Model Building by Deletion
A significant result of the score test (AKA Lagrange multiplier test) indicates that the assumption of proportional odds does not hold, in which case multinomial logistic regression must be used instead of ordinal logistic regression.
use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): omodel mlogit BP_cat i.MJ gndr i.EDUC_cat i.race_eth indfmpir ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
i.race_eth _Irace_eth_1-4 (naturally coded; _Irace_eth_1 omitted)
omodel is not supported by svy with vce(linearized); see help svy estimation
for a list of Stata estimation commands that are supported by svy
r(322);
end of do-file
r(322);Because the omodel function does not support survey data, multinomial logistic regression must be used instead of ordinal logistic regression.
This model includes
use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat i.race_eth indfmpir ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
i.race_eth _Irace_eth_1-4 (naturally coded; _Irace_eth_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,611
Number of PSUs = 214 Population size = 298,035,877
Subpop. no. obs = 19,441
Subpop. size = 133,015,285
Design df = 109
F( 42, 68) = 40.76
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0363006 .0590408 -0.61 0.540 -.1533175 .0807163
_IMJ_2 | -.0041059 .095461 -0.04 0.966 -.1933065 .1850947
_IMJ_3 | .2927856 .1938602 1.51 0.134 -.091439 .6770103
_IMJ_4 | .2812502 .1370245 2.05 0.043 .009672 .5528283
gndr | .8140264 .0505111 16.12 0.000 .7139151 .9141377
_IEDUC_cat_2 | .1516013 .1161073 1.31 0.194 -.0785195 .3817221
_IEDUC_cat_3 | -.017117 .1250915 -0.14 0.891 -.2650442 .2308102
_IEDUC_cat_4 | .0454332 .1247408 0.36 0.716 -.2017989 .2926654
_IEDUC_cat_5 | -.2973366 .1229019 -2.42 0.017 -.5409241 -.0537491
_Irace_eth_2 | .2211874 .0558161 3.96 0.000 .1105616 .3318132
_Irace_eth_3 | -.0091752 .0773064 -0.12 0.906 -.162394 .1440436
_Irace_eth_4 | -.1036474 .0715179 -1.45 0.150 -.2453935 .0380988
indfmpir | .0185861 .0188483 0.99 0.326 -.0187707 .0559428
ridageyr | .0276735 .0029498 9.38 0.000 .0218271 .0335198
_cons | -2.629235 .1775434 -14.81 0.000 -2.98112 -2.27735
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0717828 .0613775 -1.17 0.245 -.193431 .0498654
_IMJ_2 | -.0207673 .1008804 -0.21 0.837 -.220709 .1791743
_IMJ_3 | -.1160899 .188758 -0.62 0.540 -.4902021 .2580223
_IMJ_4 | -.0415222 .1224468 -0.34 0.735 -.2842077 .2011634
gndr | .8953139 .0512615 17.47 0.000 .7937153 .9969124
_IEDUC_cat_2 | .0865293 .1222806 0.71 0.481 -.155827 .3288855
_IEDUC_cat_3 | .194514 .1116929 1.74 0.084 -.0268577 .4158857
_IEDUC_cat_4 | .1925886 .1157982 1.66 0.099 -.0369196 .4220967
_IEDUC_cat_5 | -.1571692 .1188222 -1.32 0.189 -.3926711 .0783326
_Irace_eth_2 | .2986566 .0692219 4.31 0.000 .161461 .4358521
_Irace_eth_3 | -.144648 .0743 -1.95 0.054 -.2919082 .0026123
_Irace_eth_4 | -.1083194 .0726097 -1.49 0.139 -.2522294 .0355906
indfmpir | .0075501 .0172596 0.44 0.663 -.0266577 .041758
ridageyr | .0472498 .0024167 19.55 0.000 .04246 .0520395
_cons | -3.295797 .1650432 -19.97 0.000 -3.622907 -2.968687
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | -.0026754 .0748614 -0.04 0.972 -.1510482 .1456974
_IMJ_2 | .0110587 .1231324 0.09 0.929 -.2329856 .2551031
_IMJ_3 | .0035981 .2455748 0.01 0.988 -.4831231 .4903193
_IMJ_4 | -.019241 .1720499 -0.11 0.911 -.3602384 .3217563
gndr | 1.028395 .0692889 14.84 0.000 .8910664 1.165723
_IEDUC_cat_2 | .1713297 .1515171 1.13 0.261 -.1289723 .4716318
_IEDUC_cat_3 | .0999278 .1446825 0.69 0.491 -.1868283 .3866839
_IEDUC_cat_4 | .1405151 .1638887 0.86 0.393 -.1843069 .4653372
_IEDUC_cat_5 | -.3236899 .1825127 -1.77 0.079 -.6854242 .0380444
_Irace_eth_2 | .9097815 .0784406 11.60 0.000 .7543147 1.065248
_Irace_eth_3 | .0009256 .0951241 0.01 0.992 -.1876072 .1894584
_Irace_eth_4 | .1253278 .1033056 1.21 0.228 -.0794205 .3300761
indfmpir | -.0561899 .0224183 -2.51 0.014 -.1006222 -.0117576
ridageyr | .0848647 .0027754 30.58 0.000 .079364 .0903655
_cons | -5.585293 .1804015 -30.96 0.000 -5.942843 -5.227743
------------------------------------------------------------------------------Store full model ORs in a matrix for future:
data.0 <- c(-.0363006,-.0041059,.2927856,.2812502,-.0717828,-.0207673,-.1160899,-.0415222,-.0026754,.0110587,.0035981,-.019241)
m.0 <- matrix(data.0, nrow = 3, ncol = 4, byrow = TRUE)
m.0 [,1] [,2] [,3] [,4]
[1,] -0.0363006 -0.0041059 0.2927856 0.2812502
[2,] -0.0717828 -0.0207673 -0.1160899 -0.0415222
[3,] -0.0026754 0.0110587 0.0035981 -0.0192410OR.0 <- exp(m.0)
OR.0 [,1] [,2] [,3] [,4]
[1,] 0.9643504 0.9959025 1.3401554 1.3247850
[2,] 0.9307330 0.9794469 0.8903952 0.9593280
[3,] 0.9973282 1.0111201 1.0036046 0.9809429As gender is a political variable, it will not be considered for removal.
