Mediation and Moderation Analysis
Mediation, moderation and moderated mediation
We estimate mediation, moderation and moderated-mediation models with bootstrap confidence intervals, and interpret them correctly.
Indirect effects are among the most frequently misreported results in the social sciences. The two recurring problems are inference drawn from the Sobel test rather than bootstrapped intervals, and causal language applied to cross-sectional data.
\nWe estimate the models with bias-corrected bootstrap intervals, report the indirect effect with its interval rather than a chain of significant paths, and state plainly what the design can and cannot support. Where your design is within-subject or repeated-measures, PROCESS is not appropriate and we use MEMORE instead.
\nSoftware: PROCESS, MEMORE, Mplus or R, according to the design.
Models We Estimate
Simple and multiple mediation
Single and parallel mediators, and serial models where the theory supports an ordering.
Moderation
Interaction effects with proper centring, simple slopes and regions of significance.
Moderated mediation
Conditional indirect effects with the index of moderated mediation reported.
How the Process Works
- Initial assessmentYou share your research question, your dataset and your target journal. Together we decide which analysis is appropriate.
- Proposal and planScope, timeline and fee are agreed, and what you will receive is set out in writing.
- AnalysisAssumption checks, model specification and estimation, with interim findings shared with you.
- Delivery and handoverAPA-format tables, a written interpretation and a handover conversation. We do not request authorship for standard analysis work.
Request This Analysis
Send us your model diagram or a description of the variables and design, and we will tell you which estimator fits.
Or write to us directly: info@badeinstitute.com
