Search

Missing data are ubiquitous in psychological research. They may come about as an unwanted result of coding or computer error, participants' non-response or absence, or missing values may be intentional, as in planned missing designs. We discuss the effects of missing data on χ²-based goodness-of-fit indices in Structural Equation Modeling (SEM), specifically on the Root Mean Squared Error of Approximation (RMSEA). We use simulations to show that naive implementations of the RMSEA have a downward bias in the presence of missing data and, thus, overestimate model goodness-of-fit. Unfortunately, many state-of-the-art software packages report the biased form of RMSEA. As a consequence, the scientific community may have been accepting a much larger fraction of models with non-acceptable model fit. We propose a bias-correction for the RMSEA based on information-theoretic considerations that take into account the expected misfit of a person with fully observed data. The corrected RMSEA is asymptotically independent of the proportion of missing data for misspecified models. Importantly, results of the corrected RMSEA computation are identical to naive RMSEA if there are no missing data.

Here, we take a computational approach to understand the mechanisms underlying face perception biases in depression. Thirty participants diagnosed with major depressive disorder and 30 healthy control participants took part in three studies involving recognition of identity and emotion in faces. We used signal detection theory to determine whether any perceptual biases exist in depression aside from decisional biases. We found lower sensitivity to happiness in general, and lower sensitivity to both happiness and sadness with ambiguous stimuli. Our use of highly-controlled face stimuli ensures that such asymmetry is truly perceptual in nature, rather than the result of studying expressions with inherently different discriminability. We found no systematic effect of depression on the perceptual interactions between face expression and identity. We also found that decisional strategies used in our task were different for people with depression and controls, but in a way that was highly specific to the stimulus set presented. We show through simulation that the observed perceptual effects, as well as other biases found in the literature, can be explained by a computational model in which channels encoding positive expressions are selectively suppressed. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Ordinal variables, while extremely common in Psychology, are almost exclusively analysed with statistical models that falsely assume them to be metric. This practice can lead to distorted effect size estimates, inflated error rates, and other problems. We argue for the application of ordinal models that make appropriate assumptions about the variables under study. In this tutorial article, we first explain the three major ordinal model classes; the cumulative, sequential and adjacent category models. We then show how to fit ordinal models in a fully Bayesian framework with the R package brms, using data sets on stem cell opinions and marriage time courses. Two appendices provide detailed mathematical derivations of the models, and a third practical example that connects ordinal models to Signal Detection Theory with confidence rating data. Ordinal models provide better theoretical interpretation and numerical inference from ordinal data, and we recommend their widespread adoption in Psychology.

Background. Studies investigating the link between depressive symptoms and inflammation have yielded inconsistent results, which may be due to two factors. First, studies differed regarding the specific inflammatory markers studied and covariates accounted for. Second, specific depressive symptoms may be differentially related to inflammation. We address both challenges using network psychometrics.Methods. We estimated seven regularized mixed graphical models in the NESDA data (N=2321) to explore shared variances among 1) depression severity, modeled via depression sum-score, 9 DSM-5 symptoms, or 28 individual depressive symptoms; 2) inflammatory markers C-reactive protein (CRP), interleukin 6 (IL-6), and tumor necrosis factor α (TNF-α); 3) before and after adjusting for sex, age, body mass index (BMI), exercise, smoking, alcohol, and chronic diseases.Results. The depression sum-score was related to both IL-6 and CRP before, and only to IL-6 after covariate adjustment. When modeling the DSM-5 symptoms and CRP in a conceptual replication of Jokela et al. (2016), CRP was associated with ‘sleep problems’, ‘energy level’, and ‘weight/appetite changes’; only the first two links survived covariate adjustment. In a conservative model with all 38 variables, symptoms and markers were unrelated. Following recent psychometric work, we re-estimated the full model without regularization: the depressive symptoms ‘insomnia, ‘hypersomnia’, and ‘aches and pain’ showed unique positive relations to all inflammatory markers.Conclusions. We found evidence for differential relations between markers, depressive symptoms, and covariates. Associations between symptoms and markers were attenuated after covariate adjustment; BMI and sex consistently showed strong relations with inflammatory markers.

The prevalence of obesity has driven searches for cognitive or behavioural economic factors related to Body Mass Index (BMI). One compelling candidate is delay discounting: those who prefer smaller sooner rewards over larger but later rewards are hypothesised to have higher BMI. The findings in the literature are mixed however, with meta analyses suggesting only a very small correlation between discounting and BMI. Here we present novel empirical data (N=381) and Bayesian analyses which suggest no such relationship between discounting of either monetary or weight loss rewards and BMI. We also find evidence against our novel proposal that discounting moderates the rate of BMI gain over time. We also present our data in the context of a random effects Bayesian meta-analytical result which does suggest the presence of a small correlation overall. The strength of the correlation is so weak (2.25% shared variance) that its practical significance may be minor to non existent. However because we found decisive evidence for unaccounted for study-level variance, due to study heterogeneity, we argue that we should treat such meta-analytic correlations with extreme caution. While the relationship between discounting and health outcomes such as BMI remain theoretically appealing, our empirical and meta-analytic results suggest we should be cautious in inferring a correlational, let alone a causal, role for discounting processes in driving BMI or moderating BMI gain with age.

This article introduces the NetworkToolbox package for R. Network analysis offers an intuitive perspective on complex phenomena via models depicted by nodes (variables) and edges (correlations). The ability of networks to model complexity has made them the standard approach for modeling the intricate interactions in the brain. Similarly, networks have become an increasingly attractive model for studying the complexity of psychological and psychopathological phenomena. NetworkToolbox aims to provide psychological researchers with state-of-the-art methods and measures for estimating and analyzing brain, cognitive, and psychometric data. In this article, I introduce NetworkToolbox and provide a manual for applying the package’s functions to psychometric (personality) data.

Psychological data often violate the normality assumptions made by commonly used statistical methods. These violations are addressed in a variety of ways such as transformations or assuming the employed method is robust to violations. Here we argue that data transformations are unnecessary at best and severely misleading at worst. An alternative approach is to use a Bayesian model that permits skewness and other perturbations to classical assumptions (e.g., heteroskedasticity). Through simulation, we demonstrate that a Bayesian skew-normal model has optimal frequentist properties (i.e., "type 1" error, "power", unbiasedness) compared to normal-assumptive models with or without transformation. Furthermore, the Bayesian skew-normal model has greater predictive utility, as indicated by posterior predictive checking and approximate leave-one-out cross-validation. After an applied example, we discuss practical implications of our findings for psychological science in general, and specifically how Bayesian modeling can improve reproducibility in psychology.