use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.race_eth indfmpir ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.race_eth _Irace_eth_1-4 (naturally coded; _Irace_eth_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,616
Number of PSUs = 214 Population size = 298,070,653
Subpop. no. obs = 19,446
Subpop. size = 133,050,061
Design df = 109
F( 30, 80) = 63.62
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0259609 .0577603 -0.45 0.654 -.14044 .0885181
_IMJ_2 | .0306447 .0950866 0.32 0.748 -.1578138 .2191032
_IMJ_3 | .3333996 .1935347 1.72 0.088 -.0501798 .716979
_IMJ_4 | .3272004 .1364079 2.40 0.018 .0568443 .5975564
gndr | .8226529 .0505841 16.26 0.000 .7223969 .9229089
_Irace_eth_2 | .2382433 .0555762 4.29 0.000 .128093 .3483936
_Irace_eth_3 | .0338704 .0723626 0.47 0.641 -.10955 .1772907
_Irace_eth_4 | -.1184785 .072962 -1.62 0.107 -.2630869 .0261299
indfmpir | -.0193956 .0175245 -1.11 0.271 -.0541285 .0153374
ridageyr | .0278214 .0028775 9.67 0.000 .0221182 .0335246
_cons | -2.602814 .1288933 -20.19 0.000 -2.858276 -2.347352
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0580872 .0609093 -0.95 0.342 -.1788074 .062633
_IMJ_2 | .0137353 .1007186 0.14 0.892 -.1858857 .2133562
_IMJ_3 | -.0749648 .1900303 -0.39 0.694 -.4515987 .301669
_IMJ_4 | .004502 .1232108 0.04 0.971 -.2396978 .2487018
gndr | .9007623 .0502784 17.92 0.000 .8011122 1.000412
_Irace_eth_2 | .3113405 .0696749 4.47 0.000 .1732471 .4494338
_Irace_eth_3 | -.1445504 .0723536 -2.00 0.048 -.2879529 -.0011479
_Irace_eth_4 | -.1335835 .0719351 -1.86 0.066 -.2761565 .0089896
indfmpir | -.0225824 .0150605 -1.50 0.137 -.0524318 .0072669
ridageyr | .0473137 .0024486 19.32 0.000 .0424607 .0521667
_cons | -3.152063 .1258539 -25.05 0.000 -3.401502 -2.902625
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .0114964 .074848 0.15 0.878 -.1368499 .1598427
_IMJ_2 | .0560183 .1198869 0.47 0.641 -.1815936 .2936302
_IMJ_3 | .0541847 .2459139 0.22 0.826 -.4332086 .541578
_IMJ_4 | .034036 .1746708 0.19 0.846 -.3121559 .3802279
gndr | 1.037176 .069356 14.95 0.000 .8997147 1.174637
_Irace_eth_2 | .9294336 .0778645 11.94 0.000 .7751087 1.083759
_Irace_eth_3 | .0289344 .0920223 0.31 0.754 -.1534507 .2113195
_Irace_eth_4 | .1064711 .1011617 1.05 0.295 -.0940281 .3069703
indfmpir | -.1023022 .0192978 -5.30 0.000 -.1405498 -.0640545
ridageyr | .08526 .0028421 30.00 0.000 .0796271 .0908929
_cons | -5.49039 .1445096 -37.99 0.000 -5.776804 -5.203977
------------------------------------------------------------------------------data.E <- c(-.0259609,.0306447,.3333996,.3272004,-.0580872,.0137353,-.0749648,.004502,.0114964,.0560183,.0541847,.034036)
m.E <- matrix(data.E, nrow = 3, ncol = 4, byrow = TRUE)
#m.E
OR.E <- exp(m.E)
OR.E [,1] [,2] [,3] [,4]
[1,] 0.9743732 1.031119 1.3957049 1.387079
[2,] 0.9435677 1.013830 0.9277761 1.004512
[3,] 1.0115627 1.057617 1.0556796 1.034622D_OR <- (OR.0-OR.E)/OR.0
100*D_OR [,1] [,2] [,3] [,4]
[1,] -1.039334 -3.536146 -4.145003 -4.702227
[2,] -1.378981 -3.510472 -4.198245 -4.709975
[3,] -1.427270 -4.598560 -5.188795 -5.472176100*mean(abs(D_OR))[1] 3.658932use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat indfmpir ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,611
Number of PSUs = 214 Population size = 298,035,877
Subpop. no. obs = 19,441
Subpop. size = 133,015,285
Design df = 109
F( 33, 77) = 41.53
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0302624 .0577596 -0.52 0.601 -.14474 .0842153
_IMJ_2 | .0128038 .0943648 0.14 0.892 -.174224 .1998317
_IMJ_3 | .3171098 .1904571 1.66 0.099 -.06037 .6945896
_IMJ_4 | .2975759 .1347681 2.21 0.029 .0304699 .5646818
gndr | .8071392 .0503878 16.02 0.000 .7072722 .9070062
_IEDUC_cat_2 | .1850557 .1147055 1.61 0.110 -.0422868 .4123982
_IEDUC_cat_3 | .0171188 .1219637 0.14 0.889 -.2246092 .2588469
_IEDUC_cat_4 | .0789331 .1191632 0.66 0.509 -.1572444 .3151107
_IEDUC_cat_5 | -.2719444 .1168893 -2.33 0.022 -.5036152 -.0402737
indfmpir | .0150317 .019297 0.78 0.438 -.0232143 .0532776
ridageyr | .0277618 .0028977 9.58 0.000 .0220187 .033505
_cons | -2.645423 .1597037 -16.56 0.000 -2.961951 -2.328896
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0594694 .0603219 -0.99 0.326 -.1790253 .0600866
_IMJ_2 | .0109809 .0994986 0.11 0.912 -.186222 .2081838
_IMJ_3 | -.0724136 .1901414 -0.38 0.704 -.4492676 .3044405
_IMJ_4 | -.0099068 .1210962 -0.08 0.935 -.2499155 .230102
gndr | .8837196 .0512372 17.25 0.000 .7821692 .98527
_IEDUC_cat_2 | .1735089 .1222237 1.42 0.159 -.0687345 .4157522
_IEDUC_cat_3 | .2918334 .1103453 2.64 0.009 .0731327 .5105342
_IEDUC_cat_4 | .2928367 .1127233 2.60 0.011 .0694229 .5162506
_IEDUC_cat_5 | -.0640441 .1165805 -0.55 0.584 -.2951029 .1670147
indfmpir | .0044031 .0166787 0.26 0.792 -.0286537 .0374598
ridageyr | .0475489 .0023812 19.97 0.000 .0428295 .0522683
_cons | -3.390665 .1441988 -23.51 0.000 -3.676462 -3.104867
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | -.0225278 .0728994 -0.31 0.758 -.1670121 .1219565
_IMJ_2 | .0359059 .1198813 0.30 0.765 -.2016949 .2735068
_IMJ_3 | .0292962 .2390438 0.12 0.903 -.4444809 .5030733
_IMJ_4 | -.029817 .1732218 -0.17 0.864 -.373137 .3135031
gndr | 1.002373 .0692083 14.48 0.000 .865204 1.139541
_IEDUC_cat_2 | .3068448 .147112 2.09 0.039 .0152737 .5984159
_IEDUC_cat_3 | .2151521 .1359232 1.58 0.116 -.0542432 .4845474
_IEDUC_cat_4 | .2627966 .1538556 1.71 0.090 -.0421401 .5677334
_IEDUC_cat_5 | -.2173916 .1733188 -1.25 0.212 -.5609038 .1261207
indfmpir | -.0846118 .0218372 -3.87 0.000 -.1278925 -.0413311
ridageyr | .083982 .0028078 29.91 0.000 .0784171 .089547
_cons | -5.412274 .1606606 -33.69 0.000 -5.730698 -5.09385
------------------------------------------------------------------------------data.R <- c(-.0302624, .0128038,.3171098,.2975759,-.0594694 ,.0109809, -.0724136,-.0099068,-.0225278,.0359059,.0292962,-.029817)
m.R <- matrix(data.R, nrow = 3, ncol = 4, byrow = TRUE)
#m.R
OR.R <- exp(m.R)
OR.R [,1] [,2] [,3] [,4]
[1,] 0.9701909 1.012886 1.3731533 1.3465906
[2,] 0.9422644 1.011041 0.9301461 0.9901421
[3,] 0.9777241 1.036558 1.0297296 0.9706231D_OR <- (OR.0-OR.R)/OR.0
100*D_OR [,1] [,2] [,3] [,4]
[1,] -0.6056467 -1.705348 -2.462245 -1.645969
[2,] -1.2389522 -3.225755 -4.464415 -3.212048
[3,] 1.9656639 -2.515846 -2.603114 1.052027100*mean(abs(D_OR))[1] 2.224752use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat i.race_eth ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
i.race_eth _Irace_eth_1-4 (naturally coded; _Irace_eth_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,182
Number of PSUs = 214 Population size = 306,497,065
Subpop. no. obs = 21,012
Subpop. size = 141,476,473
Design df = 109
F( 39, 71) = 45.05
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0343907 .0569046 -0.60 0.547 -.1471738 .0783925
_IMJ_2 | .0116518 .0944596 0.12 0.902 -.1755641 .1988676
_IMJ_3 | .3211071 .1864153 1.72 0.088 -.0483619 .6905761
_IMJ_4 | .2006977 .1335903 1.50 0.136 -.0640739 .4654693
