Daniel Heck
@danielheck.bsky.social
Professor of Psychological Methods @Phillips-Universtät Marburg
Mathematical psychology | Cognitive modeling | Psychometrics | Bayesian statistics
Personal: www.dwheck.de
Team: https://www.uni-marburg.de/de/fb04/team-heck
Mathematical psychology | Cognitive modeling | Psychometrics | Bayesian statistics
Personal: www.dwheck.de
Team: https://www.uni-marburg.de/de/fb04/team-heck
Reposted by Daniel Heck
Does it make sense to preregister simulation studies?
This question has sparked a lot of debate.
▶️We* work through the why, when, and how
▶️We discuss different phases of methodological research to clarify where preregistration might (or might not) add value
📝 Preprint: doi.org/10.31234/osf...
This question has sparked a lot of debate.
▶️We* work through the why, when, and how
▶️We discuss different phases of methodological research to clarify where preregistration might (or might not) add value
📝 Preprint: doi.org/10.31234/osf...
February 4, 2026 at 10:40 AM
Does it make sense to preregister simulation studies?
This question has sparked a lot of debate.
▶️We* work through the why, when, and how
▶️We discuss different phases of methodological research to clarify where preregistration might (or might not) add value
📝 Preprint: doi.org/10.31234/osf...
This question has sparked a lot of debate.
▶️We* work through the why, when, and how
▶️We discuss different phases of methodological research to clarify where preregistration might (or might not) add value
📝 Preprint: doi.org/10.31234/osf...
Yes, it's a classic paper experimental researchers should know about.
I was surprised that ANOVA of proportions has still been used relatively often in the literature on the truth effect.
I was surprised that ANOVA of proportions has still been used relatively often in the literature on the truth effect.
February 2, 2026 at 3:02 PM
Yes, it's a classic paper experimental researchers should know about.
I was surprised that ANOVA of proportions has still been used relatively often in the literature on the truth effect.
I was surprised that ANOVA of proportions has still been used relatively often in the literature on the truth effect.
Finally published in Memory & Cognition:
Multinomial models of the repetition-based truth effect: Investigating the role of prior knowledge
link.springer.com/article/10.3...
Multinomial models of the repetition-based truth effect: Investigating the role of prior knowledge
link.springer.com/article/10.3...
February 2, 2026 at 3:00 PM
Finally published in Memory & Cognition:
Multinomial models of the repetition-based truth effect: Investigating the role of prior knowledge
link.springer.com/article/10.3...
Multinomial models of the repetition-based truth effect: Investigating the role of prior knowledge
link.springer.com/article/10.3...
Great, thanks for the detailed and constructive feedback and suggestions! You are right that the current framing might not highlight the core issue of aggregation clearly enough.
The paper is currently submitted, but we will consider PCI: Psych in the future, thanks for the hint.
The paper is currently submitted, but we will consider PCI: Psych in the future, thanks for the hint.
January 29, 2026 at 9:04 AM
Great, thanks for the detailed and constructive feedback and suggestions! You are right that the current framing might not highlight the core issue of aggregation clearly enough.
The paper is currently submitted, but we will consider PCI: Psych in the future, thanks for the hint.
The paper is currently submitted, but we will consider PCI: Psych in the future, thanks for the hint.
Thanks for this remark, we will mention in a revision that the field has already been changing. That's good to see!
January 29, 2026 at 9:00 AM
Thanks for this remark, we will mention in a revision that the field has already been changing. That's good to see!
Agreed!👍
However, it took me a while to understand that all main effects in the probit GLMM are simply effects on the response bias c, whereas all interactions with the factor indicating noise/signal of the stimuli can be interpreted as (main or interaction) effects on the discriminability d'.
However, it took me a while to understand that all main effects in the probit GLMM are simply effects on the response bias c, whereas all interactions with the factor indicating noise/signal of the stimuli can be interpreted as (main or interaction) effects on the discriminability d'.
January 29, 2026 at 8:59 AM
Agreed!👍
However, it took me a while to understand that all main effects in the probit GLMM are simply effects on the response bias c, whereas all interactions with the factor indicating noise/signal of the stimuli can be interpreted as (main or interaction) effects on the discriminability d'.
