Testing a nonlinear regression specification: A nonregular case

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

A statistical test of whether an additive nonlinear term in the response function should be omitted from a nonlinear regression specification is considered. The regularity conditions used to obtain the asymptotic distributions of the customary test statistics are violated when the null hypothesis of omission is true. Moreover, standard iterative algorithms are likely to perform poorly when the data support the null hypothesis. Methods designed to circumvent these mathematical and computational difficulties are described and illustrated.

Original languageEnglish (US)
Pages (from-to)523-530
Number of pages8
JournalJournal of the American Statistical Association
Volume72
Issue number359
DOIs
StatePublished - Jan 1 1977

Fingerprint

Nonlinear Regression
Null hypothesis
Specification
Testing
Response Function
Statistical test
Regularity Conditions
Iterative Algorithm
Asymptotic distribution
Test Statistic
Likely
Term
Regularity
Statistical tests
Nonlinear regression
Test statistic
Standards

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

@article{d54e5b08596e43fd91d52ea3d04d33d4,
title = "Testing a nonlinear regression specification: A nonregular case",
abstract = "A statistical test of whether an additive nonlinear term in the response function should be omitted from a nonlinear regression specification is considered. The regularity conditions used to obtain the asymptotic distributions of the customary test statistics are violated when the null hypothesis of omission is true. Moreover, standard iterative algorithms are likely to perform poorly when the data support the null hypothesis. Methods designed to circumvent these mathematical and computational difficulties are described and illustrated.",
author = "Gallant, {Andrew Ronald}",
year = "1977",
month = "1",
day = "1",
doi = "10.1080/01621459.1977.10480606",
language = "English (US)",
volume = "72",
pages = "523--530",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",
number = "359",

}

Testing a nonlinear regression specification : A nonregular case. / Gallant, Andrew Ronald.

In: Journal of the American Statistical Association, Vol. 72, No. 359, 01.01.1977, p. 523-530.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Testing a nonlinear regression specification

T2 - A nonregular case

AU - Gallant, Andrew Ronald

PY - 1977/1/1

Y1 - 1977/1/1

N2 - A statistical test of whether an additive nonlinear term in the response function should be omitted from a nonlinear regression specification is considered. The regularity conditions used to obtain the asymptotic distributions of the customary test statistics are violated when the null hypothesis of omission is true. Moreover, standard iterative algorithms are likely to perform poorly when the data support the null hypothesis. Methods designed to circumvent these mathematical and computational difficulties are described and illustrated.

AB - A statistical test of whether an additive nonlinear term in the response function should be omitted from a nonlinear regression specification is considered. The regularity conditions used to obtain the asymptotic distributions of the customary test statistics are violated when the null hypothesis of omission is true. Moreover, standard iterative algorithms are likely to perform poorly when the data support the null hypothesis. Methods designed to circumvent these mathematical and computational difficulties are described and illustrated.

UR - http://www.scopus.com/inward/record.url?scp=0009921231&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0009921231&partnerID=8YFLogxK

U2 - 10.1080/01621459.1977.10480606

DO - 10.1080/01621459.1977.10480606

M3 - Article

AN - SCOPUS:0009921231

VL - 72

SP - 523

EP - 530

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

IS - 359

ER -