请教SPSS中的ordinal logistics regression的位置模型和尺度模型

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

各位高手好！兄弟我利用SPSS做有序的LOGISTICS回归，选项中有LOCATION和SCALE两种模型，如果只做LOCATION，模型拟合情况不太好，将同样的变量选择进入SCALE模型后，模型拟合度变好，且输出表格中多了SCALE模型的结果，而且LOCATION模型中的回归结果与仅做LOCATION模型的结果也不一样。我想请问的是，LOCATION和SCALE模型的结果应如何解释，报告回归结果是应看LOCATION模型的系数和显著性，还是要看SCALE模型的系数和显著性？期盼有高人指点。

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏1 回帖

关键词：regression LOGISTICS regressio logistic ordinal logistics 模型 而且 如何

我也想知道啊 期待高手啊.......

I got a idea but I'm not sure whether it is right or not. The coefficients estimated by the common GLM, such as probit, logit, multinomial logit, ordered logit, etc, are "Location/Scale". We cannot estimate the locations and scales separetely, due to the identification problem. That causes many difficulties. For example, we cannot compare the estimated coefficients directly, because the difference between coefficients may be caused by the different scales, rather than the different locations. Some researches figured out some ways to overcome this problem. Now, under certain assumptions, location and scale can be estimated separately. In STATA, we use the commands - oglm or hetprob - to fullfil this goals. Therefore, I guess the Location and Scale in the SPSS are used to fullfil similar goals.

If you are want to learn about this issue a bit more, there are some reference you may be interested in.

Allison, Paul D. . 1999. “Comparing Logit and Probit Coefficients Across Groups.” Sociological Methods & Research  28(2): 186-208.
Williams, Richard. 2009. “Using Heterogeneous Choice Models to Compare Logit and Probit Coefficients across Groups.” Sociological Methods & Research 37(4): 531-559.
Buis, Maarten L.. 2011. “The Consequences of Unobserved Heterogeneity in a Sequential Logit Model.” Research in Social Stratification and Mobility 29(3): 247-262.
Logistic Regression: Why We Cannot Do What We Think We Can Do, and What We Can Do about It

Hope they are helpful, and hope I made a right guess.

Kirin_guess, thanks for your kind help. I appreciate it.