Estimating Mediated Effects with Survival Data

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Summary. Mediation analyses help identify variables in the causal sequence relating predictor variables to outcome variables . In many studies, outcomes are time until an event occurs and survival analyses are applied . This study examines the point and interval estimates of the mediated effect using two methods of survival analyses : the log-survival time and log-hazard time models . The results show that, under the condition of no censored data, the assumption that mediated effects calculated by the product of coefficients method (a(3) and those calculated by the difference in coefficients method (tr - r') are identical does apply to log-survival time survival analyses but not to log-hazard time survival analyses . The standard error of the mediated effect can be calculated with the delta formula, the second order Taylor series formula, and the unbiased formula . Consistent with ordinary least squares regression, the three formulas yield similar results . Although the logsurvival time model and the log-hazard time model utilize different estimation methods, the results of the significant tests, using the ratio of ap to se,p, were comparable between the two methods . However, the significance tests based on the empirical standard error appear
Original languageEnglish (US)
Title of host publicationNew Developments in Psychometrics
PublisherSpringer Japan
Pages405-412
Number of pages8
DOIs
StatePublished - Apr 30 2003

Publication series

NameNew Developments in Psychometrics

Fingerprint

hazard
estimation method
effect
method
test
product

Cite this

Tein, J-Y., & MacKinnon, D. P. (2003). Estimating Mediated Effects with Survival Data. In New Developments in Psychometrics (pp. 405-412). (New Developments in Psychometrics). Springer Japan. https://doi.org/10.1007/978-4-431-66996-8_46

Estimating Mediated Effects with Survival Data. / Tein, Jenn-Yun; MacKinnon, David P.

New Developments in Psychometrics. Springer Japan, 2003. p. 405-412 (New Developments in Psychometrics).

Research output: Chapter in Book/Report/Conference proceedingChapter

Tein, J-Y & MacKinnon, DP 2003, Estimating Mediated Effects with Survival Data. in New Developments in Psychometrics. New Developments in Psychometrics, Springer Japan, pp. 405-412. https://doi.org/10.1007/978-4-431-66996-8_46
Tein J-Y, MacKinnon DP. Estimating Mediated Effects with Survival Data. In New Developments in Psychometrics. Springer Japan. 2003. p. 405-412. (New Developments in Psychometrics). https://doi.org/10.1007/978-4-431-66996-8_46
Tein, Jenn-Yun ; MacKinnon, David P. / Estimating Mediated Effects with Survival Data. New Developments in Psychometrics. Springer Japan, 2003. pp. 405-412 (New Developments in Psychometrics).
@inbook{6a8e5c86f65545d996c551aa24a52ad1,
title = "Estimating Mediated Effects with Survival Data",
abstract = "Summary. Mediation analyses help identify variables in the causal sequence relating predictor variables to outcome variables . In many studies, outcomes are time until an event occurs and survival analyses are applied . This study examines the point and interval estimates of the mediated effect using two methods of survival analyses : the log-survival time and log-hazard time models . The results show that, under the condition of no censored data, the assumption that mediated effects calculated by the product of coefficients method (a(3) and those calculated by the difference in coefficients method (tr - r') are identical does apply to log-survival time survival analyses but not to log-hazard time survival analyses . The standard error of the mediated effect can be calculated with the delta formula, the second order Taylor series formula, and the unbiased formula . Consistent with ordinary least squares regression, the three formulas yield similar results . Although the logsurvival time model and the log-hazard time model utilize different estimation methods, the results of the significant tests, using the ratio of ap to se,p, were comparable between the two methods . However, the significance tests based on the empirical standard error appear",
author = "Jenn-Yun Tein and MacKinnon, {David P.}",
year = "2003",
month = "4",
day = "30",
doi = "10.1007/978-4-431-66996-8_46",
language = "English (US)",
series = "New Developments in Psychometrics",
publisher = "Springer Japan",
pages = "405--412",
booktitle = "New Developments in Psychometrics",
address = "Japan",

}

TY - CHAP

T1 - Estimating Mediated Effects with Survival Data

AU - Tein, Jenn-Yun

AU - MacKinnon, David P.

PY - 2003/4/30

Y1 - 2003/4/30

N2 - Summary. Mediation analyses help identify variables in the causal sequence relating predictor variables to outcome variables . In many studies, outcomes are time until an event occurs and survival analyses are applied . This study examines the point and interval estimates of the mediated effect using two methods of survival analyses : the log-survival time and log-hazard time models . The results show that, under the condition of no censored data, the assumption that mediated effects calculated by the product of coefficients method (a(3) and those calculated by the difference in coefficients method (tr - r') are identical does apply to log-survival time survival analyses but not to log-hazard time survival analyses . The standard error of the mediated effect can be calculated with the delta formula, the second order Taylor series formula, and the unbiased formula . Consistent with ordinary least squares regression, the three formulas yield similar results . Although the logsurvival time model and the log-hazard time model utilize different estimation methods, the results of the significant tests, using the ratio of ap to se,p, were comparable between the two methods . However, the significance tests based on the empirical standard error appear

AB - Summary. Mediation analyses help identify variables in the causal sequence relating predictor variables to outcome variables . In many studies, outcomes are time until an event occurs and survival analyses are applied . This study examines the point and interval estimates of the mediated effect using two methods of survival analyses : the log-survival time and log-hazard time models . The results show that, under the condition of no censored data, the assumption that mediated effects calculated by the product of coefficients method (a(3) and those calculated by the difference in coefficients method (tr - r') are identical does apply to log-survival time survival analyses but not to log-hazard time survival analyses . The standard error of the mediated effect can be calculated with the delta formula, the second order Taylor series formula, and the unbiased formula . Consistent with ordinary least squares regression, the three formulas yield similar results . Although the logsurvival time model and the log-hazard time model utilize different estimation methods, the results of the significant tests, using the ratio of ap to se,p, were comparable between the two methods . However, the significance tests based on the empirical standard error appear

UR - http://www.mendeley.com/research/estimating-mediated-effects-survival-data

U2 - 10.1007/978-4-431-66996-8_46

DO - 10.1007/978-4-431-66996-8_46

M3 - Chapter

T3 - New Developments in Psychometrics

SP - 405

EP - 412

BT - New Developments in Psychometrics

PB - Springer Japan

ER -