我不是Stata用户所以我正在尝试重现在R中给我的Stata结果。我想使用具有补充log-log功能的GLM。我的stata代码是：
glm c IndA fia，family（二项 s ）链接（cloglog）偏移量（偏移量）
R代码是：
glmt <- glm(data=dataset, c ~ IndA + fia, offset = offset,
family = binomial(link = cloglog))
这产生了与Stata输出不同的结果。我认为区别在于变量 s ，它包含在Stata的二项式系列中（代码中的粗体）。如何将此变量合并到R代码中？
我的数据集如下：
IndA s c itot fia offset
1 23 0 61 0.442622951 -0.494296322
1 25 0 58 0.431034483 -0.544727175
1 27 0 59 0.389830508 -0.527632742
1 31 3 51 0.37254902 -0.673344553
1 28 2 53 0.41509434 -0.634878272
1 26 0 55 0.436363636 -0.597837001
...
其中IndA是一个虚拟变量，在数据集中稍后为0。 c是s（n - （n + 1））的差异。
R输出如下所示：
Call:
glm(formula = c ~ IndA + fia, family = binomial(link = cloglog),
data = dataset, offset = offset)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.2697 -0.9707 -0.8304 1.3688 1.6390
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 1.9633 1.9185 1.023 0.306
IndA -0.3174 0.3357 -0.945 0.344
fia -5.1155 4.8163 -1.062 0.288
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 136.81 on 101 degrees of freedom
Residual deviance: 134.71 on 99 degrees of freedom
AIC: 140.71
Number of Fisher Scoring iterations: 5
Stata输出有点乱，但看起来像这样：
. glm c IndA fia, family(binomial s ) link(cloglog) offset(offset)
Iteration 0: log likelihood = -144.17967
Iteration 1: log likelihood = -133.66053
Iteration 2: log likelihood = -133.58996
Iteration 3: log likelihood = -133.58992
Iteration 4: log likelihood = -133.58992
Generalized linear models No. of obs = 102
Optimization : ML Residual df = 99
Scale parameter = 1
Deviance = 179.1806126 (1/df) Deviance = 1.809905
Pearson = 203.965157 (1/df) Pearson = 2.060254
Variance function: V(u) = u*(1-u/s) [Binomial]
Link function : g(u) = ln(-ln(1-u/s)) [Complementary log-log]
AIC = 2.678234
Log likelihood = -133.5899239 BIC = -278.6917
OIM
c Coef. Std. Err. z P>|z| [95% Conf. Interval]
IndA -.7284992 .2308676 -3.16 0.002 -1.180991 -.2760071
fia -7.147842 3.185532 -2.24 0.025 -13.39137 -.9043128
_cons .4404201 1.265651 0.35 0.728 -2.040211 2.921051
offset (offset)
此输出适用于整个数据集：
IndA s c itot fia offset
1 23 0 61 0.442622951 -0.494296322
1 25 0 58 0.431034483 -0.544727175
1 27 0 59 0.389830508 -0.527632742
1 31 3 51 0.37254902 -0.673344553
1 28 2 53 0.41509434 -0.634878272
1 26 0 55 0.436363636 -0.597837001
1 26 0 52 0.461538462 -0.653926467
1 27 0 53 0.433962264 -0.634878272
1 29 1 50 0.42 -0.693147181
1 28 0 52 0.423076923 -0.653926467
1 28 0 56 0.392857143 -0.579818495
1 30 4 50 0.4 -0.693147181
1 26 0 57 0.421052632 -0.562118918
1 26 1 56 0.428571429 -0.579818495
1 25 0 58 0.431034483 -0.544727175
1 26 0 56 0.428571429 -0.579818495
1 29 3 54 0.388888889 -0.616186139
1 26 3 58 0.413793103 -0.544727175
1 23 0 62 0.435483871 -0.478035801
1 23 0 62 0.435483871 -0.478035801
1 25 0 59 0.423728814 -0.527632742
1 27 3 54 0.425925926 -0.616186139
1 24 0 60 0.433333333 -0.510825624
1 25 0 60 0.416666667 -0.510825624
1 25 0 60 0.416666667 -0.510825624
1 26 0 57 0.421052632 -0.562118918
1 27 0 55 0.418181818 -0.597837001
1 27 0 53 0.433962264 -0.634878272