[Correction Notice: An Erratum for this article was reported in Vol 115(6) of Journal of Personality and Social Psychology (see record 2018-63189-002). In the article, the original supplemental material has been revised to include a clarifying note to the Tests of model fit over larger sample (Samples C-G) section and post-peer review analyses added to the Post-peer review Factor Analytic Tests section. All versions of this article have been corrected.] Philosophers and behavioral scientists refer to wisdom as unbiased reasoning that guides one toward a balance of interests and promotes a good life. However, major instruments developed to test wisdom appear biased, and it is unclear whether they capture balance-related tendencies. We examined whether shifting from global, de-contextualized reports to state-level reports about concrete situations provides a less biased method to assess wise reasoning (e.g., intellectual humility, recognition of uncertainty and change, consideration of the broader context at hand and perspectives of others, integration of these perspectives or compromise), which may be aligned with the notion of balancing interests. Results of a large-scale psychometric investigation (N = 4,463) revealed that the novel Situated WIse Reasoning Scale (SWIS) is reliable and appears independent of psychological biases (attribution bias, bias blind spot, self-deception, and impression management), whereas global wisdom reports are subject to such biases. Moreover, SWIS scores were positively related to indices of living well (e.g., adaptive emotion regulation, mindfulness), and balancing of cooperative and self-protective interests, goals (influence-vs.-adjustment), and causal inferences about conflict (attribution to the self-vs.-other party). In contrast, global wisdom reports were unrelated or negatively related to balance-related measures. Notably, people showed modest within-person consistency in wise reasoning across situations or over time, suggesting that a single-shot measurement may be insufficient for whole understanding of trait-level wisdom. We discuss theoretical and practical implications for research on wisdom, judgment and decision making, well-being, and prosociality. (PsycINFO Database Record (c) 2018 APA, all rights reserved).

Previous studies found that an automatic meaning-composition process affects the processing of morphologically complex words, and related this operation to conceptual combination. However, research on embodied cognition demonstrates that concepts are more than just lexical meanings, rather being also grounded in perceptual experience. Therefore, perception-based information should also be involved in mental operations on concepts, such as conceptual combination. Consequently, we should expect to find perceptual effects in the processing of morphologically complex words. In order to investigate this hypothesis, we present the first fully-implemented and data-driven model of perception-based (more specifically, vision-based) conceptual combination, and use the predictions of such a model to investigate processing times for compound words in four large-scale behavioral experiments employing three paradigms (naming, lexical decision, and timed sensibility judgments). We observe facilitatory effects of vision-based compositionality in all three paradigms, over and above a strong language-based (lexical and semantic) baseline, thus demonstrating for the first time perceptually grounded effects at the sub-lexical level. This suggests that perceptually-grounded information is not only utilized according to specific task demands but rather automatically activated when available.

Mixed-effects models are becoming common in psychological science. Although they have many desirable features, there is still untapped potential. It is customary to view homogeneous variance as an assumption to satisfy. We argue to move beyond that perspective, and to view modeling within-person variance as an opportunity to gain a richer understanding of psychological processes. The technique to do so is based on the mixed-effects location scale model that can simultaneously estimate mixed-effects submodels to both the mean (location) and within-person variance (scale). We develop a framework that goes beyond assessing the submodels in isolation of one another and introduce a novel Bayesian hypothesis test for mean—variance correlations in the distribution of random effects. We first present a motivating example, which makes clear how the model can characterize mean—variance relations. We then apply the method to reaction times (RTs) gathered from 2 cognitive inhibition tasks. We find there are more individual differences in the within-person variance than the mean structure, as well as a complex web of structural mean—variance relations. This stands in contrast to the dominant view of within-person variance (i.e., “noise”). The results also point toward paradoxical within-person, as opposed to between-person, effects: several people had slower and less variable incongruent responses. This contradicts the typical pattern, wherein larger means tend to be associated with more variability. We conclude with future directions, spanning from methodological to theoretical inquires, that can be answered with the presented methodology.Impact StatementTranslational Abstract: Mixed-effects models are becoming common in psychological science. Although researchers typically focus on explaining the dependent variable, characterizing individual differences in within-person variance can provide valuable insights into behavioral consistency. The technique to do so is based on an extension to the mixed-effect model that can simultaneously predict both the mean and within-person variance. We introduce Bayesian methodology to test the covariance among individual difference parameters that capture mean and within-person variance relations. This is motivated by theoretical models that predict larger means will be associated with more variability. Our approach allows for confirmatory hypothesis testing of this proposed relation in the distribution of random effects. Hence, researchers can begin to test the interplay between an experimental effect on the mean and within-person variability in their repeated measurements. We apply our method to RTs gathered from two cognitive inhibition tasks. We find there are more individual differences in the within-person variance than the mean structure, as well as a complex web of structural mean-variance relations. We also uncovered strikingly large correlations between the experimental effect on the mean and within-person variability. Substantively this novel finding suggests that cognitive inhibition has important and highly interrelated signatures on both RT and RT consistency. The method is implemented in the user-friendly R package hypMuVar.

Under labels such as unconscious processing and subliminal perception, identification of stimuli falling below the subjective threshold (whether truly unconscious or not) has been found remarkably accurate in some experiments while completely at chance in others. Here, we first identify that an apparent window of subliminal perception arises in humans under specific stimulus conditions using different experimental paradigms and analysis methods. We then show that the standard signal detection theory (SDT) model is unable to account for this window and extend it until it is. We finally compare a range of models on empirical data. The models performing best are notable for their absence of hierarchical levels, indicating that the window could be a base property of any phenomenally conscious system. The models explain previously incompatible findings in the literature, and they allow for estimations of peaks in subthreshold perception across the spectrum of stimulus saliency, which may be used in further studies of subliminal perception.

Decision making is a vital aspect of our everyday functioning, from simple perceptual demands to more complex and meaningful decisions. The strategy adopted to make such decisions is often viewed as balancing elements of speed and caution, i.e. making fast or careful decisions. Using sequential sampling models to analyse decision making data can allow us to tease apart strategic differences, such as being more or less cautious, from processing differences, which would otherwise be indistinguishable in behavioural data. Our study used a multiple object tracking task where student participants and a highly skilled military group were compared on their ability to track several items at once. Using a mathematical model of decision making (the linear ballistic accumulator), we show the underpinnings of how two groups differ in performance. Results showed a large difference between the groups on accuracy, with the Royal Australian Air Force (RAAF) group outperforming students. An interaction effect was observed between groups and level of difficulty in response times, where RAAF response times slowed at a greater rate than the student group as difficulty increased. Model results indicated that the RAAF personnel were more cautious in their decisions than students, and had faster processing in some conditions. Our study shows the strength of sequential sampling models, as well as providing a first attempt at fitting a sequential sampling model to data from a multiple object tracking task.

Uncertainty plays a critical role in reinforcement learning and decision making. However, exactly how it influences behavior remains unclear. Multiarmed-bandit tasks offer an ideal test bed, since computational tools such as approximate Kalman filters can closely characterize the interplay between trial-by-trial values, uncertainty, learning, and choice. To gain additional insight into learning and choice processes, we obtained data from subjects’ overt allocation of gaze. The estimated value and estimation uncertainty of options influenced what subjects looked at before choosing; these same quantities also influenced choice, as additionally did fixation itself. A momentary measure of uncertainty in the form of absolute prediction errors determined how long participants looked at the obtained outcomes. These findings affirm the importance of uncertainty in multiple facets of behavior and help delineate its effects on decision making.

To reach sustainability transitions, we must learn to leverage social systems into tipping points, where societies exhibit positive feedback loops in the adoption of sustainable behavioural or cultural traits. For instance, future sustainability transitions are often portrayed as having roughly ‘S-shaped’ timelines, introducing pro-environmental lifestyles at an increasingly fast rate until a new sustainable state is reached. However, much less is known about the most efficient ways to reach such transitions, or how self-reinforcing systemic transformations might be instigated through means of policy. This research article models behaviour as the function of the person and their environment. We employ an agent-based model to study the emergence of social tipping points through five interconnected factors which have been previously identified to constitute an ecological approach to human behaviour. These are 1. affordances (action opportunities provided by the physical environment), 2. asocial learning and habituation, 3. personal states (such as intentions and habits), 4. cultural niche construction, and 5. social learning in a social network. Our model suggests that even a linear introduction of pro-environmental affordances to a social system can have non-linear self-reinforcing effects on the emergence of collective pro-environmental behaviour patterns. We also conduct empirical validation with a case study, validating the model against data of the evolution of cycling and driving behaviours in central Copenhagen. Our model gives further evidence and justification for policies that make pro-environmental behaviour psychologically salient, easy, and the path of least resistance.

To date, the application of semantic network methodologies to study cognitive processes in psychological phenomena has been limited in scope. One barrier to broader application is the lack of resources for researchers unfamiliar with the approach. Another barrier, for both the unfamiliar and knowledgeable researcher, is the tedious and laborious preprocessing of semantic data. We aim to minimize these barriers by offering a comprehensive semantic network analysis pipeline (preprocessing, estimating, and analyzing networks), and an associated R tutorial that uses a suite of R packages to accommodate the pipeline. Two of these packages, SemNetDictionaries and SemNetCleaner, promote an efficient, reproducible, and transparent approach to preprocessing linguistic data. The third package, SemNeT, provides methods and measures for estimating and statistically comparing semantic networks via a point-and-click graphical user interface. Using real-world data, we present a start-to-finish pipeline from raw data to semantic network analysis results. This article aims to provide resources for researchers, both the unfamiliar and knowledgeable, that reduce some of the barriers for conducting semantic network analysis.

In recent years, open science practices have become increasingly popular in psychology and related sciences. These practices aim to increase rigour and transparency in science as a potential response to the challenges posed by the replication crisis. Many of these reforms—including the increasingly usedpreregistration—have been designed for purely experimental work that tests straightforward hypotheses with standard inferential statistical analyses, such as assessing whether an experimental manipulation has an effect on a variable of interest. But psychology is a diverse field of research. The somewhat narrow focus of the prevalent discussions surrounding and templates for preregistration has led to debates on how appropriate these reforms are for areas of research with more diverse hypotheses and more intricate methods of analysis, such as cognitive modelling research within mathematical psychology. Our article attempts to bridge the gap between open science and mathematical psychology, focusing on the type of cognitive modelling that Crüwellet al.(Crüwell S, Stefan AM, Evans NJ. 2019 Robust standards in cognitive science.Comput. Brain Behav.2, 255–265) labelledmodel application, where researchers apply a cognitive model as ameasurement toolto test hypotheses about parameters of the cognitive model. Specifically, we (i) discuss several potential researcher degrees of freedom within model application, (ii) provide the first preregistration template for model application and (iii) provide an example of a preregistered model application using our preregistration template. More broadly, we hope that our discussions and concrete proposals constructively advance the mostly abstract current debate surrounding preregistration in cognitive modelling, and provide a guide for how preregistration templates may be developed in other diverse or intricate research contexts.

This dissertation deals with the problem of modeling psychopathology. Its first part focuses on statistical (data) models and introduces a number of models for cross-sectional and time series data that can be visualized as a network. This includes Mixed Graphical Models (MGMs), which allow one to include variables of different types in a statistical network model, Moderated Network Models (MNMs) which allow pairwise interactions to be moderated by other variables in the model, and time-varying Vector Autoregressive (VAR) models and MGMs that relax the standard assumption of stationarity. In addition, I discuss several methodological issues related to statistical network models such as the importance of considering predictability, model selection between AR and VAR models, and how the interpretation of the Ising model depends on its domain. The second part focuses on formal theories of psychopathology and how to develop them using data models. I first illustrate the fundamental difficulties in obtaining a formal theory with a purely statistical approach, by trying to recover an assumed bistable system for emotion dynamics with currently popular time series analyses. Next, I present a formal theory of panic disorder, based on an extensive review of the literature on the phenomenology of panic disorder and existing theories. Finally, I discuss three different ways to use data models to construct formal theories about psychopathological phenomena. Based on this discussion, I put forward an abductive framework for constructing formal theories for psychological and psychopathological phenomena.

The positive manifold of intelligence has fascinated generations of scholars in human ability. In the past century, various formal explanations have been proposed, including the dominant g factor, the revived sampling theory, and the recent multiplier effect model and mutualism model. In this article, we propose a novel idiographic explanation. We formally conceptualize intelligence as evolving networks in which new facts and procedures are wired together during development. The static model, an extension of the Fortuin–Kasteleyn model, provides a parsimonious explanation of the positive manifold and intelligence’s hierarchical factor structure. We show how it can explain the Matthew effect across developmental stages. Finally, we introduce a method for studying growth dynamics. Our truly idiographic approach offers a new view on a century-old construct and ultimately allows the fields of human ability and human learning to coalesce.

Agents that can make better use of computation, experience, time, and memory can solve a greater range of problems more effectively. A crucial ingredient for managing such finite resources is intelligently chosen abstract representations. But, how do abstractions facilitate problem solving under limited resources? What makes an abstraction useful? To answer such questions, we review several trends in recent reinforcement-learning research that provide insight into how abstractions interact with learning and decision making. During learning, abstraction can guide exploration and generalization as well as facilitate efficient tradeoffs---e.g., time spent learning versus the quality of a solution. During computation, good abstractions provide simplified models for computation while also preserving relevant information about decision-theoretic quantities. These features of abstraction are not only key for scaling up artificial problem solving, but can also shed light on what pressures shape the use of abstract representations in humans and other organisms.

A wide variety of effect size indices have been proposed for quantifying the magnitude of treatment effects in single-case designs. Commonly used measures include parametric indices such as the standardized mean difference as well as nonoverlap measures such as the percentage of nonoverlapping data, improvement rate difference, and nonoverlap of all pairs. Currently, little is known about the properties of these indices when applied to behavioral data collected by systematic direct observation, even though systematic direct observation is the most common method for outcome measurement in single-case research. This study uses Monte Carlo simulation to investigate the properties of several widely used single-case effect size measures when applied to systematic direct observation data. Results indicate that the magnitude of the nonoverlap measures and of the standardized mean difference can be strongly influenced by procedural details of the study's design, which is a significant limitation to using these indices as effect sizes for meta-analysis of single-case designs. A less widely used parametric index, the log response ratio, has the advantage of being insensitive to sample size and observation session length, although its magnitude is influenced by the use of partial interval recording. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

In eye-movement control during reading, advanced process-oriented models have been developed to reproduce behavioral data. So far, model complexity and large numbers of model parameters prevented rigorous statistical inference and modeling of interindividual differences. Here we propose a Bayesian approach to both problems for one representative computational model of sentence reading (SWIFT; Engbert et al., Psychological Review, 112, 2005, pp. 777-813). We used experimental data from 36 subjects who read the text in a normal and one of four manipulated text layouts (e.g., mirrored and scrambled letters). The SWIFT model was fitted to subjects and experimental conditions individually to investigate between-subject variability. Based on posterior distributions of model parameters, fixation probabilities and durations are reliably recovered from simulated data and reproduced for withheld empirical data, at both the experimental condition and subject levels. A subsequent statistical analysis of model parameters across reading conditions generates model-driven explanations for observable effects between conditions. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

State trace analyses assess the latent dimensionality of a cognitive process by asking whether the means of two dependent variables conform to a monotonic function across a set of conditions. Using an assumption of independence between the measures, recently proposed statistical tests address bivariate measurement error, allowing both frequentist and Bayesian analyses of monotonicity (e.g., Davis-Stober, Morey, Gretton, & Heathcote, 2016; Kalish, Dunn, Burdakov, & Sysoev, 2016). However, statistical inference can be biased by unacknowledged dependencies between measures, particularly when the data are insufficient to overwhelm an incorrect prior assumption of independence. To address this limitation, we developed a hierarchical Bayesian model that explicitly models the separate roles of subject, item, and trial-level dependencies between two measures. Assessment of monotonicity is then performed by fitting separate models that do or do not allow a non-monotonic relation between the condition effects (i.e., same versus different rank orders). The Widely Applicable Information Criterion (WAIC) and Pseudo Bayesian Model Averaging – both cross validation measures of model fit – are used for model comparison, providing an inferential conclusion regarding the dimensionality of the latent psychological space. We validated this new state trace analysis technique using model recovery simulation studies, which assumed different ground truths regarding monotonicity and the direction/magnitude of the subject- and trial-level dependence. We also provide an example application of this new technique to a visual object learning study that compared performance on a visual retrieval task (forced choice part recognition) versus a verbal retrieval task (cued recall).

Personality and demographics are important variables in social sciences, while in NLP they can aid in interpretability and removal of societal biases. However, datasets with both personality and demographic labels are scarce. To address this, we present PANDORA, the first large-scale dataset of Reddit comments labeled with three personality models (including the well-established Big 5 model) and demographics (age, gender, and location) for more than 10k users. We showcase the usefulness of this dataset on three experiments, where we leverage the more readily available data from other personality models to predict the Big 5 traits, analyze gender classification biases arising from psycho-demographic variables, and carry out a confirmatory and exploratory analysis based on psychological theories. Finally, we present benchmark prediction models for all personality and demographic variables., Proceedings of the Ninth International Workshop on Natural Language Processing for Social Media, NAACL 2021, https://www.aclweb.org/anthology/2021.socialnlp-1.12

Steinley, Hoffman, Brusco, and Sher (2017) proposed a new method for evaluating the performance of psychological network models: fixed-margin sampling. The authors investigated LASSO regularized Ising models (eLasso) by generating random datasets with the same margins as the original binary dataset, and concluded that many estimated eLasso parameters are not distinguishable from those that would be expected if the data were generated by chance. We argue that fixed-margin sampling cannot be used for this purpose, as it generates data under a particular null-hypothesis: a unidimensional factor model with interchangeable indicators (i.e., the Rasch model). We show this by discussing relevant psychometric literature and by performing simulation studies. Results indicate that while eLasso correctly estimated network models and estimated almost no edges due to chance, fixed-margin sampling performed poorly in classifying true effects as “interesting” (Steinley et al. 2017, p. 1004). Further simulation studies indicate that fixed-margin sampling offers a powerful method for highlighting local misfit from the Rasch model, but performs only moderately in identifying global departures from the Rasch model. We conclude that fixed-margin sampling is not up to the task of assessing if results from estimated Ising models or other multivariate psychometric models are due to chance.

The link between mind, brain, and behavior has mystified philosophers and scientists for millennia. Recent progress has been made by forming statistical associations between manifest variables of the brain (e.g., EEG, fMRI) and manifest variables of behavior (e.g., response times, accuracy) through hierarchical latent variable models (Turner, Forstmann, & Steyvers, 2019). Within this framework, one can make inferences about the mind in a statistically principled way, such that complex patterns of brain-behavior associations drive the inference procedure. However, previous approaches were limited in the flexibility of the linking function, which has proven prohibitive for understanding the complex dynamics exhibited by the brain. In this article, we propose a data-driven, non-parametric approach that allows complex linking functions to emerge from fitting a hierarchical latent representation of the mind to multivariate, multimodal data. Furthermore, to enforce biological plausibility, we impose both spatial and temporal structure so that the types of realizable system dynamics are constrained. To illustrate the benefits of our approach, we investigate the model’s performance in a simulation study and apply it to experimental data. In the simulation study, we verify that the model can be accurately fit to simulated data, and latent dynamics can be well recovered. In an experimental application, we simultaneously fit the model to fMRI and behavioral data from a continuous motion tracking task. We show that the model accurately recovers both neural and behavioral data, and reveals interesting latent cognitive dynamics. Finally, we provide a test of the model’s generalizability by assessing its predictive accuracy in a cross-validation test.

A key assumption of models of human cognition is that there is variability in information processing. Evidence accumulation models (EAMs) commonly assume 2 broad variabilities in information processing: within-trial variability, which is thought to reflect moment-to-moment fluctuations in perceptual processes, and between-trial variability, which is thought to reflect variability in slower-changing processes like attention, or systematic variability between the stimuli on different trials. Recently, Ratcliff, Voskuilen, and McKoon (2018) claimed to "provide direct evidence that external noise is, in fact, required to explain the data from five simple two-choice decision tasks" (p. 33), suggesting that at least some portion of the between-trial variability in information processing is due to "noise." However, we argue that Ratcliff et al. (2018) failed to distinguish between 2 different potential sources of between-trial variability: random (i.e., "external noise") and systematic (e.g., item effects). Contrary to the claims of Ratcliff et al. (2018), we show that "external noise" is not required to explain their findings, as the same trends of data can be produced when only item effects are present. Furthermore, we contend that the concept of "noise" within cognitive models merely serves as a convenience parameter for sources of variability that we know exist but are unable to account for. Therefore, we question the usefulness of experiments aimed at testing the general existence of "random" variability and instead suggest that future research should attempt to replace the random variability terms within cognitive models with actual explanations of the process. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Human decisions often deviate from economic rationality and are influenced by cognitive biases. One such bias is the memory bias according to which people prefer choice options they have a better memory of—even when the options’ utilities are comparatively low. Although this phenomenon is well supported empirically, its cognitive foundation remains elusive. Here we test two conceivable computational accounts of the memory bias against each other. On the one hand, a single-process account explains the memory bias by assuming a single biased evidence-accumulation process in favor of remembered options. On the contrary, a dual-process account posits that some decisions are driven by a purely memory-driven process and others by a utility-maximizing one. We show that both accounts are indistinguishable based on choices alone as they make similar predictions with respect to the memory bias. However, they make qualitatively different predictions about response times. We tested the qualitative and quantitative predictions of both accounts on behavioral data from a memory-based decision-making task. Our results show that a single-process account provides a better account of the data, both qualitatively and quantitatively. In addition to deepening our understanding of memory-based decision-making, our study provides an example of how to rigorously compare single- versus dual-process models using empirical data and hierarchical Bayesian parameter estimation methods.

A major problem resulting from the massive use of social media is the potential spread of incorrect information. Yet, very few studies have investigated the impact of incorrect information on individual and collective decisions. We performed experiments in which participants had to estimate a series of quantities, before and after receiving social information. Unbeknownst to them, we controlled the degree of inaccuracy of the social information through ‘virtual influencers’, who provided some incorrect information. We find that a large proportion of individuals only partially follow the social information, thus resisting incorrect information. Moreover, incorrect information can help improve group performance more than correct information, when going against a human underestimation bias. We then design a computational model whose predictions are in good agreement with the empirical data, and sheds light on the mechanisms underlying our results. Besides these main findings, we demonstrate that the dispersion of estimates varies a lot between quantities, and must thus be considered when normalizing and aggregating estimates of quantities that are very different in nature. Overall, our results suggest that incorrect information does not necessarily impair the collective wisdom of groups, and can even be used to dampen the negative effects of known cognitive biases.

Multiple learning systems allow individuals to flexibly respond to opportunities and challenges present in the environment. An evolutionarily conserved “Pavlovian” learning mechanism couples valence and action, promoting a tendency to approach cues associated with reward and to inhibit action in the face of anticipated punishment. Although this default response system may be adaptive, these hard-wired reactions can hinder the ability to learn flexible “instrumental” actions in pursuit of a goal. Such constraints on behavioral flexibility have been studied extensively in adults. However, the extent to which these valence-specific response tendencies bias instrumental learning across development remains poorly characterized. Here, we show that while Pavlovian response biases constrain flexible action learning in children and adults, these biases are attenuated in adolescents. This adolescent-specific reduction in Pavlovian bias may promote unbiased exploration of approach and avoidance responses, facilitating the discovery of rewarding behavior in the many novel contexts that adolescents encounter.

International audience; We elaborated an index, the Interference Distribution Index, which allows quantifying the relation between response times and the size of the interference effect. This index is associated with an intuitive graphical representation, the Lorenz-interference plot. We show that this index has some convenient properties in terms of sensitivity to changes in the distribution of the interference effect and to aggregation of individual data. Moreover, it turns out that this index is the only one (up to an arbitrary increasing transformation) possessing these properties. The relevance of this index is illustrated through simulations of a cognitive model of interference effects and reanalysis of experimental data.

Moral intuitions are a central motivator in human behavior. Recent work highlights the importance of moral intuitions for understanding a wide range of issues ranging from online radicalization to vaccine hesitancy. Extracting and analyzing moral content in messages, narratives, and other forms of public discourse is a critical step toward understanding how the psychological influence of moral judgments unfolds at a global scale. Extant approaches for extracting moral content are limited in their ability to capture the intuitive nature of moral sensibilities, constraining their usefulness for understanding and predicting human moral behavior. Here we introduce the extended Moral Foundations Dictionary (eMFD), a dictionary-based tool for extracting moral content from textual corpora. The eMFD, unlike previous methods, is constructed from text annotations generated by a large sample of human coders. We demonstrate that the eMFD outperforms existing approaches in a variety of domains. We anticipate that the eMFD will contribute to advance the study of moral intuitions and their influence on social, psychological, and communicative processes. All data and code for our analyses can be found here. We also created software -eMFDscore - for easy scoring of text via our eMFD. The software is available at https://github.com/medianeuroscience/emfdscore.

Generic language (e.g., “Birds fly”) conveys generalizations about categories and is essential for learning beyond our direct experience. The meaning of generic language is notoriously hard to specify, however (e.g., penguins don’t fly). Tessler and Goodman (2019) proposed a model for generics that is mathematically equivalent to Bayesian belief-updating based on a single pedagogical example, suggesting a deep connection be- tween learning from experience and learning from language. Relatedly, Csibra and Shamsudheen (2015) argue that generics are inherently pedagogical, understood by infants as referring to a member of a kind. In two experiments with adults, we quantify the exchange-rate between generics and observations by relating their belief-updating capacity, varying both the number of observations and whether they are presented pedagogically or incidentally. We find generics convey stronger generalizations than single pedagogical observations (Expt. 1), even when the property is explicitly demarcated (Expt. 2). We suggest revisions to the vague quantifier model of generics that would allow it to accommodate this intriguing exchange-rate.

Language exhibits striking systematicity in its form-meaning mappings: Similar meanings are assigned similar forms. Here we study how systematicity relates to another well-studied phenomenon, linguistic regularization, the removal of unpredictable variation in linguistic variants. Systematicity is ultimately a property of classes of form-meaning mappings, each member of which can be acted upon independently by linguistic regularization. Both are supported by a cognitive bias for simplicity, but this leaves open the question of how they interact to structure the lexicon. Using data from a recent artificial gesture learning experiment by Verhoef, Padden, and Kirby(2016), we formalize cognitive biases at the item level and the language level as inductive biases in a hierarchical Bayesian model. Simulated data from models that lack either one of those biases show how both are necessary to capture subjects' systematicity preferences. Our results bring conceptual clarity about the relationship between regularization and systematicity and promote a multi-level approach to cognitive biases in artificial language learning and language evolution.

When learning to partition the world into categories, people rely on a set of assumptions (overhypotheses) about possible category structures. Here we propose that the nature of these overhypotheses depends on the presence of a verbal label associated with a given category. We describe a computational model that demonstrates how labels can either accelerate or hinder category learning, depending on whether or not the prior beliefs imposed by their presence align with the true category structure. This account provides an explanation for the phenomena described in prior experimental work (Lupyan, Rakison, & McClelland, 2007; Brojde, Porter, & Colunga, 2011) that have remained unexplained by other models. Based on these results, we argue that the overhypothesis theory of label effects provides a way to formalize and quantify the effect of language on category learning and to develop a more precise delineation between linguistic and non-linguistic thought.

The mind and consciousness are pretty complex to observe and analyze quantitatively since the complexity of neuron and mystery of consciousness. Recent years, many scientists try to explain our mind from physical viewpoint or elucidate the physics of the mind. In the paper, we reveal a new approach to understand our mind and present a tentative theory to describe how our mind works respect to physics. Two significant concept of time, namely, the internal time of mind and external time of the outside are introduced to measure our mind quantitatively on some degree. The related law of these two types of time is similar to the special relativity proposed by Einstein. Our theory aims to provide a new perspective to understand and know ourselves and lead our life better. At the end of paper, we design an experiment to check the correctness of our results.

Contemporary theories of choice posit that decision making is a constructive process in which a decision maker uses information about the choice options to generate support for various decisions and judgments, then uses these decisions and judgments to reduce their uncertainty about their own preferences. Here we examine how these constructive processes unfold by tracking dynamic changes in preference strength. Across two experiments, we observed that mean preference strength oscillated over time and found that eliciting a choice strongly affected the pattern of oscillation. Preferences following choices oscillated between being stronger than those without prior choice (bolstering) and being weaker than those without choice (suppression). An open system model, merging epistemic uncertainty about how a person reacts to options and ontic uncertainty about how their preference is affected by choice, accounts for the oscillations resulting in both bolstering and suppression effects.

Response time information has recently attracted a significant amount of attention in the literature as it may provide meaningful information about item preknowledge. The methods that propose the use of response time information in identifying examinees with potential item preknowledge make an implicit assumption that the examinees with item preknowledge differ in their response time patterns compared to other examinees without item preknowledge. In this study, we analyzed the differences in response time of examinees with potential item preknowledge and examinees without item preknowledge based on a real experimental dataset provided by Toton and Maynes (2019). A multiple-group extension of van der Linden’s Lognormal Response Model with a gating mechanism was used to capture the differences in latent speed for control and experimental groups on disclosed and undisclosed items. The model used in the study and estimated parameters from this experimental dataset may inform future simulation studies in this area of research to simulate realistic datasets with item preknowledge behavior.

Response time (RT) information has recently attracted a significant amount of attention in the literature as it may provide meaningful information about item preknowledge. In this study, a Deterministic Gated Lognormal Response Time (DG-LNRT) model is proposed to identify examinees with potential item preknowledge using RT information. The proposed model is applied to a real experimental dataset provided by Toton and Maynes (2019) in which item preknowledge was manipulated, and its performance is demonstrated. Then, the performance of the DG-LNRT model is investigated through a simulation study. The model is estimated using the Bayesian framework via Stan. The results indicate that the proposed model is viable and has the potential to be useful in detecting cheating by using response time differences between compromised and uncompromised items.

The past few years were marked by increased online offensive strategies perpetrated by state and non-state actors to promote their political agenda, sow discord, and question the legitimacy of democratic institutions in the US and Western Europe. In 2016, the US congress identified a list of Russian state-sponsored Twitter accounts that were used to try to divide voters on a wide range of issues. Previous research used latent Dirichlet allocation (LDA) to estimate latent topics in data extracted from these accounts. However, LDA has characteristics that may limit the effectiveness of its use on data from social media: The number of latent topics must be specified by the user, interpretability of the topics can be difficult to achieve, and it does not model short-term temporal dynamics. In the current paper, we propose a new method to estimate latent topics in texts from social media termed Dynamic Exploratory Graph Analysis (DynEGA). In a Monte Carlo simulation, we compared the ability of DynEGA and LDA to estimate the number of simulated latent topics. The results show that DynEGA is substantially more accurate than several different LDA algorithms when estimating the number of simulated topics. In an applied example, we performed DynEGA on a large dataset with Twitter posts from state-sponsored right- and left-wing trolls during the 2016 US presidential election. DynEGA revealed topics that were pertinent to several consequential events in the election cycle, demonstrating the coordinated effort of trolls capitalizing on current events in the USA. This example demonstrates the potential power of our approach for revealing temporally relevant information from qualitative text data.