gndr | .8042252 .0480668 16.73 0.000 .7089584 .8994921
_IEDUC_cat_2 | .1444326 .1107856 1.30 0.195 -.0751408 .364006
_IEDUC_cat_3 | .0392992 .1153619 0.34 0.734 -.1893442 .2679427
_IEDUC_cat_4 | .0769147 .1179903 0.65 0.516 -.1569383 .3107676
_IEDUC_cat_5 | -.2852786 .1151712 -2.48 0.015 -.5135442 -.057013
_Irace_eth_2 | .1985129 .0590286 3.36 0.001 .0815202 .3155056
_Irace_eth_3 | -.0486006 .0773644 -0.63 0.531 -.2019343 .1047332
_Irace_eth_4 | -.1196564 .0739327 -1.62 0.108 -.2661886 .0268757
ridageyr | .0277068 .0028107 9.86 0.000 .0221362 .0332775
_cons | -2.593922 .1723039 -15.05 0.000 -2.935423 -2.252422
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0769204 .0594379 -1.29 0.198 -.1947245 .0408837
_IMJ_2 | .0046071 .1014352 0.05 0.964 -.1964341 .2056484
_IMJ_3 | -.1511093 .1818947 -0.83 0.408 -.5116186 .2094
_IMJ_4 | -.139263 .1208684 -1.15 0.252 -.3788203 .1002943
gndr | .8783655 .0498407 17.62 0.000 .7795828 .9771482
_IEDUC_cat_2 | .0574299 .1098378 0.52 0.602 -.160265 .2751248
_IEDUC_cat_3 | .1931061 .1034981 1.87 0.065 -.0120237 .398236
_IEDUC_cat_4 | .1724506 .1046426 1.65 0.102 -.0349477 .3798489
_IEDUC_cat_5 | -.1883123 .1077381 -1.75 0.083 -.4018457 .0252211
_Irace_eth_2 | .3097918 .062821 4.93 0.000 .1852827 .4343009
_Irace_eth_3 | -.1655951 .0728905 -2.27 0.025 -.3100616 -.0211285
_Irace_eth_4 | -.1049987 .0693156 -1.51 0.133 -.2423799 .0323824
ridageyr | .0477974 .0024178 19.77 0.000 .0430053 .0525895
_cons | -3.262964 .1605002 -20.33 0.000 -3.58107 -2.944858
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | -.0139708 .0753943 -0.19 0.853 -.1633999 .1354583
_IMJ_2 | .0652252 .1224475 0.53 0.595 -.1774618 .3079122
_IMJ_3 | .0515842 .2378352 0.22 0.829 -.4197974 .5229658
_IMJ_4 | .0095457 .1716543 0.06 0.956 -.3306675 .3497589
gndr | 1.013071 .0682633 14.84 0.000 .8777756 1.148367
_IEDUC_cat_2 | .0659118 .1435377 0.46 0.647 -.2185752 .3503988
_IEDUC_cat_3 | .0104667 .1369104 0.08 0.939 -.2608853 .2818188
_IEDUC_cat_4 | -.0044705 .1489166 -0.03 0.976 -.2996184 .2906774
_IEDUC_cat_5 | -.5406974 .1599247 -3.38 0.001 -.8576629 -.2237319
_Irace_eth_2 | .9608631 .0779787 12.32 0.000 .8063118 1.115414
_Irace_eth_3 | .0401777 .0951914 0.42 0.674 -.1484887 .228844
_Irace_eth_4 | .1529818 .0942471 1.62 0.107 -.0338128 .3397764
ridageyr | .085122 .0025912 32.85 0.000 .0799863 .0902578
_cons | -5.645881 .1837036 -30.73 0.000 -6.009975 -5.281786
------------------------------------------------------------------------------data.I <- c(-.0343907,.0116518,.3211071,.2006977,-.0769204, .0046071,-.1511093,-.139263,-.0139708,.0652252, .0515842,.0095457)
m.I <- matrix(data.I, nrow = 3, ncol = 4, byrow = TRUE)
#m.I
OR.I <- exp(m.I)
OR.I [,1] [,2] [,3] [,4]
[1,] 0.9661939 1.011720 1.3786532 1.2222552
[2,] 0.9259636 1.004618 0.8597537 0.8699992
[3,] 0.9861263 1.067399 1.0529378 1.0095914D_OR <- (OR.0-OR.I)/OR.0
100*D_OR [,1] [,2] [,3] [,4]
[1,] -0.1911725 -1.588251 -2.872637 7.739353
[2,] 0.5124425 -2.569907 3.441332 9.311606
[3,] 1.1231846 -5.566035 -4.915607 -2.920504100*mean(abs(D_OR))[1] 3.562669use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat i.race_eth indfmpir pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
i.race_eth _Irace_eth_1-4 (naturally coded; _Irace_eth_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,611
Number of PSUs = 214 Population size = 298,035,877
Subpop. no. obs = 19,441
Subpop. size = 133,015,285
Design df = 109
F( 39, 71) = 22.01
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0425472 .0571736 -0.74 0.458 -.1558633 .070769
_IMJ_2 | -.1292948 .0940237 -1.38 0.172 -.3156468 .0570571
_IMJ_3 | .165799 .1917079 0.86 0.389 -.2141599 .5457578
_IMJ_4 | .1728389 .1330092 1.30 0.197 -.0907809 .4364588
gndr | .7733103 .0509535 15.18 0.000 .672322 .8742985
_IEDUC_cat_2 | .0648448 .1145152 0.57 0.572 -.1621206 .2918102
_IEDUC_cat_3 | -.133651 .1232426 -1.08 0.281 -.3779139 .1106118
_IEDUC_cat_4 | -.1168279 .1196253 -0.98 0.331 -.3539213 .1202656
_IEDUC_cat_5 | -.4484931 .122007 -3.68 0.000 -.690307 -.2066793
_Irace_eth_2 | .190842 .0561899 3.40 0.001 .0794756 .3022085
_Irace_eth_3 | -.093504 .076247 -1.23 0.223 -.2446232 .0576151
_Irace_eth_4 | -.1598845 .0710775 -2.25 0.026 -.3007578 -.0190112
indfmpir | .0627348 .0183474 3.42 0.001 .0263708 .0990989
_cons | -1.504913 .1160773 -12.96 0.000 -1.734974 -1.274851
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0772219 .0595323 -1.30 0.197 -.195213 .0407691
_IMJ_2 | -.2250079 .1003543 -2.24 0.027 -.4239067 -.026109
_IMJ_3 | -.3344453 .1839809 -1.82 0.072 -.6990895 .0301989
_IMJ_4 | -.2296752 .1278576 -1.80 0.075 -.4830848 .0237344
gndr | .8239349 .0503844 16.35 0.000 .7240746 .9237952
_IEDUC_cat_2 | -.0635798 .1192945 -0.53 0.595 -.3000176 .1728581
_IEDUC_cat_3 | -.0025006 .1121302 -0.02 0.982 -.224739 .2197379
_IEDUC_cat_4 | -.0776092 .1155348 -0.67 0.503 -.3065954 .151377
_IEDUC_cat_5 | -.4158074 .1178728 -3.53 0.001 -.6494274 -.1821873
_Irace_eth_2 | .2402457 .0669965 3.59 0.001 .1074608 .3730306
_Irace_eth_3 | -.2986046 .0707857 -4.22 0.000 -.4388995 -.1583097
_Irace_eth_4 | -.2073988 .0708371 -2.93 0.004 -.3477957 -.067002
indfmpir | .0810614 .0176597 4.59 0.000 .0460605 .1160622
_cons | -1.318593 .130153 -10.13 0.000 -1.576552 -1.060634
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | -.0113785 .0681792 -0.17 0.868 -.1465075 .1237504
_IMJ_2 | -.3217282 .1166797 -2.76 0.007 -.5529835 -.0904729
_IMJ_3 | -.3938527 .236956 -1.66 0.099 -.8634918 .0757864
_IMJ_4 | -.3773404 .1627376 -2.32 0.022 -.699881 -.0547998
gndr | .9040703 .0663395 13.63 0.000 .7725875 1.035553
_IEDUC_cat_2 | -.0842657 .1438086 -0.59 0.559 -.3692897 .2007583
_IEDUC_cat_3 | -.2293515 .1392577 -1.65 0.102 -.5053558 .0466528
_IEDUC_cat_4 | -.3029697 .155296 -1.95 0.054 -.6107613 .0048219
_IEDUC_cat_5 | -.7619093 .1732619 -4.40 0.000 -1.105309 -.4185098
_Irace_eth_2 | .7950888 .0731534 10.87 0.000 .6501011 .9400764
_Irace_eth_3 | -.3031253 .0865927 -3.50 0.001 -.4747492 -.1315014
_Irace_eth_4 | -.0570139 .0988357 -0.58 0.565 -.2529031 .1388754
indfmpir | .0628632 .0223667 2.81 0.006 .0185332 .1071933
_cons | -1.842705 .1477248 -12.47 0.000 -2.13549 -1.549919
------------------------------------------------------------------------------data.A <- c(-.0425472, -.1292948,.165799,.1728389,-.0772219,-.2250079,-.3344453,-.2296752,-.0113785,-.3217282,-.3938527,-.3773404)
m.A <- matrix(data.A, nrow = 3, ncol = 4, byrow = TRUE)
#m.A
OR.A <- exp(m.A)
OR.A [,1] [,2] [,3] [,4]
[1,] 0.9583452 0.8787149 1.1803358 1.1886746
[2,] 0.9256844 0.7985099 0.7157350 0.7947917
[3,] 0.9886860 0.7248952 0.6744534 0.6856826D_OR <- (OR.0-OR.A)/OR.0
100*D_OR [,1] [,2] [,3] [,4]
[1,] 0.6227131 11.76698 11.92545 10.27415
[2,] 0.5424335 18.47338 19.61603 17.15121
[3,] 0.8665338 28.30770 32.79690 30.09964100*mean(abs(D_OR))[1] 15.20359Of the variables above, race has the least average change in OR; it will be excluded from the next cycle.
