FTFelix Thoemmes

Felix Thoemmes

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Felix Thoemmes has a total of 51 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Systematic Literature Review of Causal Mediation Analysis in Social Psychology

    by Jiwoo Kim, Felix Thoemmes

    Mediation analysis has been widely used in social psychology to investigate the mechanisms linking independent and dependent variables. However, conducting a proper mediation analysis is challenging, as its causal interpretation requires strong assumptions and careful study design. This systematic literature review examined mediation studies published in the Journal of Personality and Social Psychology and provides recommendations for applied researchers conducting causal mediation analysis. Results revealed a near-exclusive reliance on linear regression frameworks, with no study incorporating treatment and mediator interactions or distinguishing between causally defined natural and controlled direct effects. Only a few studies conducted sensitivity analyses. Temporal ordering was also inconsistently reported, and in rare cases the order of mediator and dependent variable was reversed, raising concerns about the validity of the mediation model. These patterns highlight a persistent gap between methodological recommendations and current research practices. We provide recommendations to strengthen study design, clarify assumptions, and adopt more rigorous analytic strategies to enhance the credibility of mediation research.

  • Open Access

    Revisiting equivalent structural models through the mediation models’ Markov equivalence class

    by Dakota W. Cintron, Felix J. Thoemmes

    Rational inquiry in psychology aims not just to describe phenomena but also to explain them. Researchers continue to use the classic trivariate mediation model as a key tool in this pursuit. The problem of equivalent structural models is revisited as a Markov equivalence class (MEC) ( Andersson et al., 1997 ), illustrated using the trivariate mediation model. We highlight how the MEC complicates causal interpretation of the classic mediation model. To illustrate this, we examine two recent articles from a widely read psychology journal that apply the classic trivariate mediation model, one demonstrating strong causal control, the other weak causal control. We also introduce an R Shiny application that helps users understand this linkage and explore how causal assumptions shape the number of equivalent models in the MEC of a mediation model.

  • Open Access

    Bias and sensitivity analyses for linear front-door models

    by Felix Thoemmes, Yongnam Kim

    The front-door model allows unbiased estimation of a total effect in the presence of unobserved confounding. This guarantee of unbiasedness hinges on a set of assumptions that can be violated in practice. We derive formulas that quantify the amount of bias for specific violations, and contrast them with bias that would be realized from a naive estimator of the effect. Some violations result in simple, monotonic increases in bias, while others lead to more complex bias, consisting of confounding bias, collider bias, and bias amplification. In some instances, these sources of bias can (partially) cancel each other out. We present ways to conduct sensitivity analyses for all violations, and provide code that performs sensitivity analyses for the linear front-door model. We finish with an applied example of the effect of math self-efficacy on educational achievement.

  • Derailment and depression in college: Tests of 3-year predictive capacity and moderation by self-reflection, brooding, perfectionism, and cognitive flexibility.

    by Kaylin Ratner, Anthony L. Burrow, Jane Mendle, Felix Thoemmes

    While rich with opportunities for self-exploration, the transition to and through college is stressful, often associated with the onset or exacerbation of mental illness. Attending to these characteristics, this preregistered study asked whether derailment-or difficulties reconciling perceived identity change-in freshman year predicts senior depressive symptoms, and how individual risks for depression relate to this association. Derailment and depressive symptoms evidenced significant 3-year stability, and these constructs had positive cross-sectional associations in both freshman and senior year. Freshman derailment failed to predict senior depressive symptoms for the average student, but individual differences in self-reflection moderated the association: freshman derailment positively predicted senior depression among those lowest in self-reflection. Together, this study suggests derailment and depressive symptoms are consistently related at critical points of transition, and some individual differences in cognition may help predict their long-term association. While useful for understanding nuances between derailment and depression, these findings also inform ways of attending to and supporting college students through periods of transition. (PsycInfo Database Record (c) 2023 APA, all rights reserved).


  • Invariance of the Derailment Scale-6: Testing the measurement and correlates of derailment across adulthood.

