CoChat
Sign Up
Cited by 17,383Advanced Causal Inference TechniquesPsychometric Methodologies and TestingStatistical Methods and Bayesian Inference

SPSS and SAS procedures for estimating indirect effects in simple mediation models

Open access
Preacher, Kristopher J., Hayes, Andrew F.·2004·Behavior Research Methods, Instruments, & Computers
DOI 10.3758/bf03206553
Open this paper in CoChat — free accountOpen PDF

Abstract

Researchers often conduct mediation analysis in order to indirectly assess the effect of a proposed cause on some outcome through a proposed mediator. The utility of mediation analysis stems from its ability to go beyond the merely descriptive to a more functional understanding of the relationships among variables. A necessary component of mediation is a statistically and practically significant indirect effect. Although mediation hypotheses are frequently explored in psychological research, formal significance tests of indirect effects are rarely conducted. After a brief overview of mediation, we argue the importance of directly testing the significance of indirect effects and provide SPSS and SAS macros that facilitate estimation of the indirect effect with a normal theory approach and a bootstrap approach to obtaining confidence intervals, as well as the traditional approach advocated by Baron and Kenny (1986). We hope that this discussion and the macros will enhance the frequency of formal mediation tests in the psychology literature. Electronic copies of these macros may be downloaded from the Psychonomic Society's Web archive at www.psychonomic.org/archive/.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.3758/bf03206553

Cites

46 of 46 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • The study results demonstrate mediation only according to the original Baron and Kenny approach, but not according to more recent discussions which would require an independent interaction term to be significant.

    “one can argue that our study results demonstrate mediation only according to the original Baron and Kenny approach, but not according to more recent discussions [74,75], which would require an independent interaction term to be significant.”

    From Exploring emotional well-being, spiritual, religious and personal beliefs and telomere length in chronic pain patients—A pilot study with cross-sectional design (2024) · cites this paper as [74]

    In “The Need for a Formal Test” (page unavailable), Preacher and Hayes contrast the Baron and Kenny criteria with a formal test of the indirect effect. They say the relevant significance test concerns the product of the mediation paths (ab), not an independent interaction term. Their paper also cannot establish whether the citing study’s own results satisfy either approach. (LLM, conf=0.88)

    Checked by a reasoning model — may be wrong.

    Experimental — accuracy not yet measured

  • The multiple mediation analysis used 1,000 bootstrap samples and 95% bias-corrected confidence intervals around the standardised estimate of the effect.

    “The multiple mediation analysis used 1,000 bootstrap samples and 95 % bias-corrected confidence intervals (CI) around the standardised estimate of the effect as recommend by Preacher and Hayes (2004).”

    From The role of experiential avoidance and cognitive fusion in the development of anxiety symptoms among family carers of people with dementia (2024) · cites this paper as Preacher and Hayes (2004)

    In “Estimating the Size and Significance of the Indirect Effect” (page unavailable), the paper illustrates 1,000 bootstrap samples and a 95% confidence interval for an indirect effect. It presents 1,000 as an illustrative choice, not a recommendation, and describes percentile bounds rather than bias-corrected bounds. Its procedures address simple mediation and do not support the claim about multiple mediation or a standardised estimate. (LLM, conf=0.91)

    Experimental — accuracy not yet measured

  • Mediation was deemed significant if the 95% bias corrected and accelerated (BCa) bootstrap confidence intervals for the indirect effects based on 20000 bootstrapped samples did not include zero.

    “Mediation was deemed as significant if the 95% bias corrected and accelerated (BCa) bootstrap confidence intervals (CI) for the indirect effects based on 20000 bootstrapped samples did not include zero (see Preacher & Hayes, 2004).”

    From The relationship between low levels of mindfulness skills and pathological worry: The mediating role of psychological inflexibility. [La relación entre habilidades de mindfulness y preocupación patológica: El rol mediador de la inflexibilidad psicológica] (2014) · cites this paper as Preacher & Hayes, 2004

    In “Estimating the Size and Significance of the Indirect Effect” (page unavailable), the paper says an indirect effect is significant when its 95% bootstrap confidence interval excludes zero. It illustrates percentile-based intervals using 1,000 resamples, not BCa intervals or a prescribed 20,000 resamples. The citing authors’ choice of 20,000 resamples may be their own, but that detail is not established by this source. (LLM, conf=0.89)

    Experimental — accuracy not yet measured

  • Bootstrapping is a nonparametric procedure that acknowledges the fact that the product distribution often violates normality.

    “Bootstrapping is a nonparametric procedure that acknowledges this fact [61, 62].”

    From Acceptance as a Mediator for Change in Acceptance and Commitment Therapy for Persons with Chronic Pain? (2016) · cites this paper as [61]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. [19]19. Lockwood, C. M., & MacKinnon, D. P. (1998). Bootstrapping the standard error of the mediated effect. Proceedings of the 23rd annual meeting of SAS Users Group International (pp. 997-1002).

Cited by

Sorted by citations

Related

Authors
Preacher, K. J., & Hayes, A. F.
Affiliations
  • University of North Carolina at Chapel Hill (US)
  • The Ohio State University (US)
Corresponding author
Preacher, Kristopher J.
ORCID
Venue
Behavior Research Methods, Instruments, & Computers
Publisher
Springer Science and Business Media LLC
Published
2004-11
Volume
36
Issue
4
Pages
717-731
ISSN
  • 0743-3808 (ISSN-L)
  • 1532-5970
DOI
10.3758/bf03206553
OpenAlex
W1995031937
Semantic Scholar
8432aa15730df7432e1ca82ce48f0a90da2a86ec
Cited by
17,383
References
41
Citations by year
2026: 642 · 2025: 893 · 2024: 965 · 2023: 974 · 2022: 1187 · 2021: 1157 · 2020: 1100 · 2019: 1104 · 2018: 1149 · 2017: 1053 · 2016: 1089 · 2015: 1089 · 2014: 1063 · 2013: 988 · 2012: 840
FWCI
84.75
Type
journal article
Language
en
Open access
Licence
www.springer.com/tdm (tdm)
Locations
  • Behavior Research Methods, Instruments, & Computers (publishedVersion)
  • PubMed (publishedVersion)
  • KU ScholarWorks (University of Kansas) (submittedVersion, other-oa)
Topics
  • Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
  • Social Sciences › Decision Sciences › Management Science and Operations Research › Psychometric Methodologies and Testing
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
Keywords
mediation analysis, indirect effects, bootstrap confidence intervals, SAS macro
MeSH terms
Humans, Models, Psychological, Psychology

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review