OR.1 <- OR.Ruse "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr indfmpir ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,616
Number of PSUs = 214 Population size = 298,070,653
Subpop. no. obs = 19,446
Subpop. size = 133,050,061
Design df = 109
F( 21, 89) = 74.46
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0198477 .055943 -0.35 0.723 -.130725 .0910295
_IMJ_2 | .0475298 .0937731 0.51 0.613 -.1383254 .2333851
_IMJ_3 | .3575815 .190307 1.88 0.063 -.0196008 .7347639
_IMJ_4 | .3436251 .134084 2.56 0.012 .0778748 .6093753
gndr | .8161761 .0504202 16.19 0.000 .716245 .9161073
indfmpir | -.0249385 .0179398 -1.39 0.167 -.0604946 .0106175
ridageyr | .0278504 .0028397 9.81 0.000 .0222223 .0334786
_cons | -2.576898 .1167656 -22.07 0.000 -2.808323 -2.345472
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0386858 .0595199 -0.65 0.517 -.1566524 .0792807
_IMJ_2 | .0536644 .0994729 0.54 0.591 -.1434877 .2508166
_IMJ_3 | -.0218283 .1911706 -0.11 0.909 -.4007223 .3570656
_IMJ_4 | .0459204 .1212766 0.38 0.706 -.1944459 .2862867
gndr | .8864468 .0501944 17.66 0.000 .7869631 .9859304
indfmpir | -.0235913 .0144778 -1.63 0.106 -.0522858 .0051033
ridageyr | .047544 .0024258 19.60 0.000 .0427361 .052352
_cons | -3.163172 .1097447 -28.82 0.000 -3.380683 -2.945662
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | -.0025774 .0729153 -0.04 0.972 -.1470932 .1419384
_IMJ_2 | .0900549 .1159895 0.78 0.439 -.1398326 .3199424
_IMJ_3 | .088156 .2390929 0.37 0.713 -.3857183 .5620304
_IMJ_4 | .0311304 .1756605 0.18 0.860 -.3170231 .3792839
gndr | 1.008343 .0693471 14.54 0.000 .8708997 1.145787
indfmpir | -.1309781 .018752 -6.98 0.000 -.1681441 -.0938122
ridageyr | .0842744 .0028685 29.38 0.000 .0785891 .0899597
_cons | -5.199184 .1254282 -41.45 0.000 -5.447779 -4.95059
------------------------------------------------------------------------------data.E <- c(-.0198477,.0475298,.3575815,.3436251,-.0386858,.0536644,-.0218283,.0459204, -.0025774,.0900549,.088156,.0311304)
m.E <- matrix(data.E, nrow = 3, ncol = 4, byrow = TRUE)
#m.E
OR.E <- exp(m.E)
OR.E [,1] [,2] [,3] [,4]
[1,] 0.9803480 1.048677 1.4298671 1.410050
[2,] 0.9620529 1.055130 0.9784082 1.046991
[3,] 0.9974259 1.094234 1.0921585 1.031620D_OR <- (OR.1-OR.E)/OR.1
100*D_OR [,1] [,2] [,3] [,4]
[1,] -1.046912 -3.533599 -4.130184 -4.712593
[2,] -2.100108 -4.360754 -5.188659 -5.741495
[3,] -2.015074 -5.564188 -6.062653 -6.284301100*mean(abs(D_OR))[1] 4.228377use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,182
Number of PSUs = 214 Population size = 306,497,065
Subpop. no. obs = 21,012
Subpop. size = 141,476,473
Design df = 109
F( 30, 80) = 50.60
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0237849 .0563017 -0.42 0.674 -.135373 .0878033
_IMJ_2 | .0333968 .0939846 0.36 0.723 -.1528777 .2196713
_IMJ_3 | .3524385 .1822094 1.93 0.056 -.0086946 .7135717
_IMJ_4 | .2235829 .1308588 1.71 0.090 -.035775 .4829408
gndr | .7963692 .0479554 16.61 0.000 .7013231 .8914153
_IEDUC_cat_2 | .1876414 .1082255 1.73 0.086 -.0268581 .4021409
_IEDUC_cat_3 | .0867666 .111372 0.78 0.438 -.1339691 .3075023
_IEDUC_cat_4 | .1237756 .1114965 1.11 0.269 -.0972068 .3447579
_IEDUC_cat_5 | -.2485626 .1090789 -2.28 0.025 -.4647534 -.0323718
ridageyr | .027839 .002753 10.11 0.000 .0223826 .0332953
_cons | -2.645468 .1508878 -17.53 0.000 -2.944522 -2.346413
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.063084 .0574096 -1.10 0.274 -.176868 .0506999
_IMJ_2 | .0397107 .1003663 0.40 0.693 -.1592121 .2386335
_IMJ_3 | -.1019097 .1825296 -0.56 0.578 -.4636773 .259858
_IMJ_4 | -.1035826 .1190715 -0.87 0.386 -.3395784 .1324133
gndr | .8651325 .0496565 17.42 0.000 .7667148 .9635501
_IEDUC_cat_2 | .1513488 .1086468 1.39 0.166 -.0639856 .3666831
_IEDUC_cat_3 | .2991265 .1016349 2.94 0.004 .0976895 .5005635
_IEDUC_cat_4 | .281089 .1006702 2.79 0.006 .081564 .480614
_IEDUC_cat_5 | -.0905159 .1026045 -0.88 0.380 -.2938747 .1128429
ridageyr | .0480488 .002365 20.32 0.000 .0433615 .0527361
_cons | -3.372985 .1351686 -24.95 0.000 -3.640885 -3.105085
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | -.0446323 .0731996 -0.61 0.543 -.1897115 .1004469
_IMJ_2 | .0943642 .1191925 0.79 0.430 -.1418715 .3305998
_IMJ_3 | .0842452 .2321425 0.36 0.717 -.3758538 .5443441
_IMJ_4 | .0039705 .1705128 0.02 0.981 -.3339803 .3419213
gndr | .9814749 .0680554 14.42 0.000 .8465914 1.116358
_IEDUC_cat_2 | .1851024 .1401928 1.32 0.189 -.0927551 .4629599
_IEDUC_cat_3 | .0827937 .1272784 0.65 0.517 -.1694679 .3350553
_IEDUC_cat_4 | .0641297 .1383301 0.46 0.644 -.2100362 .3382955
_IEDUC_cat_5 | -.5209059 .1483462 -3.51 0.001 -.8149233 -.2268885
ridageyr | .0832283 .0026016 31.99 0.000 .0780719 .0883846
_cons | -5.435086 .1587374 -34.24 0.000 -5.749698 -5.120474
------------------------------------------------------------------------------data.I <- c(-.0237849, .0333968,.3524385,.2235829,-.063084,.0397107,-.1019097,-.1035826,-.0446323, .0943642,.0842452,.0039705)
m.I <- matrix(data.I, nrow = 3, ncol = 4, byrow = TRUE)
#m.I
OR.I <- exp(m.I)
OR.I [,1] [,2] [,3] [,4]
[1,] 0.9764957 1.033961 1.4225322 1.2505493
[2,] 0.9388646 1.040510 0.9031111 0.9016015
[3,] 0.9563491 1.098960 1.0878956 1.0039784D_OR <- (OR.1-OR.I)/OR.1
100*D_OR [,1] [,2] [,3] [,4]
[1,] -0.6498524 -2.080650 -3.596017 7.132181
[2,] 0.3608075 -2.914648 2.906534 8.942208
[3,] 2.1861986 -6.020077 -5.648673 -3.436478100*mean(abs(D_OR))[1] 3.82286use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat indfmpir pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,611
Number of PSUs = 214 Population size = 298,035,877
Subpop. no. obs = 19,441
Subpop. size = 133,015,285