However, it took me a while to understand that all main effects in the probit GLMM are simply effects on the response bias c, whereas all interactions with the factor indicating noise/signal of the stimuli can be interpreted as (main or interaction) effects on the discriminability d'.
Link to preprint: osf.io/preprints/ps...
January 28, 2026 at 8:49 AM
Link to preprint: osf.io/preprints/ps...
New preprint by @semihaktepe.bsky.social 🎉
We compare ANOVA/SDT/GLMM for binary judgments in 20 datasets of the truth effect. #lme4
Main conclusion:
"GLMMs are a theoretically sound and practically robust method and thus superior for analyzing binary judgments in social and cognitive psychology.”
We compare ANOVA/SDT/GLMM for binary judgments in 20 datasets of the truth effect. #lme4
Main conclusion:
"GLMMs are a theoretically sound and practically robust method and thus superior for analyzing binary judgments in social and cognitive psychology.”
January 28, 2026 at 8:10 AM
New preprint by @semihaktepe.bsky.social 🎉
We compare ANOVA/SDT/GLMM for binary judgments in 20 datasets of the truth effect. #lme4
Main conclusion:
"GLMMs are a theoretically sound and practically robust method and thus superior for analyzing binary judgments in social and cognitive psychology.”
We compare ANOVA/SDT/GLMM for binary judgments in 20 datasets of the truth effect. #lme4
Main conclusion:
"GLMMs are a theoretically sound and practically robust method and thus superior for analyzing binary judgments in social and cognitive psychology.”
Interesting postdoc position on modeling psychological resilience at the DKFZ Hector Cancer Institute, Mannheim:
jobs.dkfz.de/en/jobs/1682...
--> dynamic/latent-variable models, Bayesian hierarchical models, causal inference, time-series analysis, cognitive modelling
jobs.dkfz.de/en/jobs/1682...
--> dynamic/latent-variable models, Bayesian hierarchical models, causal inference, time-series analysis, cognitive modelling
Postdoctoral Researcher in Quantitative Psychology / Data Science in Psychology
jobs.dkfz.de
December 17, 2025 at 3:47 PM
Interesting postdoc position on modeling psychological resilience at the DKFZ Hector Cancer Institute, Mannheim:
jobs.dkfz.de/en/jobs/1682...
--> dynamic/latent-variable models, Bayesian hierarchical models, causal inference, time-series analysis, cognitive modelling
jobs.dkfz.de/en/jobs/1682...
--> dynamic/latent-variable models, Bayesian hierarchical models, causal inference, time-series analysis, cognitive modelling
Reposted by Daniel Heck
The latest version of Sigmund, the #AI #research assistant, is able to fully control the #OpenSesame interface. This allows you to build #psychology experiments entirely through conversation. This is no AI slop, but real functionality with real benefits! Tutorial 👉 osdoc.cogsci.nl/4.1/tutorial...
November 13, 2025 at 9:40 AM
The latest version of Sigmund, the #AI #research assistant, is able to fully control the #OpenSesame interface. This allows you to build #psychology experiments entirely through conversation. This is no AI slop, but real functionality with real benefits! Tutorial 👉 osdoc.cogsci.nl/4.1/tutorial...
Reposted by Daniel Heck
November 12, 2025 at 12:32 AM
The method worked better than simple aggregation for validation words such as "fifty-fifty chance", "never" or "always" (indicated in the plot by the black intervals compared to the gray interval areas).
November 5, 2025 at 3:26 PM
The method worked better than simple aggregation for validation words such as "fifty-fifty chance", "never" or "always" (indicated in the plot by the black intervals compared to the gray interval areas).
New paper by @matzekloft.bsky.social in Psychometrika🎉
We developed a model for aggregating response intervals which are obtained if a participant judges that the word "most" covers the range [86% - 97%].
The model aggregates such intervals while considering people's proficiency & item difficulty.
We developed a model for aggregating response intervals which are obtained if a participant judges that the word "most" covers the range [86% - 97%].
The model aggregates such intervals while considering people's proficiency & item difficulty.
The final paper of my dissertation has just been published in Psychometrika (@pmetricsoc.bsky.social). Many thanks to @bsiepe.bsky.social and @danielheck.bsky.social, who co-authored it with me.
www.cambridge.org/core/journal...
www.cambridge.org/core/journal...
