***
## freqescovacat 66.45 1.72 2.00 76 7.7e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
confint(modelo2multi.bin)
## 2.5 % 97.5 %
## (Intercept) -0.5689 0.20005
## paisUru 0.2485 0.73689
## escolmaerecat23cat2-high school 0.2104 0.83131
## escolmaerecat23cat3-elementary school 0.6982 1.39886
## freqescovacat2-veces al d\xeda -0.5478 -0.02116
## freqescovacat3-veces al d\xeda -0.7525 -0.16159
exp(modelo2multi.bin$coefficients)
## (Intercept)
## 0.8316
## paisUru
## 1.6367
## escolmaerecat23cat2-high school
## 1.6835
## escolmaerecat23cat3-elementary school
## 2.8534
## freqescovacat2-veces al d\xeda
## 0.7524
## freqescovacat3-veces al d\xeda
## 0.6332
modelo3multi.bin <- svyglm(prevoms ~ pais + escolmae13cat + freqescovacat, disenio_urubra,
family = quasibinomial())
summary(modelo3multi.bin)
##
## Call:
## svyglm(formula = prevoms ~ pais + escolmae13cat + freqescovacat,
## disenio_urubra, family = quasibinomial())
##
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec,
## 1), data = brasuru, nest = TRUE)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.325 0.194 -1.67 0.0985 .
## paisUru 0.342 0.130 2.62 0.0105 *
## escolmae13cat2-high school 0.537 0.243 2.21 0.0299 *
## escolmae13cat3-Elementary School 1.134 0.174 6.52 7e-09 ***
## freqescovacat2-veces al d\xeda -0.283 0.136 -2.08 0.0412 *
## freqescovacat3-veces al d\xeda -0.459 0.148 -3.09 0.0028 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 1.004)
##
## Number of Fisher Scoring iterations: 4
confint(modelo3multi.bin)
## 2.5 % 97.5 %
## (Intercept) -0.70653 0.05582
## paisUru 0.08656 0.59768
## escolmae13cat2-high school 0.06140 1.01323
## escolmae13cat3-Elementary School 0.79289 1.47466
## freqescovacat2-veces al d\xeda -0.55025 -0.01595
## freqescovacat3-veces al d\xeda -0.74926 -0.16815
exp(modelo3multi.bin$coefficients)
## (Intercept) paisUru
## 0.7223 1.4079
## escolmae13cat2-high school escolmae13cat3-Elementary School
## 1.7114 3.1074
## freqescovacat2-veces al d\xeda freqescovacat3-veces al d\xeda
## 0.7534 0.6321
anova(modelo3multi.bin)
## Anova table: (Rao-Scott LRT)
## svyglm(formula = prevoms ~ pais, disenio_urubra, family = quasibinomial())
## stats DEff df ddf p
## pais 9.85 3.22 1.00 80 0.086 .
## escolmae13cat 197.91 2.51 2.00 78 5.1e-11 ***
## freqescovacat 63.94 1.67 2.00 76 5.9e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
modelo4multi.bin <- svyglm(prevoms ~ pais + escolmaerecat23cat + freqescovacat +
tipoesc, disenio_urubra, family = quasibinomial())
summary(modelo4multi.bin)
##
## Call:
## svyglm(formula = prevoms ~ pais + escolmaerecat23cat + freqescovacat +
## tipoesc, disenio_urubra, family = quasibinomial())
##
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec,
## 1), data = brasuru, nest = TRUE)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.356 0.195 -1.83 0.07193
## paisUru 0.485 0.120 4.05 0.00012
## escolmaerecat23cat2-high school 0.391 0.190 2.06 0.04322
## escolmaerecat23cat3-elementary school 0.841 0.218 3.86 0.00024
## freqescovacat2-veces al d\xeda -0.248 0.133 -1.87 0.06542
## freqescovacat3-veces al d\xeda -0.443 0.153 -2.89 0.00508
## tipoesc2-Publica 0.379 0.158 2.40 0.01907
##
## (Intercept) .
## paisUru ***
## escolmaerecat23cat2-high school *
## escolmaerecat23cat3-elementary school ***
## freqescovacat2-veces al d\xeda .
## freqescovacat3-veces al d\xeda **
## tipoesc2-Publica *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 1.002)
##
## Number of Fisher Scoring iterations: 4
anova(modelo4multi.bin)
