Increasing Statistical Power in Mediation Models Without Increasing Sample Size

Matthew S. Fritz, Matthew G. Cox, David Mackinnon

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

Inadequate statistical power to detect treatment effects in health research is a problem that is compounded when testing for mediation. In general, the preferred strategy for increasing power is to increase the sample size, but there are many situations where additional participants cannot be recruited, necessitating the use of other methods to increase statistical power. Many of these other strategies, commonly applied to analysis of variance and multiple regression models, can be applied to mediation models with similar results. Additional predictors or blocking variables will increase or decrease statistical power, however, depending on whether these variables are related to the mediator, the outcome, or both. The effect of these two methods on the power for tests of mediation is illustrated through the use of simulations. Implications for health researchers using these methods are discussed.

Original languageEnglish (US)
Pages (from-to)343-366
Number of pages24
JournalEvaluation and the Health Professions
Volume38
Issue number3
DOIs
StatePublished - Sep 18 2015

Fingerprint

Sample Size
Health
Analysis of Variance
Research Personnel
Research

Keywords

  • mediation
  • sample size
  • statistical power

ASJC Scopus subject areas

  • Health Policy

Cite this

Increasing Statistical Power in Mediation Models Without Increasing Sample Size. / Fritz, Matthew S.; Cox, Matthew G.; Mackinnon, David.

In: Evaluation and the Health Professions, Vol. 38, No. 3, 18.09.2015, p. 343-366.

Research output: Contribution to journalArticle

@article{b3ba7a7abed74de7b957e9a7e3c8d275,
title = "Increasing Statistical Power in Mediation Models Without Increasing Sample Size",
abstract = "Inadequate statistical power to detect treatment effects in health research is a problem that is compounded when testing for mediation. In general, the preferred strategy for increasing power is to increase the sample size, but there are many situations where additional participants cannot be recruited, necessitating the use of other methods to increase statistical power. Many of these other strategies, commonly applied to analysis of variance and multiple regression models, can be applied to mediation models with similar results. Additional predictors or blocking variables will increase or decrease statistical power, however, depending on whether these variables are related to the mediator, the outcome, or both. The effect of these two methods on the power for tests of mediation is illustrated through the use of simulations. Implications for health researchers using these methods are discussed.",
keywords = "mediation, sample size, statistical power",
author = "Fritz, {Matthew S.} and Cox, {Matthew G.} and David Mackinnon",
year = "2015",
month = "9",
day = "18",
doi = "10.1177/0163278713514250",
language = "English (US)",
volume = "38",
pages = "343--366",
journal = "Evaluation and the Health Professions",
issn = "0163-2787",
publisher = "SAGE Publications Inc.",
number = "3",

}

TY - JOUR

T1 - Increasing Statistical Power in Mediation Models Without Increasing Sample Size

AU - Fritz, Matthew S.

AU - Cox, Matthew G.

AU - Mackinnon, David

PY - 2015/9/18

Y1 - 2015/9/18

N2 - Inadequate statistical power to detect treatment effects in health research is a problem that is compounded when testing for mediation. In general, the preferred strategy for increasing power is to increase the sample size, but there are many situations where additional participants cannot be recruited, necessitating the use of other methods to increase statistical power. Many of these other strategies, commonly applied to analysis of variance and multiple regression models, can be applied to mediation models with similar results. Additional predictors or blocking variables will increase or decrease statistical power, however, depending on whether these variables are related to the mediator, the outcome, or both. The effect of these two methods on the power for tests of mediation is illustrated through the use of simulations. Implications for health researchers using these methods are discussed.

AB - Inadequate statistical power to detect treatment effects in health research is a problem that is compounded when testing for mediation. In general, the preferred strategy for increasing power is to increase the sample size, but there are many situations where additional participants cannot be recruited, necessitating the use of other methods to increase statistical power. Many of these other strategies, commonly applied to analysis of variance and multiple regression models, can be applied to mediation models with similar results. Additional predictors or blocking variables will increase or decrease statistical power, however, depending on whether these variables are related to the mediator, the outcome, or both. The effect of these two methods on the power for tests of mediation is illustrated through the use of simulations. Implications for health researchers using these methods are discussed.

KW - mediation

KW - sample size

KW - statistical power

UR - http://www.scopus.com/inward/record.url?scp=84939215299&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84939215299&partnerID=8YFLogxK

U2 - 10.1177/0163278713514250

DO - 10.1177/0163278713514250

M3 - Article

VL - 38

SP - 343

EP - 366

JO - Evaluation and the Health Professions

JF - Evaluation and the Health Professions

SN - 0163-2787

IS - 3

ER -