Investigating treatment mediators when simple random assignment to a control group is not possible

Brian Doss, David C. Atkins

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

As the study of psychological treatments begins to mature beyond questions of efficacy, investigators are focusing more on treatment mediators in an effort to aid revision and dissemination of existing treatments. Often, such studies are conducted without wait-list or no-treatment control groups. However, designs without a control group constrain the examination of treatment mediators and preclude statements about the overall effects of treatment. We review the limitations of treatment-treatment and single-treatment designs for studying mediation and suggest design modifications to strengthen future investigations. Specifically, we present the benefits and limitations of regression discontinuity and interrupted time series in testing mediation when ethical or organizational considerations preclude simple random assignment to a no-treatment condition.

Original languageEnglish
Pages (from-to)321-336
Number of pages16
JournalClinical Psychology: Science and Practice
Volume13
Issue number4
DOIs
StatePublished - Dec 1 2006
Externally publishedYes

Fingerprint

Control Groups
Research Personnel
Psychology
Interrupted Time Series Analysis

Keywords

  • Control group
  • Mediation
  • Regression discontinuity
  • Study design
  • Treatment

ASJC Scopus subject areas

  • Psychology(all)
  • Clinical Psychology

Cite this

Investigating treatment mediators when simple random assignment to a control group is not possible. / Doss, Brian; Atkins, David C.

In: Clinical Psychology: Science and Practice, Vol. 13, No. 4, 01.12.2006, p. 321-336.

Research output: Contribution to journalArticle

@article{b9e0ae3e1b664dc3b9a49894c952bebd,
title = "Investigating treatment mediators when simple random assignment to a control group is not possible",
abstract = "As the study of psychological treatments begins to mature beyond questions of efficacy, investigators are focusing more on treatment mediators in an effort to aid revision and dissemination of existing treatments. Often, such studies are conducted without wait-list or no-treatment control groups. However, designs without a control group constrain the examination of treatment mediators and preclude statements about the overall effects of treatment. We review the limitations of treatment-treatment and single-treatment designs for studying mediation and suggest design modifications to strengthen future investigations. Specifically, we present the benefits and limitations of regression discontinuity and interrupted time series in testing mediation when ethical or organizational considerations preclude simple random assignment to a no-treatment condition.",
keywords = "Control group, Mediation, Regression discontinuity, Study design, Treatment",
author = "Brian Doss and Atkins, {David C.}",
year = "2006",
month = "12",
day = "1",
doi = "10.1111/j.1468-2850.2006.00045.x",
language = "English",
volume = "13",
pages = "321--336",
journal = "Clinical Psychology: Science and Practice",
issn = "0969-5893",
publisher = "Wiley-Blackwell",
number = "4",

}

TY - JOUR

T1 - Investigating treatment mediators when simple random assignment to a control group is not possible

AU - Doss, Brian

AU - Atkins, David C.

PY - 2006/12/1

Y1 - 2006/12/1

N2 - As the study of psychological treatments begins to mature beyond questions of efficacy, investigators are focusing more on treatment mediators in an effort to aid revision and dissemination of existing treatments. Often, such studies are conducted without wait-list or no-treatment control groups. However, designs without a control group constrain the examination of treatment mediators and preclude statements about the overall effects of treatment. We review the limitations of treatment-treatment and single-treatment designs for studying mediation and suggest design modifications to strengthen future investigations. Specifically, we present the benefits and limitations of regression discontinuity and interrupted time series in testing mediation when ethical or organizational considerations preclude simple random assignment to a no-treatment condition.

AB - As the study of psychological treatments begins to mature beyond questions of efficacy, investigators are focusing more on treatment mediators in an effort to aid revision and dissemination of existing treatments. Often, such studies are conducted without wait-list or no-treatment control groups. However, designs without a control group constrain the examination of treatment mediators and preclude statements about the overall effects of treatment. We review the limitations of treatment-treatment and single-treatment designs for studying mediation and suggest design modifications to strengthen future investigations. Specifically, we present the benefits and limitations of regression discontinuity and interrupted time series in testing mediation when ethical or organizational considerations preclude simple random assignment to a no-treatment condition.

KW - Control group

KW - Mediation

KW - Regression discontinuity

KW - Study design

KW - Treatment

UR - http://www.scopus.com/inward/record.url?scp=34248344994&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34248344994&partnerID=8YFLogxK

U2 - 10.1111/j.1468-2850.2006.00045.x

DO - 10.1111/j.1468-2850.2006.00045.x

M3 - Article

AN - SCOPUS:34248344994

VL - 13

SP - 321

EP - 336

JO - Clinical Psychology: Science and Practice

JF - Clinical Psychology: Science and Practice

SN - 0969-5893

IS - 4

ER -