Design df = 109
F( 30, 80) = 21.35
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0251341 .0560716 -0.45 0.655 -.1362662 .085998
_IMJ_2 | -.1032825 .0932865 -1.11 0.271 -.2881733 .0816083
_IMJ_3 | .1999113 .1884477 1.06 0.291 -.173586 .5734085
_IMJ_4 | .1994269 .1314965 1.52 0.132 -.0611948 .4600486
gndr | .7642288 .0507867 15.05 0.000 .6635713 .8648863
_IEDUC_cat_2 | .1211163 .1133522 1.07 0.288 -.1035442 .3457767
_IEDUC_cat_3 | -.070009 .1206141 -0.58 0.563 -.3090623 .1690442
_IEDUC_cat_4 | -.053951 .1147216 -0.47 0.639 -.2813256 .1734235
_IEDUC_cat_5 | -.392255 .116271 -3.37 0.001 -.6227003 -.1618098
indfmpir | .0626615 .0187375 3.34 0.001 .0255243 .0997987
_cons | -1.581047 .1074383 -14.72 0.000 -1.793986 -1.368108
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0465438 .0585697 -0.79 0.429 -.162627 .0695395
_IMJ_2 | -.1791071 .0995009 -1.80 0.075 -.3763146 .0181004
_IMJ_3 | -.2729503 .1848875 -1.48 0.143 -.6393914 .0934907
_IMJ_4 | -.1801132 .1267181 -1.42 0.158 -.4312644 .071038
gndr | .808747 .0502412 16.10 0.000 .7091707 .9083234
_IEDUC_cat_2 | .0630373 .1180616 0.53 0.594 -.170957 .2970316
_IEDUC_cat_3 | .1471284 .1103672 1.33 0.185 -.0716159 .3658727
_IEDUC_cat_4 | .0749667 .1117204 0.67 0.504 -.1464596 .2963929
_IEDUC_cat_5 | -.2685009 .1145995 -2.34 0.021 -.4956334 -.0413685
indfmpir | .084636 .017027 4.97 0.000 .050889 .118383
_cons | -1.514393 .1152079 -13.14 0.000 -1.742732 -1.286055
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .0007932 .0667942 0.01 0.991 -.1315907 .1331771
_IMJ_2 | -.2554685 .1128074 -2.26 0.026 -.4790491 -.0318879
_IMJ_3 | -.3034009 .229149 -1.32 0.188 -.7575667 .1507649
_IMJ_4 | -.3234544 .160236 -2.02 0.046 -.6410369 -.0058718
gndr | .8728168 .0663742 13.15 0.000 .7412653 1.004368
_IEDUC_cat_2 | .1296841 .1381046 0.94 0.350 -.1440347 .4034029
_IEDUC_cat_3 | -.0051845 .1314302 -0.04 0.969 -.2656748 .2553058
_IEDUC_cat_4 | -.0701819 .1452312 -0.48 0.630 -.3580255 .2176617
_IEDUC_cat_5 | -.5451717 .162787 -3.35 0.001 -.8678103 -.222533
indfmpir | .0460368 .0216097 2.13 0.035 .003207 .0888666
_cons | -1.93226 .1260928 -15.32 0.000 -2.182172 -1.682348
------------------------------------------------------------------------------data.A <- c(-.0251341,-.1032825, .1999113,.1994269,-.0465438,-.1791071,-.2729503,-.1801132,.0007932,-.2554685,-.3034009,-.3234544)
m.A <- matrix(data.A, nrow = 3, ncol = 4, byrow = TRUE)
#m.A
OR.A <- exp(m.A)
OR.A [,1] [,2] [,3] [,4]
[1,] 0.9751791 0.9018722 1.2212944 1.2207030
[2,] 0.9545228 0.8360164 0.7611306 0.8351757
[3,] 1.0007935 0.7745535 0.7383031 0.7236450D_OR <- (OR.1-OR.A)/OR.1
100*D_OR [,1] [,2] [,3] [,4]
[1,] -0.5141472 10.96016 11.05914 9.348618
[2,] -1.3009497 17.31136 18.17085 15.650930
[3,] -2.3595061 25.27641 28.30127 25.445322100*mean(abs(D_OR))[1] 13.80822Of the variables above, income has the least average change in OR; it will be excluded from the next cycle.
OR.2 <- OR.Ic(111,222,333,444,222,222,333,444,333,222,333,444)
use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr ridageyr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,190
Number of PSUs = 214 Population size = 306,545,053
Subpop. no. obs = 21,020
Subpop. size = 141,524,461
Design df = 109
F( 18, 92) = 93.07
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0227526 .0547765 -0.42 0.679 -.1313178 .0858126
_IMJ_2 | .0797134 .0935508 0.85 0.396 -.1057013 .265128
_IMJ_3 | .4072004 .181229 2.25 0.027 .0480105 .7663903
_IMJ_4 | .2865253 .1303029 2.20 0.030 .0282692 .5447813
gndr | .7982729 .0478785 16.67 0.000 .7033792 .8931666
ridageyr | .0265302 .0027076 9.80 0.000 .0211638 .0318966
_cons | -2.599187 .1109633 -23.42 0.000 -2.819113 -2.379262
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0499135 .0560524 -0.89 0.375 -.1610075 .0611805
_IMJ_2 | .0921278 .0995805 0.93 0.357 -.1052375 .2894932
_IMJ_3 | -.0401177 .1842881 -0.22 0.828 -.4053708 .3251354
_IMJ_4 | -.0355165 .119337 -0.30 0.767 -.2720385 .2010055
gndr | .8633234 .0488686 17.67 0.000 .7664675 .9601793
ridageyr | .0470443 .0023972 19.62 0.000 .0422931 .0517955
_cons | -3.200029 .1069927 -29.91 0.000 -3.412085 -2.987973
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | -.0548756 .0730604 -0.75 0.454 -.1996788 .0899277
_IMJ_2 | .1690277 .1130569 1.50 0.138 -.0550474 .3931028
_IMJ_3 | .1638104 .2305652 0.71 0.479 -.2931623 .6207831
_IMJ_4 | .0852999 .1726447 0.49 0.622 -.2568764 .4274762
gndr | .9855432 .067661 14.57 0.000 .8514412 1.119645
ridageyr | .0818487 .0025662 31.90 0.000 .0767627 .0869348
_cons | -5.475991 .1225016 -44.70 0.000 -5.718785 -5.233197
------------------------------------------------------------------------------data.E <- c(-.0227526,.0797134,.4072004,.2865253,-.0499135,.0921278,-.0401177,-.0355165,-.0548756,.1690277,.1638104,.0852999)
m.E <- matrix(data.E, nrow = 3, ncol = 4, byrow = TRUE)
#m.E
OR.E <- exp(m.E)
OR.E [,1] [,2] [,3] [,4]
[1,] 0.9775043 1.082977 1.5026052 1.3317919
[2,] 0.9513117 1.096505 0.9606764 0.9651068
[3,] 0.9466029 1.184153 1.1779909 1.0890436D_OR <- (OR.2-OR.E)/OR.2
100*D_OR [,1] [,2] [,3] [,4]
[1,] -0.1032833 -4.740597 -5.628908 -6.496550
[2,] -1.3257613 -5.381520 -6.374106 -7.043606
[3,] 1.0191016 -7.752150 -8.281616 -8.472815100*mean(abs(D_OR))[1] 5.218334use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,182
Number of PSUs = 214 Population size = 306,497,065
Subpop. no. obs = 21,012
Subpop. size = 141,476,473
Design df = 109
F( 27, 83) = 23.82
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0095226 .054527 -0.17 0.862 -.1175933 .0985482
_IMJ_2 | -.092097 .0923199 -1.00 0.321 -.275072 .0908781
