Estimation Accuracies of Total Effects Using Supplementary Variables  

Oleh:

Kuroki, Manabu ; Hayashi, Takahiro

Jenis: Article from Proceeding
Dalam koleksi: 12th ANQ Congress in Singapore, 5-8 Agustus 2014, page 1-12.
Topik: Causal risk difference; Intermediate variable; Multicollinearity; Recursive regression model; Total effect
Fulltext: QP1-1.1-P0240.pdf (224.3KB)

Isi artikel

This paper considers the problem of estimating total effects using a set of supplementary variables, including intermediate variables. When a univariate treatment is linearly associated with a response through a univariate intermediate variable, the regression coefficient of the treatment on the response in a single regression model can be rendered more accurately by a recursive regression model based on the intermediate variable, developed from asymptotic theory. However, such situations are frequently not found in practice. This paper focuses on situations in which a set of treatments is associated with a response through a set of supplementary variables in linear as well as discrete models. We show that the causal effect can be estimated more accurately from the supplementary variables. Unlike previous studies, our results can be derived without assuming Gaussian error terms in linear models or dichotomous variables in discrete models. From our results, we can judge from graph structures the situations under which the causal effect can be estimated more accurately by supplementary variables, and reliably evaluate the total effects from observed data

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali