TY - JOUR
A1 - Schad, Daniel
A1 - Nicenboim, Bruno
A1 - Bürkner, Paul-Christian
A1 - Betancourt, Michael
A1 - Vasishth, Shravan
T1 - Workflow techniques for the robust use of bayes factors
JF - Psychological methods
N2 - Inferences about hypotheses are ubiquitous in the cognitive sciences. Bayes factors provide one general way to compare different hypotheses by their compatibility with the observed data. Those quantifications can then also be used to choose between hypotheses. While Bayes factors provide an immediate approach to hypothesis testing, they are highly sensitive to details of the data/model assumptions and it's unclear whether the details of the computational implementation (such as bridge sampling) are unbiased for complex analyses. Hem, we study how Bayes factors misbehave under different conditions. This includes a study of errors in the estimation of Bayes factors; the first-ever use of simulation-based calibration to test the accuracy and bias of Bayes factor estimates using bridge sampling; a study of the stability of Bayes factors against different MCMC draws and sampling variation in the data; and a look at the variability of decisions based on Bayes factors using a utility function. We outline a Bayes factor workflow that researchers can use to study whether Bayes factors are robust for their individual analysis. Reproducible code is available from haps://osf.io/y354c/.
Translational Abstract
In psychology and related areas, scientific hypotheses are commonly tested by asking questions like "is [some] effect present or absent." Such hypothesis testing is most often carried out using frequentist null hypothesis significance testing (NIIST). The NHST procedure is very simple: It usually returns a p-value, which is then used to make binary decisions like "the effect is present/abscnt." For example, it is common to see studies in the media that draw simplistic conclusions like "coffee causes cancer," or "coffee reduces the chances of geuing cancer." However, a powerful and more nuanced alternative approach exists: Bayes factors. Bayes factors have many advantages over NHST. However, for the complex statistical models that arc commonly used for data analysis today, computing Bayes factors is not at all a simple matter. In this article, we discuss the main complexities associated with computing Bayes factors. This is the first article to provide a detailed workflow for understanding and computing Bayes factors in complex statistical models. The article provides a statistically more nuanced way to think about hypothesis testing than the overly simplistic tendency to declare effects as being "present" or "absent".
KW - Bayes factors
KW - Bayesian model comparison
KW - prior
KW - posterior
KW - simulation-based calibration
Y1 - 2022
U6 - https://doi.org/10.1037/met0000472
SN - 1082-989X
SN - 1939-1463
VL - 28
IS - 6
SP - 1404
EP - 1426
PB - American Psychological Association
CY - Washington
ER -
TY - JOUR
A1 - Stone, Kate
A1 - Lago, Sol
A1 - Schad, Daniel
T1 - Divergence point analyses of visual world data
BT - applications to bilingual research
JF - Bilingualism : language and cognition
N2 - Much work has shown that differences in the timecourse of language processing are central to comparing native (L1) and non-native (L2) speakers. However, estimating the onset of experimental effects in timecourse data presents several statistical problems including multiple comparisons and autocorrelation. We compare several approaches to tackling these problems and illustrate them using an L1-L2 visual world eye-tracking dataset. We then present a bootstrapping procedure that allows not only estimation of an effect onset, but also of a temporal confidence interval around this divergence point. We describe how divergence points can be used to demonstrate timecourse differences between speaker groups or between experimental manipulations, two important issues in evaluating L2 processing accounts. We discuss possible extensions of the bootstrapping procedure, including determining divergence points for individual speakers and correlating them with individual factors like L2 exposure and proficiency. Data and an analysis tutorial are available at https://osf.io/exbmk/.
