The design and analysis of state-trace experiments

Melissa Prince, Scott Brown, Andrew Heathcote

    Research output: Contribution to journalArticlepeer-review

    48 Citations (Scopus)


    State-trace analysis (Bamber, 1979) addresses a question of interest in many areas of psychological research: Does 1 or more than 1 latent (i.e., not directly observed) variable mediate an interaction between 2 experimental manipulations? There is little guidance available on how to design an experiment suited to state-trace analysis, despite its increasing use, and existing statistical methods for state-trace analysis are problematic. We provide a framework for designing and refining a state-trace experiment and statistical procedures for the analysis of accuracy data using Klugkist, Kato, and Hoijtink's (2005) method of estimating Bayes factors. The statistical procedures provide estimates of the evidence favoring 1 versus more than 1 latent variable, as well as evidence that can be used to refine experimental methodology.

    Original languageEnglish
    Pages (from-to)78-99
    Number of pages22
    JournalPsychological Methods
    Issue number1
    Publication statusPublished - Mar 2012


    • Bayes factors
    • Dimensional analysis
    • Inequality constraints
    • Model selection
    • State-trace analysis


    Dive into the research topics of 'The design and analysis of state-trace experiments'. Together they form a unique fingerprint.

    Cite this