The Interval Consensus Model: Aggregating Continuous Bounded Interval Responses | Psychometrika | Cambridge Core
The Interval Consensus Model: Aggregating Continuous Bounded Interval Responses
www.cambridge.org
November 5, 2025 at 3:20 PM
New paper by @matzekloft.bsky.social in Psychometrika🎉
We developed a model for aggregating response intervals which are obtained if a participant judges that the word "most" covers the range [86% - 97%].
The model aggregates such intervals while considering people's proficiency & item difficulty.
We developed a model for aggregating response intervals which are obtained if a participant judges that the word "most" covers the range [86% - 97%].
The model aggregates such intervals while considering people's proficiency & item difficulty.
Halloween in our lab 🎃
Beware of divergent transitions!
Beware of divergent transitions!
Actually, we recently spooked away some ghosts with warnings and other scary things (with @semihaktepe.bsky.social)
October 31, 2025 at 2:26 PM
Halloween in our lab 🎃
Beware of divergent transitions!
Beware of divergent transitions!
Reposted by Daniel Heck
There still seems to be a lot of confusion about significance testing in psych. No, p-values *don’t* become useless at large N. This flawed point also used to be framed as "too much power". But power isn't the problem – it's 1) unbalanced error rates and 2) the (lack of a) SESOI. 1/ >
October 31, 2025 at 8:13 AM
There still seems to be a lot of confusion about significance testing in psych. No, p-values *don’t* become useless at large N. This flawed point also used to be framed as "too much power". But power isn't the problem – it's 1) unbalanced error rates and 2) the (lack of a) SESOI. 1/ >
Reposted by Daniel Heck
Simulation studies have a conflict of interest problem. The same team:
- develops a new method
- designs a simulation study to evaluate it
However, the new method has to show good performance to get published.
We propose living synthetic benchmarks to address the issue (doi.org/10.48550/arX...).
- develops a new method
- designs a simulation study to evaluate it
However, the new method has to show good performance to get published.
We propose living synthetic benchmarks to address the issue (doi.org/10.48550/arX...).
October 23, 2025 at 4:02 PM
Simulation studies have a conflict of interest problem. The same team:
- develops a new method
- designs a simulation study to evaluate it
However, the new method has to show good performance to get published.
We propose living synthetic benchmarks to address the issue (doi.org/10.48550/arX...).
- develops a new method
- designs a simulation study to evaluate it
However, the new method has to show good performance to get published.
We propose living synthetic benchmarks to address the issue (doi.org/10.48550/arX...).
Welcome to @lauragroot.bsky.social as a new member in the Psychological Methods Lab at @unimarburg.bsky.social! 🎉
Laura just started as a PhD student and will work on Bayesian statistics and multinomial processing tree (MPT) modeling.
Laura just started as a PhD student and will work on Bayesian statistics and multinomial processing tree (MPT) modeling.
October 30, 2025 at 9:13 PM
Welcome to @lauragroot.bsky.social as a new member in the Psychological Methods Lab at @unimarburg.bsky.social! 🎉
Laura just started as a PhD student and will work on Bayesian statistics and multinomial processing tree (MPT) modeling.
Laura just started as a PhD student and will work on Bayesian statistics and multinomial processing tree (MPT) modeling.
Reposted by Daniel Heck
We built the openESM database:
▶️60 openly available experience sampling datasets (16K+ participants, 740K+ obs.) in one place
▶️Harmonized (meta-)data, fully open-source software
▶️Filter & search all data, simply download via R/Python
Find out more:
🌐 openesmdata.org
📝 doi.org/10.31234/osf...
▶️60 openly available experience sampling datasets (16K+ participants, 740K+ obs.) in one place
▶️Harmonized (meta-)data, fully open-source software
▶️Filter & search all data, simply download via R/Python
Find out more:
🌐 openesmdata.org
📝 doi.org/10.31234/osf...
October 22, 2025 at 7:34 PM
We built the openESM database:
▶️60 openly available experience sampling datasets (16K+ participants, 740K+ obs.) in one place
▶️Harmonized (meta-)data, fully open-source software
▶️Filter & search all data, simply download via R/Python
Find out more:
🌐 openesmdata.org
📝 doi.org/10.31234/osf...