KW - divergence point analyses
KW - non-parametric approaches
KW - bootstrapping
KW - visual world eye-tracking
KW - bilingualism
Y1 - 2021
U6 - https://doi.org/10.1017/S1366728920000607
SN - 1366-7289
SN - 1469-1841
VL - 24
IS - 5
SP - 833
EP - 841
PB - Cambridge Univ. Press
CY - Cambridge
ER -
TY - JOUR
A1 - Rabovsky, Milena
A1 - Schad, Daniel
A1 - Abdel Rahman, Rasha
T1 - Semantic richness and density effects on language production
BT - Electrophysiological and behavioral evidence
JF - Journal of experimental psychology : Learning, memory, and cognition
N2 - Language production ultimately aims to convey meaning. Yet words differ widely in the richness and density of their semantic representations, and these differences impact conceptual and lexical processes during speech planning. Here, we replicated the recent finding that semantic richness, measured as the number of associated semantic features according to semantic feature production norms, facilitates object naming. In contrast, intercorrelational semantic feature density, measured as the degree of intercorrelation of a concept's features, presumably resulting in the coactivation of closely related concepts, has an inhibitory influence. We replicated the behavioral effects and investigated their relative time course and electrophysiological correlates. Both the facilitatory effect of high semantic richness and the inhibitory influence of high feature density were reflected in an increased posterior positivity starting at about 250 ms, in line with previous reports of posterior positivities in paradigms employing contextual manipulations to induce semantic interference during language production. Furthermore, amplitudes at the same posterior electrode sites were positively correlated with object naming times between about 230 and 380 ms. The observed effects follow naturally from the assumption of conceptual facilitation and simultaneous lexical competition and are difficult to explain by language production theories dismissing lexical competition.
KW - picture naming
KW - ERPs
KW - semantic richness
KW - semantic features
KW - lexical
KW - competition
Y1 - 2021
U6 - https://doi.org/10.1037/xlm0000940
SN - 0278-7393
SN - 1939-1285
VL - 47
IS - 3
SP - 508
EP - 517
PB - American Psychological Association
CY - Washington
ER -
TY - JOUR
A1 - Schad, Daniel
A1 - Betancourt, Michael
A1 - Vasishth, Shravan
T1 - Toward a principled Bayesian workflow in cognitive science
JF - Psychological methods
N2 - Experiments in research on memory, language, and in other areas of cognitive science are increasingly being analyzed using Bayesian methods. This has been facilitated by the development of probabilistic programming languages such as Stan, and easily accessible front-end packages such as brms. The utility of Bayesian methods, however, ultimately depends on the relevance of the Bayesian model, in particular whether or not it accurately captures the structure of the data and the data analyst's domain expertise. Even with powerful software, the analyst is responsible for verifying the utility of their model. To demonstrate this point, we introduce a principled Bayesian workflow (Betancourt, 2018) to cognitive science. Using a concrete working example, we describe basic questions one should ask about the model: prior predictive checks, computational faithfulness, model sensitivity, and posterior predictive checks. The running example for demonstrating the workflow is data on reading times with a linguistic manipulation of object versus subject relative clause sentences. This principled Bayesian workflow also demonstrates how to use domain knowledge to inform prior distributions. It provides guidelines and checks for valid data analysis, avoiding overfitting complex models to noise, and capturing relevant data structure in a probabilistic model. Given the increasing use of Bayesian methods, we aim to discuss how these methods can be properly employed to obtain robust answers to scientific questions.
KW - workflow
KW - prior predictive checks
KW - posterior predictive checks
KW - model
KW - building
KW - Bayesian data analysis
Y1 - 2021
U6 - https://doi.org/10.1037/met0000275
SN - 1082-989X
SN - 1939-1463
VL - 26
IS - 1
SP - 103
EP - 126
PB - American Psychological Association
CY - Washington
ER -
TY - JOUR
A1 - Schad, Daniel
A1 - Vasishth, Shravan
T1 - The posterior probability of a null hypothesis given a statistically significant result
JF - The quantitative methods for psychology
N2 - When researchers carry out a null hypothesis significance test, it is tempting to assume that a statistically significant result lowers Prob(H0), the probability of the null hypothesis being true. Technically, such a statement is meaningless for various reasons: e.g., the null hypothesis does not have a probability associated with it. However, it is possible to relax certain assumptions to compute the posterior probability Prob(H0) under repeated sampling. We show in a step-by-step guide that the intuitively appealing belief, that Prob(H0) is low when significant results have been obtained under repeated sampling, is in general incorrect and depends greatly on: (a) the prior probability of the null being true; (b) type-I error rate, (c) type-II error rate, and (d) replication of a result. Through step-by-step simulations using open-source code in the R System of Statistical Computing, we show that uncertainty about the null hypothesis being true often remains high despite a significant result. To help the reader develop intuitions about this common misconception, we provide a Shiny app (https://danielschad.shinyapps.io/probnull/). We expect that this tutorial will help researchers better understand and judge results from null hypothesis significance tests.
KW - Null hypothesis significance testing
KW - Bayesian inference
KW - statistical
KW - power
Y1 - 2022
U6 - https://doi.org/10.20982/tqmp.18.2.p011
SN - 1913-4126
SN - 2292-1354
VL - 18
IS - 2
SP - 130
EP - 141
PB - University of Montreal, Department of Psychology
CY - Montreal
ER -
TY - JOUR
A1 - Chen, Hao
A1 - Nebe, Stephan
A1 - Mojtahedzadeh, Negin
A1 - Kuitunen-Paul, Soren
A1 - Garbusow, Maria
A1 - Schad, Daniel
A1 - Rapp, Michael A.
A1 - Huys, Quentin J. M.
A1 - Heinz, Andreas
A1 - Smolka, Michael N.
T1 - Susceptibility to interference between Pavlovian and instrumental control is associated with early hazardous alcohol use
JF - Addiction biology
N2 - Pavlovian-to-instrumental transfer (PIT) tasks examine the influence of Pavlovian stimuli on ongoing instrumental behaviour. Previous studies reported associations between a strong PIT effect, high-risk drinking and alcohol use disorder. This study investigated whether susceptibility to interference between Pavlovian and instrumental control is linked to risky alcohol use in a community sample of 18-year-old male adults. Participants (N = 191) were instructed to 'collect good shells' and 'leave bad shells' during the presentation of appetitive (monetary reward), aversive (monetary loss) or neutral Pavlovian stimuli. We compared instrumental error rates (ER) and functional magnetic resonance imaging (fMRI) brain responses between the congruent and incongruent conditions, as well as among high-risk and low-risk drinking groups. On average, individuals showed a substantial PIT effect, that is, increased ER when Pavlovian cues and instrumental stimuli were in conflict compared with congruent trials. Neural PIT correlates were found in the ventral striatum and the dorsomedial and lateral prefrontal cortices (lPFC). Importantly, high-risk drinking was associated with a stronger behavioural PIT effect, a decreased lPFC response and an increased neural response in the ventral striatum on the trend level. Moreover, high-risk drinkers showed weaker connectivity from the ventral striatum to the lPFC during incongruent trials. Our study links interference during PIT to drinking behaviour in healthy, young adults. High-risk drinkers showed higher susceptibility to Pavlovian cues, especially when they conflicted with instrumental behaviour, indicating lower interference control abilities. Increased activity in the ventral striatum (bottom-up), decreased lPFC response (top-down), and their altered interplay may contribute to poor interference control in the high-risk drinkers.
KW - high‐risk drinking
KW - interference control
KW - Pavlovian‐to‐instrumental transfer
Y1 - 2020
U6 - https://doi.org/10.1111/adb.12983
SN - 1355-6215
SN - 1369-1600
VL - 26
IS - 4
SP - 1
EP - 14
PB - Wiley
CY - Hoboken
ER -
TY - JOUR
A1 - Schad, Daniel
A1 - Vasishth, Shravan
A1 - Hohenstein, Sven
A1 - Kliegl, Reinhold
T1 - How to capitalize on a priori contrasts in linear (mixed) models
BT - a tutorial
JF - Journal of memory and language
N2 - Factorial experiments in research on memory, language, and in other areas are often analyzed using analysis of variance (ANOVA). However, for effects with more than one numerator degrees of freedom, e.g., for experimental factors with more than two levels, the ANOVA omnibus F-test is not informative about the source of a main effect or interaction. Because researchers typically have specific hypotheses about which condition means differ from each other, a priori contrasts (i.e., comparisons planned before the sample means are known) between specific conditions or combinations of conditions are the appropriate way to represent such hypotheses in the statistical model. Many researchers have pointed out that contrasts should be "tested instead of, rather than as a supplement to, the ordinary 'omnibus' F test" (Hays, 1973, p. 601). In this tutorial, we explain the mathematics underlying different kinds of contrasts (i.e., treatment, sum, repeated, polynomial, custom, nested, interaction contrasts), discuss their properties, and demonstrate how they are applied in the R System for Statistical Computing (R Core Team, 2018). In this context, we explain the generalized inverse which is needed to compute the coefficients for contrasts that test hypotheses that are not covered by the default set of contrasts. A detailed understanding of contrast coding is crucial for successful and correct specification in linear models (including linear mixed models). Contrasts defined a priori yield far more useful confirmatory tests of experimental hypotheses than standard omnibus F-tests. Reproducible code is available from https://osf.io/7ukf6/.
