Boolean logit and probit in Stata

被引：7

作者：

Braumoeller, Bear F. ^{[1
]}

机构：

[1] Harvard Univ, Dept Govt, Cambridge, MA 02138 USA

来源：

STATA JOURNAL | 2004年 / 4卷 / 04期

关键词：

st0078; mlboolean; dichotomous dependent variable; Boolean; logit; probit; multiple causal paths; complexity; random utility;

D O I：

10.1177/1536867X0400400406

中图分类号：

O1 [数学]; C [社会科学总论];

学科分类号：

03 ; 0303 ; 0701 ; 070101 ;

摘要：

This paper introduces new statistical models, Boolean logit and probit, that allow researchers to model binary outcomes as the results of Boolean interactions among independent causal processes. Each process (or "causal path") is modeled as the unobserved outcome in a standard logit or probit equation, and the dependent variable is modeled as the observed product of their Boolean interaction. Up to five causal paths can be modeled, in any combination-A and B and C produce Y, A and (B or [C and D]) produce Y, etc.

引用

页码：436 / 441

页数：6