Page 1Page 2Page 3Page 4Page 5Page 6Page 7Page 8Page 9Page 10Page 11GVPT 722Examples of Multinomial Logit Models in Stata(1) Model of the 1992 presidential vote (1=Bush, 2=Clinton, 3=Perot), 1992 NES(a) Perot as comparison category. mlogit presvote3 partyid ideology commit wordgod white sex age income educ, base(3) Iteration 0: log likelihood = -1167.5375Iteration 1: log likelihood = -785.43762Iteration 2: log likelihood = -738.53413Iteration 3: log likelihood = -732.23139Iteration 4: log likelihood = -732.03481Iteration 5: log likelihood = -732.03457Multinomial logistic regression Number of obs = 1117 LR chi2(18) = 871.01 Prob > chi2 = 0.0000Log likelihood = -732.03457 Pseudo R2 = 0.3730------------------------------------------------------------------------------ presvote3 | Coef. Std. Err. z P>|z| [95% Conf. Interval]-------------+----------------------------------------------------------------bush | partyid | .437251 .0594439 7.36 0.000 .320743 .553759 ideology | .3564197 .0883975 4.03 0.000 .1831637 .5296756 commit | 1.118237 .3566156 3.14 0.002 .4192838 1.817191 wordgod | .2886454 .3253095 0.89 0.375 -.3489495 .9262402 white | -1.274872 .6575163 -1.94 0.053 -2.56358 .0138365 sex | .5220334 .1993014 2.62 0.009 .1314098 .912657 age | .0128939 .0062051 2.08 0.038 .0007321 .0250556 income | -.0119689 .0197983 -0.60 0.545 -.0507729 .0268351 educ | .1520729 .0666235 2.28 0.022 .0214933 .2826526 _cons | -4.037983 .9286208 -4.35 0.000 -5.858046 -2.21792-------------+----------------------------------------------------------------clinton | partyid | -.5659878 .0593468 -9.54 0.000 -.6823054 -.4496702 ideology | -.2824546 .082094 -3.44 0.001 -.4433559 -.1215534 commit | -.4039828 .3676766 -1.10 0.272 -1.124616 .3166501 wordgod | -.287159 .2787296 -1.03 0.303 -.8334589 .259141 white | -2.880011 .617463 -4.66 0.000 -4.090217 -1.669806 sex | .5037801 .200119 2.52 0.012 .111554 .8960062 age | .0235076 .0063796 3.68 0.000 .0110037 .0360114 income | -.0308004 .0187528 -1.64 0.100 -.0675552 .0059544 educ | .1676501 .0673501 2.49 0.013 .0356464 .2996538 _cons | 5.487674 .8351432 6.57 0.000 3.850824 7.124525------------------------------------------------------------------------------(Outcome presvote3==perot is the comparison group)(b) Clinton as comparison category. mlogit presvote3 partyid ideology commit wordgod white sex age income educ, base(2)Iteration 0: log likelihood = -1167.5375Iteration 1: log likelihood = -785.43762Iteration 2: log likelihood = -738.53413Iteration 3: log likelihood = -732.23139Iteration 4: log likelihood = -732.03481Iteration 5: log likelihood = -732.03457Multinomial logistic regression Number of obs = 1117 LR chi2(18) = 871.01 Prob > chi2 = 0.0000Log likelihood = -732.03457 Pseudo R2 = 0.3730------------------------------------------------------------------------------ presvote3 | Coef. Std. Err. z P>|z| [95% Conf. Interval]-------------+----------------------------------------------------------------bush | partyid | 1.003239 .0657779 15.25 0.000 .8743166 1.132161 ideology | .6388743 .0929887 6.87 0.000 .4566197 .8211288 commit | 1.52222 .394552 3.86 0.000 .7489124 2.295528 wordgod | .5758044 .3375976 1.71 0.088 -.0858748 1.237484 white | 1.605139 .4004365 4.01 0.000 .8202984 2.38998 sex | .0182533 .2160687 0.08 0.933 -.4052336 .4417401 age | -.0106137 .0066241 -1.60 0.109 -.0235967 .0023693 income | .0188315 .0203642 0.92 0.355 -.0210816 .0587446 educ | -.0155772 .0721162 -0.22 0.829 -.1569223 .125768 _cons | -9.525657 .8230031 -11.57 0.000 -11.13871 -7.9126-------------+----------------------------------------------------------------perot | partyid | .5659878 .0593468 9.54 0.000 .4496702 .6823054 ideology | .2824546 .082094 3.44 0.001 .1215534 .4433559 commit | .4039828 .3676766 1.10 0.272 -.3166501 1.124616 wordgod | .287159 .2787296 1.03 0.303 -.259141 .8334589 white | 2.880011 .617463 4.66 0.000 1.669806 4.090217 sex | -.5037801 .200119 -2.52 0.012 -.8960062 -.111554 age | -.0235076 .0063796 -3.68 0.000 -.0360114 -.0110037 income | .0308004 .0187528 1.64 0.100 -.0059544 .0675552 educ | -.1676501 .0673501 -2.49 0.013 -.2996538 -.0356464 _cons | -5.487674 .8351432 -6.57 0.000 -7.124525 -3.850824------------------------------------------------------------------------------(Outcome presvote3==clinton is the comparison group)(c) The effect of independent variables on the odds of supporting one candidate asopposed to another:. listcoef, helpmlogit (N=1117): Factor Change in the Odds of presvote3 Variable: partyid (sd=2.1406531) Odds comparing|Group 1 vs Group 2| b z P>|z| e^b e^bStdX------------------+---------------------------------------------bush -clinton | 1.00324 15.252 0.000 2.7271 8.5642bush -perot | 0.43725 7.356 0.000 1.5484 2.5498clinton -bush | -1.00324 -15.252 0.000 0.3667 0.1168clinton -perot | -0.56599 -9.537 0.000 0.5678 0.2977perot -bush | -0.43725 -7.356 0.000 0.6458 0.3922perot -clinton | 0.56599 9.537 0.000 1.7612 3.3588----------------------------------------------------------------Variable: ideology (sd=1.4309962) Odds comparing|Group 1 vs Group 2| b
View Full Document