KW - contrasts
KW - null hypothesis significance testing
KW - linear models
KW - a priori
KW - hypotheses
Y1 - 2019
U6 - https://doi.org/10.1016/j.jml.2019.104038
SN - 0749-596X
SN - 1096-0821
VL - 110
PB - Elsevier
CY - San Diego
ER -
TY - JOUR
A1 - Nebe, Stephan
A1 - Kroemer, Nils B.
A1 - Schad, Daniel
A1 - Bernhardt, Nadine
A1 - Sebold, Miriam
A1 - Mueller, Dirk K.
A1 - Scholl, Lucie
A1 - Kuitunen-Paul, Sören
A1 - Heinz, Andreas
A1 - Rapp, Michael A.
A1 - Huys, Quentin J. M.
A1 - Smolka, Michael N.
T1 - No association of goal-directed and habitual control with alcohol consumption in young adults
JF - Addiction biology
N2 - Alcohol dependence is a mental disorder that has been associated with an imbalance in behavioral control favoring model-free habitual over model-based goal-directed strategies. It is as yet unknown, however, whether such an imbalance reflects a predisposing vulnerability or results as a consequence of repeated and/or excessive alcohol exposure. We, therefore, examined the association of alcohol consumption with model-based goal-directed and model-free habitual control in 188 18-year-old social drinkers in a two-step sequential decision-making task while undergoing functional magnetic resonance imaging before prolonged alcohol misuse could have led to severe neurobiological adaptations. Behaviorally, participants showed a mixture of model-free and model-based decision-making as observed previously. Measures of impulsivity were positively related to alcohol consumption. In contrast, neither model-free nor model-based decision weights nor the trade-off between them were associated with alcohol consumption. There were also no significant associations between alcohol consumption and neural correlates of model-free or model-based decision quantities in either ventral striatum or ventromedial prefrontal cortex. Exploratory whole-brain functional magnetic resonance imaging analyses with a lenient threshold revealed early onset of drinking to be associated with an enhanced representation of model-free reward prediction errors in the posterior putamen. These results suggest that an imbalance between model-based goal-directed and model-free habitual control might rather not be a trait marker of alcohol intake per se.
KW - alcohol
KW - goal-directed
KW - reinforcement learning
Y1 - 2017
U6 - https://doi.org/10.1111/adb.12490
SN - 1355-6215
SN - 1369-1600
VL - 23
IS - 1
SP - 379
EP - 393
PB - Wiley
CY - Hoboken
ER -
TY - JOUR
A1 - Obst, Elisabeth
A1 - Schad, Daniel
A1 - Huys, Quentin J. M.
A1 - Sebold, Miriam
A1 - Nebe, Stephan
A1 - Sommer, Christian
A1 - Smolka, Michael N.
A1 - Zimmermann, Ulrich S.