_IMJ_3 | .2123234 .1799733 1.18 0.241 -.1443779 .5690246
_IMJ_4 | .1139486 .1261739 0.90 0.368 -.136124 .3640213
gndr | .7601668 .0481627 15.78 0.000 .6647099 .8556236
_IEDUC_cat_2 | .1392047 .1065898 1.31 0.194 -.0720529 .3504622
_IEDUC_cat_3 | .0484727 .1098658 0.44 0.660 -.1692778 .2662232
_IEDUC_cat_4 | .0613196 .1087664 0.56 0.574 -.1542519 .2768911
_IEDUC_cat_5 | -.2458461 .1098681 -2.24 0.027 -.4636011 -.0280911
_cons | -1.514204 .0966629 -15.66 0.000 -1.705786 -1.322621
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0340005 .0561847 -0.61 0.546 -.1453568 .0773557
_IMJ_2 | -.1676792 .0996981 -1.68 0.095 -.3652776 .0299192
_IMJ_3 | -.3426363 .1779154 -1.93 0.057 -.6952588 .0099862
_IMJ_4 | -.2926804 .1235392 -2.37 0.020 -.537531 -.0478298
gndr | .8006361 .0486398 16.46 0.000 .7042336 .8970387
_IEDUC_cat_2 | .069679 .1049868 0.66 0.508 -.1384014 .2777594
_IEDUC_cat_3 | .2403985 .1031933 2.33 0.022 .0358727 .4449242
_IEDUC_cat_4 | .1850364 .1003522 1.84 0.068 -.0138584 .3839312
_IEDUC_cat_5 | -.0856343 .1031875 -0.83 0.408 -.2901485 .11888
_cons | -1.367496 .1014487 -13.48 0.000 -1.568564 -1.166428
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .0039307 .0680466 0.06 0.954 -.1309356 .1387969
_IMJ_2 | -.2306821 .1127283 -2.05 0.043 -.454106 -.0072583
_IMJ_3 | -.3168202 .2276053 -1.39 0.167 -.7679264 .1342861
_IMJ_4 | -.3181186 .1572283 -2.02 0.045 -.6297401 -.0064972
gndr | .8692388 .0643211 13.51 0.000 .7417564 .9967213
_IEDUC_cat_2 | .059932 .135 0.44 0.658 -.2076337 .3274976
_IEDUC_cat_3 | .0129754 .1260003 0.10 0.918 -.2367531 .2627039
_IEDUC_cat_4 | -.0606929 .1313257 -0.46 0.645 -.3209762 .1995904
_IEDUC_cat_5 | -.5030832 .1425577 -3.53 0.001 -.7856279 -.2205385
_cons | -1.81337 .1220381 -14.86 0.000 -2.055246 -1.571495
------------------------------------------------------------------------------data.A <- c(-.0095226,-.092097,.2123234,.1139486,-.0340005,-.1676792,-.3426363,-.2926804,.0039307,-.2306821,-.3168202,-.3181186)
m.A <- matrix(data.A, nrow = 3, ncol = 4, byrow = TRUE)
#m.A
OR.A <- exp(m.A)
OR.A [,1] [,2] [,3] [,4]
[1,] 0.9905226 0.9120167 1.2365477 1.1206945
[2,] 0.9665710 0.8456251 0.7098964 0.7462606
[3,] 1.0039384 0.7939918 0.7284617 0.7275165D_OR <- (OR.2-OR.A)/OR.2
100*D_OR [,1] [,2] [,3] [,4]
[1,] -1.436449 11.79388 13.07418 10.38382
[2,] -2.951056 18.72973 21.39435 17.22944
[3,] -4.976150 27.75061 33.03937 27.53664100*mean(abs(D_OR))[1] 15.85797Of the variables above, education has the least average change in OR; it will be excluded from the next cycle.
OR.3 <- OR.Euse "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr i.EDUC_cat pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
i.EDUC_cat _IEDUC_cat_1-5 (naturally coded; _IEDUC_cat_1 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,182
Number of PSUs = 214 Population size = 306,497,065
Subpop. no. obs = 21,012
Subpop. size = 141,476,473
Design df = 109
F( 27, 83) = 23.82
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | -.0095226 .054527 -0.17 0.862 -.1175933 .0985482
_IMJ_2 | -.092097 .0923199 -1.00 0.321 -.275072 .0908781
_IMJ_3 | .2123234 .1799733 1.18 0.241 -.1443779 .5690246
_IMJ_4 | .1139486 .1261739 0.90 0.368 -.136124 .3640213
gndr | .7601668 .0481627 15.78 0.000 .6647099 .8556236
_IEDUC_cat_2 | .1392047 .1065898 1.31 0.194 -.0720529 .3504622
_IEDUC_cat_3 | .0484727 .1098658 0.44 0.660 -.1692778 .2662232
_IEDUC_cat_4 | .0613196 .1087664 0.56 0.574 -.1542519 .2768911
_IEDUC_cat_5 | -.2458461 .1098681 -2.24 0.027 -.4636011 -.0280911
_cons | -1.514204 .0966629 -15.66 0.000 -1.705786 -1.322621
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | -.0340005 .0561847 -0.61 0.546 -.1453568 .0773557
_IMJ_2 | -.1676792 .0996981 -1.68 0.095 -.3652776 .0299192
_IMJ_3 | -.3426363 .1779154 -1.93 0.057 -.6952588 .0099862
_IMJ_4 | -.2926804 .1235392 -2.37 0.020 -.537531 -.0478298
gndr | .8006361 .0486398 16.46 0.000 .7042336 .8970387
_IEDUC_cat_2 | .069679 .1049868 0.66 0.508 -.1384014 .2777594
_IEDUC_cat_3 | .2403985 .1031933 2.33 0.022 .0358727 .4449242
_IEDUC_cat_4 | .1850364 .1003522 1.84 0.068 -.0138584 .3839312
_IEDUC_cat_5 | -.0856343 .1031875 -0.83 0.408 -.2901485 .11888
_cons | -1.367496 .1014487 -13.48 0.000 -1.568564 -1.166428
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .0039307 .0680466 0.06 0.954 -.1309356 .1387969
_IMJ_2 | -.2306821 .1127283 -2.05 0.043 -.454106 -.0072583
_IMJ_3 | -.3168202 .2276053 -1.39 0.167 -.7679264 .1342861
_IMJ_4 | -.3181186 .1572283 -2.02 0.045 -.6297401 -.0064972
gndr | .8692388 .0643211 13.51 0.000 .7417564 .9967213
_IEDUC_cat_2 | .059932 .135 0.44 0.658 -.2076337 .3274976
_IEDUC_cat_3 | .0129754 .1260003 0.10 0.918 -.2367531 .2627039
_IEDUC_cat_4 | -.0606929 .1313257 -0.46 0.645 -.3209762 .1995904
_IEDUC_cat_5 | -.5030832 .1425577 -3.53 0.001 -.7856279 -.2205385
_cons | -1.81337 .1220381 -14.86 0.000 -2.055246 -1.571495
------------------------------------------------------------------------------data.A <- c(-.0095226,-.092097,.2123234,.1139486,-.0340005,-.1676792,-.3426363,-.2926804,.0039307,-.2306821,-.3168202,-.3181186)
m.A <- matrix(data.A, nrow = 3, ncol = 4, byrow = TRUE)
#m.A
OR.A <- exp(m.A)
OR.A [,1] [,2] [,3] [,4]
[1,] 0.9905226 0.9120167 1.2365477 1.1206945
[2,] 0.9665710 0.8456251 0.7098964 0.7462606
[3,] 1.0039384 0.7939918 0.7284617 0.7275165D_OR <- (OR.3-OR.A)/OR.3
100*D_OR [,1] [,2] [,3] [,4]
[1,] -1.331790 15.78612 17.70641 15.85063
[2,] -1.604029 22.87996 26.10453 22.67585
[3,] -6.056979 32.94854 38.16067 33.19675100*mean(abs(D_OR))[1] 19.52519Of the variables above, no average change of less than 10% is found. Therefore the model from deletion cycle 3 is the final model.