Recently, the whole of Europe, including Poland, have been significantly affected by COVID-19 and its social and economic consequences which are already causing dozens of billions of euros monthly losses in Poland alone. Social behaviour has a fundamental impact on the dynamics of spread of infectious diseases such as SARS-CoV-2, challenging the existing health infrastructure and social organization. Modelling and understanding mechanisms of social behaviour (e.g. panic and social distancing) and its contextualization with regard to Poland can contribute to better response to the outbreak on a national and local level. In the presented study we aim to investigate the impact of the COVID-19 on society by: (i) measuring the relevant activity in internet news and social media; (ii) analysing attitudes and demographic patterns in Poland. In the end, we are going to implement computational social science and digital epidemiology research approach to provide urgently needed information on social dynamics during the outbreak. This study is an ad hoc reaction only and our goal is to signal the main areas of possible research to be done in future and cover issues with direct or indirect relation to public health.

Computational implementations of Theory of Mind (ToM), the ability to attribute mental states to others, has been used to investigate a variety of issues. This includes the effect of framing effects on, or inter-species differences in, ability to do ToM (Devaine et al., 2014a, 2017), ToM in autists (d’Arc et al., 2018), or providing an explanation for the apparent limits on human ability to do ToM recursively (Devaine et al., 2014b). It has been implemented in the VBA package for Matlab (Daunizeau et al., 2014), but not in any free and open-source software. Therefore this thesis presents the Theory of Mind simulation using Python (tomsup) package. The tomsup package provides accessible tools for running agent-based models in a game theory context, and allows the implementation of a computational model of ToM, either in agent-based models or in interaction with a human player. The implementation of the ToM model was originally proposed by Yoshida et al. (2008), and was developed by drawing on the Free Energy Principle (Friston, 2010) to its current form as it is in Devaine et al. (2017), where it is generalized to any 2-player game which can be operationalized as a 2-by-2 payoff matrix. Importantly, the ToM implementation introduces a sophistication level k, which determines how many recursive simulations of its opponent it can perform, hereby assuming bounded rationality (Kahneman, 2003). An agent using the ToM model, denoted as k-ToM, uses a variational Bayes Laplace approximation (Daunizeau, 2017b) on a turn-by-turn basis to infer its opponent’s model parameters and sophistication level, based on which it predicts the opponent’s choice and acts accordingly. An agent-based model simulation using the competitive matching pennies game was done to perform a prelim- inary investigation of the behaviour of the k-ToM model. Most importantly, it was found that k-ToM’s prior beliefs about its opponent have a notable effect on its performance, even over many trials, warranting further research into how its priors should be formed. Various ways are suggested in which the tomsup package and the k-ToM model could be applied and developed further, as well as a discussion on how to make it broadly available, so as to scaffold future research using computational ToM models.

Evidence accumulation models (EAMs) have become the dominant explanation of how the decision-making process operates, proposing that decisions are the result of a process of evidence accumulation. The primary use of EAMs has been as "measurement tools" of the underlying decision-making process, where researchers apply EAMs to empirical data to estimate participants' task ability (i.e., the "drift rate"), response caution (i.e., the "decision threshold"), and the time taken for other processes (i.e., the "non-decision time"), making EAMs a powerful tool for discriminating between competing psychological theories. Recent studies have brought into question the mapping between the latent parameters of EAMs and the theoretical constructs that they are thought to represent, showing that emphasizing urgent responding -- which intuitively should selectively influence decision threshold -- may also influence drift rate and/or non-decision time. However, these findings have been mixed, leading to differences in opinion between experts in the field. The current study aims to provide a more conclusive answer to the implications of emphasizing urgent responding, providing a re-analyse of 6 data sets from previous studies using two different EAMs -- the diffusion model and the linear ballistic accumulator (LBA) -- with state-of-the-art methods for model selection based inference. The findings display clear evidence for a difference in conclusions between the two models, with the diffusion model suggesting that decision threshold and non-decision time decrease when urgency is emphasized, and the LBA suggesting that decision threshold and drift rate decrease when urgency is emphasized. Furthermore, although these models disagree regarding whether non-decision time or drift rate decrease under urgency emphasis, both show clear evidence that emphasizing urgency does not selectively influence decision threshold. These findings suggest that researchers should revise their assumptions about certain experimental manipulations, the specification of certain EAMs, or perhaps both.

When faced with an imminent collision threat, human vehicle drivers respond with braking in a manner which is stereotypical, yet modulated in complex ways by many factors, including the specific traffic situation and past driver eye movements. A computational model capturing these phenomena would have high applied value, for example in virtual vehicle safety testing methods, but existing models are either simplistic or not sufficiently validated. This paper extends an existing quantitative driver model for initiation and modulation of pre-crash brake response, to handle off-road glance behavior. The resulting models are fitted to time-series data from real-world naturalistic rear-end crashes and near-crashes. A stringent parameterization and model selection procedure is presented, based on particle swarm optimization and maximum likelihood estimation. A major contribution of this paper is the resulting first-ever fit of a computational model of human braking to real near-crash and crash behavior data. The model selection results also permit novel conclusions regarding behavior and accident causation: Firstly, the results indicate that drivers have partial visual looming perception during off-road glances; that is, evidence for braking is collected, albeit at a slower pace, while the driver is looking away from the forward roadway. Secondly, the results suggest that an important causation factor in crashes without off-road glances may be a reduced responsiveness to visual looming, possibly associated with cognitive driver state (e.g., drowsiness or erroneous driver expectations). It is also demonstrated that a model parameterized on less-critical data, such as near-crashes, may also accurately reproduce driver behavior in highly critical situations, such as crashes.

What kind of dynamic decision process do humans use to make decisions? In this article, two different types of processes are reviewed and compared: Markov and quantum. Markov processes are based on the idea that at any given point in time a decision maker has a definite and specific level of support for available choice alternatives, and the dynamic decision process is represented by a single trajectory that traces out a path across time. When a response is requested, a person's decision or judgment is generated from the current location along the trajectory. By contrast, quantum processes are founded on the idea that a person's state can be represented by a superposition over different degrees of support for available choice options, and that the dynamics of this state form a wave moving across levels of support over time. When a response is requested, a decision or judgment is constructed out of the superposition by "actualizing" a specific degree or range of degrees of support to create a definite state. The purpose of this article is to introduce these two contrasting theories, review empirical studies comparing the two theories, and identify conditions that determine when each theory is more accurate and useful than the other. This article is categorized under: Economics > Individual Decision-Making Psychology > Reasoning and Decision Making Psychology > Theory and Methods.

