Some consequences of model misspecification for t testing in a structural equation

Christopher L. Skeels*, Murray D. Smith

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

This paper examines the distribution theory relevant to statistics obtained from a single structural equation from a system of simultaneous equations. Attention is focused on t statistics in misspecified models which have been estimated by instrumental variable techniques.

Original languageEnglish
Pages (from-to)251-271
Number of pages21
JournalJournal of Statistical Planning and Inference
Volume50
Issue number2
DOIs
Publication statusPublished - 01 Mar 1996
Externally publishedYes

Keywords

  • Conditional normal theory
  • Misspecified structural equations
  • Omitted variables
  • T tests

Fingerprint

Dive into the research topics of 'Some consequences of model misspecification for t testing in a structural equation'. Together they form a unique fingerprint.

Cite this