use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr ridageyr, rrr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,190
Number of PSUs = 214 Population size = 306,545,053
Subpop. no. obs = 21,020
Subpop. size = 141,524,461
Design df = 109
F( 18, 92) = 93.07
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | RRR Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | .9775043 .0535442 -0.42 0.679 .876939 1.089602
_IMJ_2 | 1.082977 .1013133 0.85 0.396 .8996934 1.303598
_IMJ_3 | 1.502605 .2723156 2.25 0.027 1.049182 2.151984
_IMJ_4 | 1.331792 .1735363 2.20 0.030 1.028673 1.724231
gndr | 2.221701 .1063717 16.67 0.000 2.020569 2.442853
ridageyr | 1.026885 .0027804 9.80 0.000 1.021389 1.032411
_cons | .074334 .0082483 -23.42 0.000 .0596588 .0926189
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | .9513117 .0533233 -0.89 0.375 .8512857 1.063091
_IMJ_2 | 1.096505 .1091905 0.93 0.357 .9001107 1.33575
_IMJ_3 | .9606764 .1770412 -0.22 0.828 .6667296 1.384218
_IMJ_4 | .9651068 .1151729 -0.30 0.767 .7618249 1.222631
gndr | 2.371028 .1158687 17.67 0.000 2.15215 2.612165
ridageyr | 1.048168 .0025127 19.62 0.000 1.0432 1.05316
_cons | .040761 .0043611 -29.91 0.000 .0329724 .0503895
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .9466029 .0691592 -0.75 0.454 .8189937 1.094095
_IMJ_2 | 1.184153 .1338767 1.50 0.138 .9464403 1.481571
_IMJ_3 | 1.177991 .2716037 0.71 0.479 .7459011 1.860384
_IMJ_4 | 1.089044 .1880177 0.49 0.622 .7734638 1.533383
gndr | 2.679267 .181282 14.57 0.000 2.343021 3.063767
ridageyr | 1.085292 .002785 31.90 0.000 1.079786 1.090826
_cons | .0041861 .0005128 -44.70 0.000 .0032837 .0053364
------------------------------------------------------------------------------
Note: _cons estimates baseline relative risk for each outcome.Record baseline RRs:
RR0 <- c(.9775043,1.082977,1.502605 ,1.331792 ,.9513117, 1.096505,.9606764,.9651068,.9466029,1.184153 ,1.177991 ,1.089044 )
RR0 <- matrix(RR0, nrow = 3, ncol = 4, byrow = TRUE)
RR0 [,1] [,2] [,3] [,4]
[1,] 0.9775043 1.082977 1.5026050 1.3317920
[2,] 0.9513117 1.096505 0.9606764 0.9651068
[3,] 0.9466029 1.184153 1.1779910 1.0890440use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr ridageyr SMK_cat, rrr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,139
Number of PSUs = 214 Population size = 306,306,164
Subpop. no. obs = 20,969
Subpop. size = 141,285,572
Design df = 109
F( 21, 89) = 77.93
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | RRR Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | .9374249 .0571627 -1.06 0.292 .8307088 1.05785
_IMJ_2 | 1.009098 .1064685 0.09 0.932 .8186839 1.243801
_IMJ_3 | 1.383885 .2492299 1.80 0.074 .968462 1.977505
_IMJ_4 | 1.2091 .1694016 1.36 0.178 .915936 1.596096
gndr | 2.209352 .1054105 16.61 0.000 2.010006 2.428469
ridageyr | 1.026261 .0027938 9.52 0.000 1.020739 1.031813
SMK_cat | 1.068204 .0285846 2.47 0.015 1.013027 1.126387
_cons | .0746479 .0082886 -23.37 0.000 .0599021 .0930235
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | .9450747 .057857 -0.92 0.358 .8370877 1.066992
_IMJ_2 | 1.089875 .1133895 0.83 0.410 .8867969 1.339457
_IMJ_3 | .9564369 .1822314 -0.23 0.816 .6556241 1.395269
_IMJ_4 | .9534254 .1162203 -0.39 0.696 .7487942 1.213979
gndr | 2.372321 .1171888 17.49 0.000 2.151065 2.616336
ridageyr | 1.048037 .0025557 19.24 0.000 1.042984 1.053115
SMK_cat | 1.006183 .0267834 0.23 0.817 .9544749 1.060692
_cons | .0409214 .0043775 -29.88 0.000 .0331034 .0505859
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .9076238 .0709204 -1.24 0.217 .7774053 1.059654
_IMJ_2 | 1.0976 .1360313 0.75 0.454 .8585507 1.403209
_IMJ_3 | 1.079162 .2554407 0.32 0.748 .6750609 1.725164
_IMJ_4 | .9788363 .1748177 -0.12 0.905 .6870383 1.394566
gndr | 2.672076 .1811746 14.50 0.000 2.336075 3.056403
ridageyr | 1.084611 .0027643 31.87 0.000 1.079146 1.090104
SMK_cat | 1.072642 .0311241 2.42 0.017 1.012695 1.136137
_cons | .0041853 .000516 -44.42 0.000 .003278 .0053438
------------------------------------------------------------------------------
Note: _cons estimates baseline relative risk for each outcome.RR_S <- c(.9374249,1.009098,1.383885 ,1.2091 ,.9450747,1.089875 ,.9564369,.9534254,.9076238,1.0976 ,1.079162,.9788363)
RR_S <- matrix(RR_S, nrow = 3, ncol = 4, byrow = TRUE)
RR_S [,1] [,2] [,3] [,4]
[1,] 0.9374249 1.009098 1.3838850 1.2091000
[2,] 0.9450747 1.089875 0.9564369 0.9534254
[3,] 0.9076238 1.097600 1.0791620 0.9788363D_RR <- (RR0-RR_S)/RR0
100*D_RR [,1] [,2] [,3] [,4]
[1,] 4.1001763 6.8218439 7.9009454 9.212550
[2,] 0.6556211 0.6046484 0.4413036 1.210374
[3,] 4.1177879 7.3092751 8.3896227 10.119674100*mean(abs(D_RR))[1] 5.073652use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr ridageyr AL_cat, rrr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,083
Number of PSUs = 214 Population size = 293,445,655
Subpop. no. obs = 18,913
Subpop. size = 128,425,062
Design df = 109
F( 21, 89) = 74.90
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | RRR Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | .8969884 .049069 -1.99 0.049 .8048219 .9997096
_IMJ_2 | .9524878 .0859745 -0.54 0.591 .7964614 1.13908
_IMJ_3 | 1.330507 .2528377 1.50 0.136 .9129489 1.939045
_IMJ_4 | 1.121004 .1480906 0.86 0.389 .8627733 1.456525
gndr | 2.361784 .1223464 16.59 0.000 2.13133 2.617156
ridageyr | 1.030069 .0029541 10.33 0.000 1.024231 1.035941
AL_cat | 1.173123 .0489471 3.83 0.000 1.080015 1.274259
_cons | .0613102 .0075926 -22.54 0.000 .0479663 .0783661
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | .8872225 .0547347 -1.94 0.055 .78511 1.002616
_IMJ_2 | 1.013692 .1019295 0.14 0.893 .8305284 1.23725
_IMJ_3 | .9123565 .1741833 -0.48 0.632 .6249314 1.331977
_IMJ_4 | .9089748 .1189676 -0.73 0.467 .7012857 1.178172
gndr | 2.52572 .1296751 18.05 0.000 2.281352 2.796263
ridageyr | 1.051498 .0027456 19.23 0.000 1.04607 1.056953
AL_cat | 1.18147 .0490749 4.01 0.000 1.088101 1.28285
_cons | .0324597 .0037792 -29.44 0.000 .0257708 .0408847
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .8909407 .0680178 -1.51 0.133 .7658352 1.036483
_IMJ_2 | 1.034994 .1271571 0.28 0.780 .8113101 1.320348
_IMJ_3 | .9619987 .2309185 -0.16 0.872 .5978028 1.548072
_IMJ_4 | .956476 .1759521 -0.24 0.809 .6642479 1.377266
gndr | 2.85697 .2004106 14.96 0.000 2.486138 3.283115
ridageyr | 1.090038 .0031562 29.77 0.000 1.083801 1.096312
AL_cat | 1.342399 .0659263 6.00 0.000 1.217893 1.479633
_cons | .0028854 .0004211 -40.07 0.000 .0021606 .0038533