## Anova table: (Rao-Scott LRT)
## svyglm(formula = prevoms ~ pais, disenio_urubra, family = quasibinomial())
## stats DEff df ddf p
## pais 9.85 3.22 1.00 80 0.086 .
## escolmaerecat23cat 175.44 1.69 2.00 78 1.4e-12 ***
## freqescovacat 66.45 1.72 2.00 76 7.7e-07 ***
## tipoesc 13.00 2.27 1.00 75 0.020 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
confint(modelo4multi.bin)
## 2.5 % 97.5 %
## (Intercept) -0.73739 0.02623
## paisUru 0.25046 0.71923
## escolmaerecat23cat2-high school 0.01833 0.76323
## escolmaerecat23cat3-elementary school 0.41380 1.26827
## freqescovacat2-veces al d\xeda -0.50842 0.01197
## freqescovacat3-veces al d\xeda -0.74359 -0.14222
## tipoesc2-Publica 0.06901 0.68938
exp(modelo4multi.bin$coefficients)
## (Intercept)
## 0.7008
## paisUru
## 1.6239
## escolmaerecat23cat2-high school
## 1.4781
## escolmaerecat23cat3-elementary school
## 2.3188
## freqescovacat2-veces al d\xeda
## 0.7802
## freqescovacat3-veces al d\xeda
## 0.6422
## tipoesc2-Publica
## 1.4611
modelo5multi.bin <- svyglm(prevoms ~ pais + escolmae13cat + freqescovacat +
tipoesc, disenio_urubra, family = quasibinomial())
summary(modelo5multi.bin)
##
## Call:
## svyglm(formula = prevoms ~ pais + escolmae13cat + freqescovacat +
## tipoesc, disenio_urubra, family = quasibinomial())
##
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec,
## 1), data = brasuru, nest = TRUE)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.444 0.209 -2.12 0.0372 *
## paisUru 0.356 0.124 2.87 0.0054 **
## escolmae13cat2-high school 0.456 0.247 1.84 0.0691 .
## escolmae13cat3-Elementary School 0.967 0.214 4.52 2.2e-05 ***
## freqescovacat2-veces al d\xeda -0.257 0.136 -1.88 0.0633 .
## freqescovacat3-veces al d\xeda -0.449 0.150 -2.99 0.0038 **
## tipoesc2-Publica 0.290 0.153 1.90 0.0615 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 1.004)
##
## Number of Fisher Scoring iterations: 4
confint(modelo5multi.bin)
## 2.5 % 97.5 %
## (Intercept) -0.853187 -0.03383
## paisUru 0.112518 0.59856
## escolmae13cat2-high school -0.028698 0.94041
## escolmae13cat3-Elementary School 0.547868 1.38578
## freqescovacat2-veces al d\xeda -0.524423 0.01025
## freqescovacat3-veces al d\xeda -0.743549 -0.15447
## tipoesc2-Publica -0.009447 0.58904
exp(modelo5multi.bin$coefficients)
## (Intercept) paisUru
## 0.6418 1.4270
## escolmae13cat2-high school escolmae13cat3-Elementary School
## 1.5775 2.6296
## freqescovacat2-veces al d\xeda freqescovacat3-veces al d\xeda
## 0.7733 0.6383
## tipoesc2-Publica
## 1.3362
anova(modelo5multi.bin)
## Anova table: (Rao-Scott LRT)
## svyglm(formula = prevoms ~ pais, disenio_urubra, family = quasibinomial())
## stats DEff df ddf p
## pais 9.85 3.22 1.00 80 0.086 .
## escolmae13cat 197.91 2.51 2.00 78 5.1e-11 ***
## freqescovacat 63.94 1.67 2.00 76 5.9e-07 ***
## tipoesc 7.20 2.01 1.00 75 0.064 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
REGRESIÓN POISSON: 1.País 2.Escomaerecat23 3.Tipoescola 4.Freqescavacat
modelo4multi.poi <- svyglm(cpodoms ~ pais + escolmaerecat23cat + freqescovacat +
tipoesc, disenio_urubra, family = quasipoisson())
summary(modelo4multi.poi)
##
## Call:
## svyglm(formula = cpodoms ~ pais + escolmaerecat23cat + freqescovacat +
## tipoesc, disenio_urubra, family = quasipoisson())
##
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec,
## 1), data = brasuru, nest = TRUE)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.2477 0.1223 -2.03 0.0463
## paisUru 0.2824 0.0653 4.32 4.7e-05
## escolmaerecat23cat2-high school 0.2291 0.1488 1.54 0.1278
## escolmaerecat23cat3-elementary school 0.4116 0.1721 2.39 0.0193
## freqescovacat2-veces al d\xeda -0.0715 0.0809 -0.88 0.3795
## freqescovacat3-veces al d\xeda -0.2334 0.0748 -3.12 0.0026
## tipoesc2-Publica 0.4618 0.1075 4.30 5.2e-05
##
## (Intercept) *
## paisUru ***
## escolmaerecat23cat2-high school
## escolmaerecat23cat3-elementary school *
## freqescovacat2-veces al d\xeda
## freqescovacat3-veces al d\xeda **
## tipoesc2-Publica ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasipoisson family taken to be 2.068)
##
## Number of Fisher Scoring iterations: 6
anova(modelo4multi.poi)
## Anova table: (Rao-Scott LRT)
## svyglm(formula = cpodoms ~ pais, disenio_urubra, family = quasipoisson())
## stats DEff df ddf p
## pais 25.8 8.36 1.00 80 0.085 .
## escolmaerecat23cat 353.4 5.69 2.00 78 2.8e-08 ***
## freqescovacat 143.6 4.64 2.00 76 2.4e-05 ***
## tipoesc 96.0 4.85 1.00 75 3.2e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
confint(modelo4multi.poi)
## 2.5 % 97.5 %
## (Intercept) -0.48740 -0.008086
## paisUru 0.15427 0.410436
## escolmaerecat23cat2-high school -0.06253 0.520721
## escolmaerecat23cat3-elementary school 0.07430 0.748827
## freqescovacat2-veces al d\xeda -0.23011 0.087048
## freqescovacat3-veces al d\xeda -0.38004 -0.086664
## tipoesc2-Publica 0.25106 0.672458
exp(modelo4multi.poi$coefficients)
## (Intercept)
## 0.7806
## paisUru
## 1.3262
## escolmaerecat23cat2-high school
## 1.2575
## escolmaerecat23cat3-elementary school
## 1.5092
## freqescovacat2-veces al d\xeda
## 0.9310
## freqescovacat3-veces al d\xeda
## 0.7919
## tipoesc2-Publica
## 1.5869
modelo5multi.poi <- svyglm(cpodoms ~ pais + escolmae13cat + freqescovacat +
tipoesc, disenio_urubra, family = quasipoisson())
summary(modelo5multi.poi)
##
## Call:
## svyglm(formula = cpodoms ~ pais + escolmae13cat + freqescovacat +
## tipoesc, disenio_urubra, family = quasipoisson())
##
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec,
## 1), data = brasuru, nest = TRUE)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.3987 0.2015 -1.98 0.05149 .
## paisUru 0.2240 0.0703 3.19 0.00209 **
## escolmae13cat2-high school 0.3838 0.2309 1.66 0.10066
## escolmae13cat3-Elementary School 0.6248 0.1782 3.51 0.00077 ***
## freqescovacat2-veces al d\xeda -0.0724 0.0821 -0.88 0.38070
## freqescovacat3-veces al d\xeda -0.2310 0.0703 -3.29 0.00154 **
## tipoesc2-Publica 0.3873 0.0928 4.17 8e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasipoisson family taken to be 2.04)
##
## Number of Fisher Scoring iterations: 6
confint(modelo5multi.poi)
## 2.5 % 97.5 %
## (Intercept) -0.79350 -0.003829
## paisUru 0.08631 0.361762
## escolmae13cat2-high school -0.06877 0.836426
## escolmae13cat3-Elementary School 0.27551 0.974056
## freqescovacat2-veces al d\xeda -0.23339 0.088549
## freqescovacat3-veces al d\xeda -0.36871 -0.093311
## tipoesc2-Publica 0.20541 0.569286
exp(modelo5multi.poi$coefficients)
## (Intercept) paisUru
## 0.6712 1.2511
## escolmae13cat2-high school escolmae13cat3-Elementary School
## 1.4679 1.8678
## freqescovacat2-veces al d\xeda freqescovacat3-veces al d\xeda
## 0.9301 0.7937
## tipoesc2-Publica
## 1.4731
anova(modelo5multi.poi)
## Anova table: (Rao-Scott LRT)
## svyglm(formula = cpodoms ~ pais, disenio_urubra, family = quasipoisson())
## stats DEff df ddf p
## pais 25.8 8.36 1.00 80 0.085 .
## escolmae13cat 430.6 8.03 2.00 78 1.7e-07 ***
## freqescovacat 140.8 4.28 2.00 76 1.0e-05 ***
## tipoesc 64.9 3.51 1.00 75 5.5e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1