    by Kaylin Ratner, Anthony L. Burrow, Felix Thoemmes, Jane Mendle

    Derailment, a newly mapped construct describing difficulty reconciling perceived changes in identity and self-direction, offers researchers a unique lens for understanding adjustment. As identity development is a lifelong process, establishing the impartiality of the primary measure for derailment across a broad age range is necessary. Across a cross-sectional sample of emerging, young, and midlife adults (N = 1,420), we tested preregistered hypotheses concerning whether the Derailment Scale-6 (DS-6) demonstrates measurement invariance, and the extent to which levels and correlates of derailment differ with age. The DS-6 exhibited configural, metric, and partial scalar invariance across the three age groups. In terms of effect size, statistical differences detected at the scalar level were hardly larger than negligible noninvariance effects, suggesting in sum that the DS-6 is adequate for assessing and comparing derailment among those roughly 18-70 years. In general, derailment was related to greater distress and identity exploration, and lower identity commitment, sense of purpose in life, and future time perspective (FTP). We further found that derailment shares a modest negative association with age; it is more positively associated with identity exploration among younger individuals; and it is more negatively associated with identity commitment, sense of purpose in life, and FTP among older individuals. With the broadest age range and most highly powered set of tests studying derailment to date, this investigation bolsters confidence in our tools for assessing derailment and has taken the initial steps toward sketching the form and function of perceived identity change over the lifespan. Public Significance Statement Feeling "off-course" in life (i.e., derailment) is associated with greater concurrent psychological distress, lower identity commitment and purpose in life, and the sense that one is running out of time in their life. This study demonstrates that the tool for assessing derailment is appropriate for use among, and the comparison of, those roughly 18 to 70-years-old. Importantly, some adjustment correlates of derailment may strengthen or weaken depending upon age.


  • Mediation Analysis Revisited Again

    by Felix Thoemmes, Gunnar Lemmer

    In a mediation model the effect of a dependent variable (DV) on an outcome is (partially) due to the DV's effect on one or multiple mediator(s) that consequently have an effect on the outcome. The use of such models as the theoretical background guiding empirical studies is widespread. Mediation models are fundamentally causal models that specify causal sequences. Unfortunately, the necessary causal assumptions are in practice often violated. In the current paper, we discuss possible improvements of causal mediation analyses, and highlight some potential pitfalls. We discuss the benefits gained by analyzing indirect effect between latent variables specified with measurement models. The validity of statistical findings can also be improved by using experimental designs. We discuss the cross-over design and the cross-over encouragement design, and how they can help improve causal conclusions. We also discuss recent advances on sensitivity analyses in the context of mediation models. Specifically, we explain how this analysis can be used to argue for the severity of unobserved confounding. Lastly, we discuss the practice of reversing the direction of the arrow between variables in a mediation model. We argue that if such reversals result in equivalent models, this practice cannot be recommended.


  • The Analysis of the Regression-Discontinuity Design in R

    by Felix Thoemmes, Wang Liao, Ze Jin

    This article describes the analysis of regression-discontinuity designs (RDDs) using the R packages rdd, rdrobust, and rddtools. We discuss similarities and differences between these packages and provide directions on how to use them effectively. We use real data from the Carolina Abecedarian Project to show how an analysis of an RDD can be performed from start to finish.


  • Reversing Arrows in Mediation Models Does Not Distinguish Plausible Models

    by Felix Thoemmes

    Reversing arrows in the classic tri-variate X-M-Y mediation models as a test to check whether one mediation model is superior to another is inadmissible. Presenting evidence that one tri-variate mediation model yields a significant indirect effect, whereas one with some reversed arrows does not, is not proof or even evidence that one model should be preferred. In fact, the significance of the indirect or any other effect can never be used to infer whether one model should be preferred over another, if the models are in the same so-called equivalence class. The practice of running several mediation models with reversed arrows to decide which model to prefer should be abandoned. The only way to choose among equivalent models is through assumptions that are either fulfilled by design features or invoked based on theory. Similar arguments about reversing arrows in mediation models have been made before, but this current work is the first to derive this result analytically for the complete (Markovian) equivalence class of the tri-variate mediation model.


  • Empirical evaluation of directional-dependence tests

    by Felix Thoemmes

    Testing of directional dependence is a method to infer causal direction that recently has attracted some attention. Previous examples by e.g. von Eye and DeShon (2012a) and extensive simulation studies by Pornprasertmanit and Little (2012) have demonstrated that under specific assumptions, directional-dependence tests can recover the true causal direction between two variables. Simulation results are important in the evaluation of any statistical method, but they are necessarily less complex than real data that come with potential irregularities (e.g. departures from linearity, presence of confounders, etc.). In this article, we evaluate the performance of directional-dependence tests using benchmark data consisting of 65 variable pairs with known causal order. We find that between 21% and 43% of all cases are correctly classified using different directional-dependence tests that rely on differences in skew, kurtosis, or a combined measure. We then examine some of the assumptions of the directional-dependence test and find that for virtually all variable pairs, some assumptions are violated. When only pairs in which assumptions are fulfilled are selected, performance of all directional-dependence tests improves. We probe whether particular features of the variable pairs impact whether a test yields a correct or incorrect result, but find no strong predictors. Our findings provide a complimentary picture to previously conducted simulation studies, and highlight the fact that directional-dependence tests are prone to causal classification errors when key assumptions are violated. Such violations are potentially common in real data.