1 27 0 55 0.418181818 -0.597837001
1 29 0 56 0.375 -0.579818495
1 31 0 53 0.358490566 -0.634878272
1 31 0 52 0.365384615 -0.653926467
1 34 0 50 0.32 -0.693147181
1 34 1 51 0.31372549 -0.673344553
1 33 5 55 0.309090909 -0.597837001
1 28 0 60 0.366666667 -0.510825624
1 28 1 58 0.379310345 -0.544727175
1 27 0 58 0.396551724 -0.544727175
1 28 0 58 0.379310345 -0.544727175
1 28 1 58 0.379310345 -0.544727175
1 27 0 59 0.389830508 -0.527632742
1 27 0 59 0.389830508 -0.527632742
1 27 0 57 0.403508772 -0.562118918
1 29 1 53 0.396226415 -0.634878272
1 28 0 55 0.4 -0.597837001
1 30 1 54 0.37037037 -0.616186139
1 29 0 54 0.388888889 -0.616186139
1 31 1 50 0.38 -0.693147181
1 30 0 57 0.350877193 -0.562118918
1 30 4 57 0.350877193 -0.562118918
1 26 0 61 0.393442623 -0.494296322
0 16 0 61 0.442622951 -0.494296322
0 17 3 58 0.431034483 -0.544727175
0 14 0 59 0.389830508 -0.527632742
0 18 0 51 0.37254902 -0.673344553
0 19 0 53 0.41509434 -0.634878272
0 19 0 55 0.436363636 -0.597837001
0 22 2 52 0.461538462 -0.653926467
0 20 0 53 0.433962264 -0.634878272
0 21 1 50 0.42 -0.693147181
0 20 4 52 0.423076923 -0.653926467
0 16 0 56 0.392857143 -0.579818495
0 20 3 50 0.4 -0.693147181
0 17 0 57 0.421052632 -0.562118918
0 18 1 56 0.428571429 -0.579818495
0 17 0 58 0.431034483 -0.544727175
0 18 1 56 0.428571429 -0.579818495
0 17 1 54 0.388888889 -0.616186139
0 16 1 58 0.413793103 -0.544727175
0 15 0 62 0.435483871 -0.478035801
0 15 0 62 0.435483871 -0.478035801
0 16 0 59 0.423728814 -0.527632742
0 19 3 54 0.425925926 -0.616186139
0 16 1 60 0.433333333 -0.510825624
0 15 0 60 0.416666667 -0.510825624
0 15 0 60 0.416666667 -0.510825624
0 17 0 57 0.421052632 -0.562118918
0 18 0 55 0.418181818 -0.597837001
0 20 2 53 0.433962264 -0.634878272
0 18 3 55 0.418181818 -0.597837001
0 15 0 56 0.375 -0.579818495
0 16 0 53 0.358490566 -0.634878272
0 17 1 52 0.365384615 -0.653926467
0 16 1 50 0.32 -0.693147181
0 15 3 51 0.31372549 -0.673344553
0 12 0 55 0.309090909 -0.597837001
0 12 0 60 0.366666667 -0.510825624
0 14 0 58 0.379310345 -0.544727175
0 15 1 58 0.396551724 -0.544727175
0 14 0 58 0.379310345 -0.544727175
0 14 0 58 0.379310345 -0.544727175
0 14 0 59 0.389830508 -0.527632742
0 14 0 59 0.389830508 -0.527632742
0 16 0 57 0.403508772 -0.562118918
0 18 1 53 0.396226415 -0.634878272
0 17 1 55 0.4 -0.597837001
0 16 0 54 0.37037037 -0.616186139
0 17 0 54 0.388888889 -0.616186139
0 19 6 50 0.38 -0.693147181
0 13 0 57 0.350877193 -0.562118918
0 13 0 57 0.350877193 -0.562118918
0 13 1 61 0.393442623 -0.494296322
希望这有帮助。
提前谢谢！
答案 0 :(得分：2)
Stata family name
familyname Description ------------------------------------------------------------------------- gaussian Gaussian (normal) igaussian inverse Gaussian binomial[varnameN|#N] Bernoulli/binomial poisson Poisson nbinomial[#k|ml] negative binomial gamma gamma
似乎 s 是数据集中的变量之一。如果没有看到您的数据结构，很难说明R glm中该变量的位置。
有关stata glm示例，请参阅此link。
二项分布可以指定为1）族（二项式），2） family（二项式#N）或3）family（二项式varnameN）。在案例2中，＃N是 二项分母N的值，试验次数。 指定族（二项式1）与指定相同 家庭（二项式）。在案例3中，varnameN是包含的变量 二项分母，允许试验数量不同 观察结果。