**published as part of "Machine Learning for Health (ML4H) at NeurIPS 2020: https://arxiv.org/abs/2011.01725**Introduction: Hierarchical model fitting has become commonplace in cognitive neuroscience. However, these techniques require us to formalise assumptions about the data-generating process at the group level, which may not be known. In computational psychiatry, we frequently encounter situations with multiple distinct groups of subjects, which makes this specification non-trivial. Specifically, we must choose whether to assume all subjects are drawn from a common population, or to model them as deriving from separate populations. These assumptions have profound implications for computational psychiatry, as they will undoubtedly affect the resulting inference (parameter recovery) and conflate or mask true group-level differences. Methods: We used simulations to test these assumptions on synthetic multi-group behavioural data from commonly used multi-armed bandit tasks. We first demonstrate the influence of hierarchical vs non hierarchical approaches on parameter recovery. We then examine recovery of group differences in parameter values under different modelling approaches: (1) modelling groups under a common shared group-level prior (assuming all participants are generated from a common distribution, and are likely to share common characteristics); (2) modelling separate groups based on symptomatology or diagnostic labels, resulting in separate group-level priors. We evaluate the robustness of these approaches to variations in data quality and in prior specification.Results: We confirm, firstly, that hierarchical approaches lead to better parameter recovery. Secondly, we argue that fitting groups separately (Model 2) provides the most robust inference across all conditions and perturbations. Thirdly, we confirm these assumptions have less of an impact with higher numbers of trials and subjects. Finally, we show that the advice to fit groups separately can fall down when extreme assumptions are made about group variance.Conclusions: Our results suggest that when dealing with data from multiple clinical groups, researchers should use a hierarchical approach to parameter fitting, but account for hypothesised group differences in the model fitting procedure. In other words, under most normal circumstances, researchers should fit patient and control groups separately.

Classical dynamic theories of decision making assume that responses are triggered by accumulating a threshold amount of information. Recently, there has been a growing appreciation that the passage of time also plays a role in triggering responses. We propose that decision processes are composed of two diffusive accumulation mechanisms---one evidence-based and one time-based---that compete in an independent race architecture. We show that this Timed Racing Diffusion Model (TRDM) provides a unified, comprehensive and quantitatively accurate explanation of key decision phenomena---including the effects of implicit and explicit deadlines and the relative speed of correct and error responses under speed-accuracy tradeoffs---without requiring additional mechanisms that have been criticized as being ad-hoc in theoretical motivation and difficult to estimate, such as trial-to-trial variability parameters, collapsing thresholds, or urgency signals. In contrast, our addition is grounded in a widely validated account of time-estimation performance, enabling the same mechanism to simultaneously account for interval estimation and decision making with an explicit deadline.

According to Spearman’s law of diminishing returns (SLODR), IQ test scores are more g saturated among those with low, compared to those with high, ability. The present simulation shows that such difference in saturation can be observed if test scores are affected by a disturbing factor, for example, low motivation, illness, or linguistic confusion, that varies in magnitude between individuals. More contemporary criteria of SLODR can also be satisfied if test scores are affected by disturbance, especially if the disturbance variable is negatively skewed. This indicates a possible threat against the validity of findings supporting SLODR and points at the importance for researchers to try to eliminate the influence of such disturbing factors from their studies.

Conflict tasks are one of the most widely studied paradigms within cognitive psychology, where participants are required to respond based on relevant sources of information while ignoring conflicting irrelevant sources of information. The flanker task, in particular, has been the focus of considerable modeling efforts, with only 3 models being able to provide a complete account of empirical choice response time distributions: the dual-stage 2-phase model (DSTP), the shrinking spotlight model (SSP), and the diffusion model for conflict tasks (DMC). Although these models are grounded in different theoretical frameworks, can provide diverging measures of cognitive control, and are quantitatively distinguishable, no previous study has compared all 3 of these models in their ability to account for empirical data. Here, we perform a comparison of the precise quantitative predictions of these models through Bayes factors, using probability density approximation to generate a pseudolikelihood estimate of the unknown probability density function, and thermodynamic integration via differential evolution to approximate the analytically intractable Bayes factors. We find that for every participant across 3 data sets from 3 separate research groups, DMC provides an inferior account of the data to DSTP and SSP, which has important theoretical implications regarding cognitive processes engaged in the flanker task, and practical implications for applying the models to flanker data. More generally, we argue that our combination of probability density approximation with marginal likelihood approximation-which we term pseudolikelihood Bayes factors-provides a crucial step forward for the future of model comparison, where Bayes factors can be calculated between any models that can be simulated. We also discuss the limitations of simulation-based methods, such as the potential for approximation error, and suggest that researchers should use analytically or numerically computed likelihood functions when they are available and computationally tractable. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Understanding the temporal dynamics of affect is crucial for our understanding human emotions in general. In this study, we empirically test a computational model of affective dynamics by analyzing a large-scale dataset of Facebook status updates using text analysis techniques. Our analyses support the central assumptions of our model: After stimulation, affective states, quantified as valence and arousal, exponentially return to an individual-specific baseline. On average, this baseline is at a slightly positive valence value and at a moderate arousal point below the midpoint. Furthermore, affective expression, in this case posting a status update on Facebook, immediately pushes arousal and valence towards the baseline by a proportional value. These results are robust to the choice of the text analysis technique and illustrate the fast timescale of affective dynamics through social media text. These outcomes are of high relevance for affective computing, the detection and modeling of collective emotions, the refinement of psychological research methodology, and the detection of abnormal, and potentially pathological, individual affect dynamics.

The Savage-Dickey density ratio is a simple method for computing the Bayes factor for an equality constraint on one or more parameters of a statistical model. In regression analysis, this includes the important scenario of testing whether one or more of the covariates have an effect on the dependent variable. However, the Savage-Dickey ratio only provides the correct Bayes factor if the prior distribution of the nuisance parameters under the nested model is identical to the conditional prior under the full model given the equality constraint. This condition is violated for multiple regression models with a Jeffreys-Zellner-Siow (JZS) prior, which is often used as a default prior in psychology. Besides linear regression models, the limitation of the Savage-Dickey ratio is especially relevant when analytical solutions for the Bayes factor are not available. This is the case for generalized linear models, nonlinear models, or cognitive process models with regression extensions. As a remedy, the correct Bayes factor can be computed using a generalized version of the Savage-Dickey density ratio.