文献来源
支付意愿空间混合logit模型：Mixed logit model in WTP space
- Talevi, M., et al. (2022). Speaking from experience: Preferences for cooking with biogas in rural India
- Appendix A. Supplementary data【数据+Stata】
示例代码
请读者自行运行1-282行，此处从略。
核心代码如下：
set seed 537363959
mixlogitwtp outcome if validated==1, rand(goodexperiencebiogas_ifbg badexperiencebiogas_ifbg noexperiencebiogas_ifbg alternisrocket maint_higher smoke_lower fuelefficiency) price(negprice) id(id) group(case)
estimates store MXL_exp
estimates restore MXL_exp
mixlbeta goodexperiencebiogas_ifbg badexperiencebiogas_ifbg noexperiencebiogas_ifbg alternisrocket maint_higher smoke_lower fuelefficiency, nrep(1000) saving(wtpexp_distribution) replace
得到结果
Iteration 0: Log likelihood = -1551.5136 (not concave)
Iteration 1: Log likelihood = -1544.5129 (not concave)
Iteration 2: Log likelihood = -1481.3067 (not concave)
Iteration 3: Log likelihood = -1461.4711 (not concave)
Iteration 4: Log likelihood = -1440.4676 (not concave)
Iteration 5: Log likelihood = -1355.3401 (not concave)
Iteration 6: Log likelihood = -1328.5785
Iteration 7: Log likelihood = -1316.8 (not concave)
Iteration 8: Log likelihood = -1315.2515 (not concave)
Iteration 9: Log likelihood = -1314.0268
Iteration 10: Log likelihood = -1313.016
Iteration 11: Log likelihood = -1312.8286 (not concave)
Iteration 12: Log likelihood = -1311.9755
Iteration 13: Log likelihood = -1311.7381
Iteration 14: Log likelihood = -1311.7282
Iteration 15: Log likelihood = -1311.7282
Mixed logit model in WTP space Number of obs = 5,484
Wald chi2(8) = 4090.87
Log likelihood = -1311.7282 Prob > chi2 = 0.0000
-------------------------------------------------------------------------------------------
outcome | Coefficient Std. err. z P>|z| [95% conf. interval]
--------------------------+----------------------------------------------------------------
Mean |
goodexperiencebiogas_ifbg | 7348.172 596.1974 12.33 0.000 6179.646 8516.697
badexperiencebiogas_ifbg | 5743.991 525.265 10.94 0.000 4714.49 6773.491
noexperiencebiogas_ifbg | 2836.663 800.9085 3.54 0.000 1266.911 4406.415
alternisrocket | -2768.394 918.7587 -3.01 0.003 -4569.128 -967.6605
maint_higher | 608.8961 404.4207 1.51 0.132 -183.7539 1401.546
smoke_lower | 4784.997 456.2536 10.49 0.000 3890.757 5679.238
fuelefficiency | 3074.038 320.1424 9.60 0.000 2446.57 3701.505
negprice | -7.587495 .144576 -52.48 0.000 -7.870858 -7.304131
--------------------------+----------------------------------------------------------------
SD |
goodexperiencebiogas_ifbg | 5654.514 831.5416 6.80 0.000 4024.723 7284.306
badexperiencebiogas_ifbg | 4184.776 558.7539 7.49 0.000 3089.639 5279.914
noexperiencebiogas_ifbg | -6880.632 763.6405 -9.01 0.000 -8377.339 -5383.924
alternisrocket | -6146.829 654.2155 -9.40 0.000 -7429.068 -4864.59
maint_higher | -3213.146 591.0612 -5.44 0.000 -4371.604 -2054.687
smoke_lower | 1388.787 384.9297 3.61 0.000 634.3389 2143.236
fuelefficiency | -520.4972 379.1623 -1.37 0.170 -1263.642 222.6472
negprice | .7824771 .1572334 4.98 0.000 .4743053 1.090649
-------------------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive
. estimates store MXL_exp
. estimates restore MXL_exp
(results MXL_exp are active now)
. mixlbeta goodexperiencebiogas_ifbg badexperiencebiogas_ifbg noexperiencebiogas_ifbg alternisrocket ma
> int_higher smoke_lower fuelefficiency, nrep(1000) saving(wtpexp_distribution) replace
(file wtpexp_distribution.dta not found)
file wtpexp_distribution.dta saved
期刊排版
（完）