T1 - Drunk decisions
BT - Alcohol shifts choice from habitual towards goal-directed control in adolescent intermediate-risk drinkers
JF - Journal of Psychopharmacology
N2 - Background: Studies in humans and animals suggest a shift from goal-directed to habitual decision-making in addiction. We therefore tested whether acute alcohol administration reduces goal-directed and promotes habitual decision-making, and whether these effects are moderated by self-reported drinking problems. Methods: Fifty-three socially drinking males completed the two-step task in a randomised crossover design while receiving an intravenous infusion of ethanol (blood alcohol level=80 mg%), or placebo. To minimise potential bias by long-standing heavy drinking and subsequent neuropsychological impairment, we tested 18- to 19-year-old adolescents. Results: Alcohol administration consistently reduced habitual, model-free decisions, while its effects on goal-directed, model-based behaviour varied as a function of drinking problems measured with the Alcohol Use Disorders Identification Test. While adolescents with low risk for drinking problems (scoring <8) exhibited an alcohol-induced numerical reduction in goal-directed choices, intermediate-risk drinkers showed a shift away from habitual towards goal-directed decision-making, such that alcohol possibly even improved their performance. Conclusions: We assume that alcohol disrupted basic cognitive functions underlying habitual and goal-directed decisions in low-risk drinkers, thereby enhancing hasty choices. Further, we speculate that intermediate-risk drinkers benefited from alcohol as a negative reinforcer that reduced unpleasant emotional states, possibly displaying a novel risk factor for drinking in adolescence.
KW - Computer-assisted Alcohol Infusion System
KW - habitual learning
KW - model-free and model-based decision-making
KW - two-stage Markov decision task
KW - subjective response to ethanol
KW - drinking problems
KW - real-life drinking behaviour
Y1 - 2018
U6 - https://doi.org/10.1177/0269881118772454
SN - 0269-8811
SN - 1461-7285
VL - 32
IS - 8
SP - 855
EP - 866
PB - Sage Publ.
CY - London
ER -
TY - JOUR
A1 - Schad, Daniel
A1 - Garbusow, Maria
A1 - Friedel, Eva
A1 - Sommer, Christian
A1 - Sebold, Miriam
A1 - Hägele, Claudia
A1 - Bernhardt, Nadine
A1 - Nebe, Stephan
A1 - Kuitunen-Paul, Sören
A1 - Liu, Shuyan
A1 - Eichmann, Uta
A1 - Beck, Anne
A1 - Wittchen, Hans-Ulrich
A1 - Walter, Henrik
A1 - Sterzer, Philipp
A1 - Zimmermann, Ulrich S.
A1 - Smolka, Michael N.
A1 - Schlagenhauf, Florian
A1 - Huys, Quentin J. M.
A1 - Heinz, Andreas
A1 - Rapp, Michael A.
T1 - Neural correlates of instrumental responding in the context of alcohol-related cues index disorder severity and relapse risk
JF - European archives of psychiatry and clinical neuroscience : official organ of the German Society for Biological Psychiatry
N2 - The influence of Pavlovian conditioned stimuli on ongoing behavior may contribute to explaining how alcohol cues stimulate drug seeking and intake. Using a Pavlovian-instrumental transfer task, we investigated the effects of alcohol-related cues on approach behavior (i.e., instrumental response behavior) and its neural correlates, and related both to the relapse after detoxification in alcohol-dependent patients. Thirty-one recently detoxified alcohol-dependent patients and 24 healthy controls underwent instrumental training, where approach or non-approach towards initially neutral stimuli was reinforced by monetary incentives. Approach behavior was tested during extinction with either alcohol-related or neutral stimuli (as Pavlovian cues) presented in the background during functional magnetic resonance imaging (fMRI). Patients were subsequently followed up for 6 months. We observed that alcohol-related background stimuli inhibited the approach behavior in detoxified alcohol-dependent patients (t = -3.86, p < .001), but not in healthy controls (t = -0.92, p = .36). This behavioral inhibition was associated with neural activation in the nucleus accumbens (NAcc) (t((30)) = 2.06, p < .05). Interestingly, both the effects were only present in subsequent abstainers, but not relapsers and in those with mild but not severe dependence. Our data show that alcohol-related cues can acquire inhibitory behavioral features typical of aversive stimuli despite being accompanied by a stronger NAcc activation, suggesting salience attribution. The fact that these findings are restricted to abstinence and milder illness suggests that they may be potential resilience factors.
KW - Alcohol dependence
KW - Human neuroimaging
KW - Nucleus accumbens
KW - Pavlovian-instrumental transfer
KW - Relapse
Y1 - 2018
U6 - https://doi.org/10.1007/s00406-017-0860-4
SN - 0940-1334
SN - 1433-8491
VL - 269
IS - 3
SP - 295
EP - 308
PB - Springer
CY - Heidelberg
ER -