------------------------------------------------------------------------------
Note: _cons estimates baseline relative risk for each outcome.RR_A <- c(.8969884,.9524878,1.330507 ,1.121004,.8872225,1.013692 ,.9123565,.9089748,.8909407,1.034994 ,.9619987,.956476)
RR_A <- matrix(RR_A, nrow = 3, ncol = 4, byrow = TRUE)
RR_A [,1] [,2] [,3] [,4]
[1,] 0.8969884 0.9524878 1.3305070 1.1210040
[2,] 0.8872225 1.0136920 0.9123565 0.9089748
[3,] 0.8909407 1.0349940 0.9619987 0.9564760D_RR <- (RR0-RR_A)/RR0
100*D_RR [,1] [,2] [,3] [,4]
[1,] 8.236884 12.049120 11.453309 15.827396
[2,] 6.736930 7.552451 5.029779 5.816144
[3,] 5.880206 12.596261 18.335649 12.172878100*mean(abs(D_RR))[1] 10.14058use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr ridageyr bmxbmi, rrr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 70,080
Number of PSUs = 214 Population size = 305,959,182
Subpop. no. obs = 20,910
Subpop. size = 140,938,589
Design df = 109
F( 21, 89) = 85.71
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | RRR Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | .9840052 .0555589 -0.29 0.776 .879827 1.100519
_IMJ_2 | 1.184859 .1127683 1.78 0.077 .9811707 1.430833
_IMJ_3 | 1.642366 .2970089 2.74 0.007 1.147648 2.350343
_IMJ_4 | 1.44872 .1941487 2.77 0.007 1.110788 1.889462
gndr | 2.317385 .1114122 17.48 0.000 2.106764 2.549063
ridageyr | 1.024799 .0028476 8.82 0.000 1.019171 1.030458
bmxbmi | 1.063045 .0047525 13.68 0.000 1.053667 1.072506
_cons | .0134204 .0022718 -25.47 0.000 .0095952 .0187705
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | .9687283 .0576282 -0.53 0.594 .8609873 1.089952
_IMJ_2 | 1.231823 .1197262 2.15 0.034 1.015986 1.493513
_IMJ_3 | 1.083761 .1985848 0.44 0.662 .7537215 1.558317
_IMJ_4 | 1.08013 .133675 0.62 0.535 .8451823 1.380391
gndr | 2.515005 .1295681 17.90 0.000 2.27088 2.785373
ridageyr | 1.04622 .0025722 18.38 0.000 1.041135 1.051331
bmxbmi | 1.077911 .0041242 19.61 0.000 1.069768 1.086116
_cons | .0047006 .0008167 -30.85 0.000 .0033312 .0066329
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .9675777 .0746639 -0.43 0.670 .830357 1.127475
_IMJ_2 | 1.399027 .1572239 2.99 0.003 1.119678 1.74807
_IMJ_3 | 1.400581 .3371062 1.40 0.164 .8692249 2.256754
_IMJ_4 | 1.292246 .2213289 1.50 0.137 .9202788 1.814559
gndr | 2.976748 .2048609 15.85 0.000 2.597194 3.411769
ridageyr | 1.084163 .0029394 29.81 0.000 1.078353 1.090005
bmxbmi | 1.104898 .0056372 19.55 0.000 1.093782 1.116128
_cons | .000206 .0000456 -38.37 0.000 .0001328 .0003193
------------------------------------------------------------------------------
Note: _cons estimates baseline relative risk for each outcome.RR_B <- c(.9840052,1.184859 , 1.642366, 1.44872,.9687283,1.231823,1.083761,1.08013,.9675777,1.399027 ,1.400581,1.292246 )
RR_B <- matrix(RR_B, nrow = 3, ncol = 4, byrow = TRUE)
RR_B [,1] [,2] [,3] [,4]
[1,] 0.9840052 1.184859 1.642366 1.448720
[2,] 0.9687283 1.231823 1.083761 1.080130
[3,] 0.9675777 1.399027 1.400581 1.292246D_RR <- (RR0-RR_B)/RR0
100*D_RR [,1] [,2] [,3] [,4]
[1,] -0.6650508 -9.407587 -9.301247 -8.779749
[2,] -1.8307985 -12.340847 -12.812285 -11.918184
[3,] -2.2157971 -18.145797 -18.895730 -18.658750100*mean(abs(D_RR))[1] 10.41432use "data\NHANES0518_new.dta", clear
svyset sdmvpsu [pw=wtmec12yr], strata(sdmvstra)
xi: svy,subpop(if include==1): mlogit BP_cat i.MJ gndr ridageyr hei2015, rrr pweight: wtmec12yr
VCE: linearized
Single unit: missing
Strata 1: sdmvstra
SU 1: sdmvpsu
FPC 1: <zero>
i.MJ _IMJ_0-4 (naturally coded; _IMJ_0 omitted)
(running mlogit on estimation sample)
Survey: Multinomial logistic regression
Number of strata = 105 Number of obs = 68,641
Number of PSUs = 214 Population size = 297,338,173
Subpop. no. obs = 19,471
Subpop. size = 132,317,580
Design df = 109
F( 21, 89) = 77.97
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
BP_cat | RRR Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Normal | (base outcome)
-------------+----------------------------------------------------------------
Elevated |
_IMJ_1 | .9800469 .0512518 -0.39 0.701 .8835545 1.087077
_IMJ_2 | 1.118512 .1096504 1.14 0.256 .9209979 1.358385
_IMJ_3 | 1.56693 .303533 2.32 0.022 1.06736 2.300322
_IMJ_4 | 1.350342 .1892399 2.14 0.034 1.022858 1.782675
gndr | 2.180578 .1090479 15.59 0.000 1.974814 2.407781
ridageyr | 1.029426 .0029668 10.06 0.000 1.023562 1.035322
hei2015 | .9926692 .002127 -3.43 0.001 .9884626 .9968938
_cons | .0995467 .0143168 -16.04 0.000 .074857 .1323796
-------------+----------------------------------------------------------------
Stage_1_HTN |
_IMJ_1 | .9278013 .0543842 -1.28 0.204 .8260391 1.0421
_IMJ_2 | 1.11908 .1175887 1.07 0.287 .9086909 1.378182
_IMJ_3 | .9262348 .1803803 -0.39 0.695 .6296399 1.362542
_IMJ_4 | .9696576 .1324793 -0.23 0.822 .7396353 1.271216
gndr | 2.255614 .1099979 16.68 0.000 2.047806 2.484509
ridageyr | 1.050978 .0028093 18.60 0.000 1.045425 1.056561
hei2015 | .9887869 .0020466 -5.45 0.000 .9847388 .9928516
_cons | .06764 .0093873 -19.41 0.000 .0513743 .0890558
-------------+----------------------------------------------------------------
Stage_2_HTN |
_IMJ_1 | .9125074 .0685265 -1.22 0.225 .7863143 1.058953
_IMJ_2 | 1.217197 .1472632 1.62 0.107 .9576832 1.547035
_IMJ_3 | 1.012749 .246387 0.05 0.959 .6253053 1.640256
_IMJ_4 | 1.010269 .1958838 0.05 0.958 .6879274 1.483651
gndr | 2.530274 .1734823 13.54 0.000 2.208777 2.898568
ridageyr | 1.089959 .003079 30.49 0.000 1.083874 1.096079
hei2015 | .9828323 .0023318 -7.30 0.000 .9782215 .9874648
_cons | .0089561 .0014716 -28.70 0.000 .0064667 .0124038
------------------------------------------------------------------------------
Note: _cons estimates baseline relative risk for each outcome.RR_H <- c(.9800469,1.118512,1.56693,1.350342,.9278013,1.11908 ,.9262348,.9696576,.9125074,1.217197 ,1.012749 ,1.010269)
RR_H <- matrix(RR_H, nrow = 3, ncol = 4, byrow = TRUE)
RR_H [,1] [,2] [,3] [,4]
[1,] 0.9800469 1.118512 1.5669300 1.3503420
[2,] 0.9278013 1.119080 0.9262348 0.9696576
[3,] 0.9125074 1.217197 1.0127490 1.0102690D_RR <- (RR0-RR_H)/RR0
100*D_RR [,1] [,2] [,3] [,4]
[1,] -0.2601114 -3.281233 -4.280899 -1.3928601
[2,] 2.4713666 -2.058814 3.585141 -0.4715333
[3,] 3.6018799 -2.790518 14.027442 7.2334084100*mean(abs(D_RR))[1] 3.787934Model 1 will be the crude model, examining BP category only as a function of MJ use category; model 2 will include gender and age as confounders; and model 3 will include gender and age, as well as all mediators (tobacco use, alcohol use, diet, & BMI).