▶️60 openly available experience sampling datasets (16K+ participants, 740K+ obs.) in one place
▶️Harmonized (meta-)data, fully open-source software
▶️Filter & search all data, simply download via R/Python
Find out more:
🌐 openesmdata.org
📝 doi.org/10.31234/osf...
Welcome to @timangelike.bsky.social as a new member of our lab at @unimarburg.bsky.social! 🎉
Tim just started as a postdoc in a Momentum project funded by @volkswagenstiftung.de:
"Bridging the Gap Between Verbal Psychological Theories and Formal Statistical Modeling with Large Language Models"
Tim just started as a postdoc in a Momentum project funded by @volkswagenstiftung.de:
"Bridging the Gap Between Verbal Psychological Theories and Formal Statistical Modeling with Large Language Models"
October 16, 2025 at 1:18 PM
Welcome to @timangelike.bsky.social as a new member of our lab at @unimarburg.bsky.social! 🎉
Tim just started as a postdoc in a Momentum project funded by @volkswagenstiftung.de:
"Bridging the Gap Between Verbal Psychological Theories and Formal Statistical Modeling with Large Language Models"
Tim just started as a postdoc in a Momentum project funded by @volkswagenstiftung.de:
"Bridging the Gap Between Verbal Psychological Theories and Formal Statistical Modeling with Large Language Models"
Reposted by Daniel Heck
Konservative 🤝Rechtsextreme = politischer Selbstmord?
In vielen EU-Ländern versuchen konservative Parteien durch Kooperationen, rechtsextreme Parteien zu „zähmen“. Die CDU-nahe Konrad-Adenauer-Stiftung hat untersucht, was das für Konsequenzen hat. Spoiler: Es funktioniert nicht. 1/6🧵
In vielen EU-Ländern versuchen konservative Parteien durch Kooperationen, rechtsextreme Parteien zu „zähmen“. Die CDU-nahe Konrad-Adenauer-Stiftung hat untersucht, was das für Konsequenzen hat. Spoiler: Es funktioniert nicht. 1/6🧵
Laut CDU-naher Studie: Rechte Parteien durch Kooperation zu zähmen, funktioniert nicht
Der Rechtsruck beschäftigt nicht nur Deutschland, in anderen EU-Ländern versuchen sich rechte Parteien in Regierungsverantwortung. Die Konrad-Adenauer-Stiftung hat untersucht, was das für Konservative...
www.spiegel.de
September 23, 2025 at 11:41 AM
Konservative 🤝Rechtsextreme = politischer Selbstmord?
In vielen EU-Ländern versuchen konservative Parteien durch Kooperationen, rechtsextreme Parteien zu „zähmen“. Die CDU-nahe Konrad-Adenauer-Stiftung hat untersucht, was das für Konsequenzen hat. Spoiler: Es funktioniert nicht. 1/6🧵
In vielen EU-Ländern versuchen konservative Parteien durch Kooperationen, rechtsextreme Parteien zu „zähmen“. Die CDU-nahe Konrad-Adenauer-Stiftung hat untersucht, was das für Konsequenzen hat. Spoiler: Es funktioniert nicht. 1/6🧵
That's interesting - thanks for comparing the methods and for letting me know!
Analytical solutions are of course more elegant than numerical integration.
Analytical solutions are of course more elegant than numerical integration.
September 23, 2025 at 8:04 PM
That's interesting - thanks for comparing the methods and for letting me know!
Analytical solutions are of course more elegant than numerical integration.
Analytical solutions are of course more elegant than numerical integration.
Reposted by Daniel Heck
For all who use Bayesian hierarchical models, have a look at our new preprint, out now together with @linushof.bsky.social @nunobusch.bsky.social and @thorstenpachur.bsky.social
osf.io/preprints/ps...
osf.io/preprints/ps...
osf.io
September 10, 2025 at 2:41 PM
For all who use Bayesian hierarchical models, have a look at our new preprint, out now together with @linushof.bsky.social @nunobusch.bsky.social and @thorstenpachur.bsky.social
osf.io/preprints/ps...
osf.io/preprints/ps...