Correlation estimation with singly truncated bivariate data

Jongho Im, Eunyong Ahn, Namseon Beck, Jae Kwang Kim, Taesung Park

Research output: Contribution to journalArticle

Abstract

Correlation coefficient estimates are often attenuated for truncated samples in the sense that the estimates are biased towards zero. Motivated by real data collected in South Sudan, we consider correlation coefficient estimation with singly truncated bivariate data. By considering a linear regression model in which a truncated variable is used as an explanatory variable, a consistent estimator for the regression slope can be obtained from the ordinary least squares method. A consistent estimator of the correlation coefficient is then obtained by multiplying the regression slope estimator by the variance ratio of the two variables. Results from two limited simulation studies confirm the validity and robustness of the proposed method. The proposed method is applied to the South Sudanese children's anthropometric and nutritional data collected by World Vision.

Original languageEnglish
Pages (from-to)1977-1988
Number of pages12
JournalStatistics in Medicine
Volume36
Issue number12
DOIs
Publication statusPublished - 2017 May 30

Fingerprint

Correlation coefficient
Consistent Estimator
Linear Models
Slope
Regression
Variance Ratio
Coefficient Estimates
Ordinary Least Squares
Linear Regression Model
Least-Squares Analysis
Least Square Method
Biased
Simulation Study
Robustness
Estimator
Zero
Estimate
Children
Vision
South Sudan

All Science Journal Classification (ASJC) codes

  • Epidemiology
  • Statistics and Probability

Cite this

Im, Jongho ; Ahn, Eunyong ; Beck, Namseon ; Kim, Jae Kwang ; Park, Taesung. / Correlation estimation with singly truncated bivariate data. In: Statistics in Medicine. 2017 ; Vol. 36, No. 12. pp. 1977-1988.
@article{926116fd61154b9a831b07b02f099fea,
title = "Correlation estimation with singly truncated bivariate data",
abstract = "Correlation coefficient estimates are often attenuated for truncated samples in the sense that the estimates are biased towards zero. Motivated by real data collected in South Sudan, we consider correlation coefficient estimation with singly truncated bivariate data. By considering a linear regression model in which a truncated variable is used as an explanatory variable, a consistent estimator for the regression slope can be obtained from the ordinary least squares method. A consistent estimator of the correlation coefficient is then obtained by multiplying the regression slope estimator by the variance ratio of the two variables. Results from two limited simulation studies confirm the validity and robustness of the proposed method. The proposed method is applied to the South Sudanese children's anthropometric and nutritional data collected by World Vision.",
author = "Jongho Im and Eunyong Ahn and Namseon Beck and Kim, {Jae Kwang} and Taesung Park",
year = "2017",
month = "5",
day = "30",
doi = "10.1002/sim.7267",
language = "English",
volume = "36",
pages = "1977--1988",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "12",

}

Im, J, Ahn, E, Beck, N, Kim, JK & Park, T 2017, 'Correlation estimation with singly truncated bivariate data', Statistics in Medicine, vol. 36, no. 12, pp. 1977-1988. https://doi.org/10.1002/sim.7267

Correlation estimation with singly truncated bivariate data. / Im, Jongho; Ahn, Eunyong; Beck, Namseon; Kim, Jae Kwang; Park, Taesung.

In: Statistics in Medicine, Vol. 36, No. 12, 30.05.2017, p. 1977-1988.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Correlation estimation with singly truncated bivariate data

AU - Im, Jongho

AU - Ahn, Eunyong

AU - Beck, Namseon

AU - Kim, Jae Kwang

AU - Park, Taesung

PY - 2017/5/30

Y1 - 2017/5/30

N2 - Correlation coefficient estimates are often attenuated for truncated samples in the sense that the estimates are biased towards zero. Motivated by real data collected in South Sudan, we consider correlation coefficient estimation with singly truncated bivariate data. By considering a linear regression model in which a truncated variable is used as an explanatory variable, a consistent estimator for the regression slope can be obtained from the ordinary least squares method. A consistent estimator of the correlation coefficient is then obtained by multiplying the regression slope estimator by the variance ratio of the two variables. Results from two limited simulation studies confirm the validity and robustness of the proposed method. The proposed method is applied to the South Sudanese children's anthropometric and nutritional data collected by World Vision.

AB - Correlation coefficient estimates are often attenuated for truncated samples in the sense that the estimates are biased towards zero. Motivated by real data collected in South Sudan, we consider correlation coefficient estimation with singly truncated bivariate data. By considering a linear regression model in which a truncated variable is used as an explanatory variable, a consistent estimator for the regression slope can be obtained from the ordinary least squares method. A consistent estimator of the correlation coefficient is then obtained by multiplying the regression slope estimator by the variance ratio of the two variables. Results from two limited simulation studies confirm the validity and robustness of the proposed method. The proposed method is applied to the South Sudanese children's anthropometric and nutritional data collected by World Vision.

UR - http://www.scopus.com/inward/record.url?scp=85013909046&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85013909046&partnerID=8YFLogxK

U2 - 10.1002/sim.7267

DO - 10.1002/sim.7267

M3 - Article

C2 - 28239899

AN - SCOPUS:85013909046

VL - 36

SP - 1977

EP - 1988

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 12

ER -