arXiv subjects

Shravan Vasishth

Publications and source records attributed to Shravan Vasishth.

At least 19 recordsLinked to original sources

arXiv2026

Deconstructing sentence disambiguation by joint latent modeling of reading paradigms: LLM surprisal is not enough

Using temporarily ambiguous garden-path sentences ("While the team trained the striker wondered ...") as a test case, we present a latent-process mixture model of human reading behavior across four different reading paradigms (eye tracking, uni- and bidirectional self-paced reading, Maze). The model distinguishes between garden-path probability, garden-path cost, and reanalysis cost, and yields more realistic processing cost estimates by taking into account trials with inattentive reading. We show that the model is able to reproduce empirical patterns with regard to rereading behavior, comprehension question responses, and grammaticality judgments. Cross-validation reveals that the mixture model also has better predictive fit to human reading patterns and end-of-trial task data than a mixture-free model based on GPT-2-derived surprisal values. We discuss implications for future work.

cs.CL↗

arXiv2023

SEAM: An Integrated Activation-Coupled Model of Sentence Processing and Eye Movements in Reading

Models of eye-movement control during reading, developed largely within psychology, usually focus on visual, attentional, lexical, and motor processes but neglect post-lexical language processing; by contrast, models of sentence comprehension processes, developed largely within psycholinguistics, generally focus only on post-lexical language processes. We present a model that combines these two research threads, by integrating eye-movement control and sentence processing. Developing such an integrated model is extremely challenging and computationally demanding, but such an integration is an important step toward complete mathematical models of natural language comprehension in reading. We combine the SWIFT model of eye-movement control (Seelig et al., 2020, doi:10.1016/j.jmp.2019.102313) with key components of the Lewis and Vasishth sentence processing model (Lewis & Vasishth, 2005, doi:10.1207/s15516709cog0000_25). This integration becomes possible, for the first time, due in part to recent advances in successful parameter identification in dynamical models, which allows us to investigate profile log-likelihoods for individual model parameters. We present a fully implemented proof-of-concept model demonstrating how such an integrated model can be achieved; our approach includes Bayesian model inference with Markov Chain Monte Carlo (MCMC) sampling as a key computational tool. The integrated Sentence-Processing and Eye-Movement Activation-Coupled Model (SEAM) can successfully reproduce eye movement patterns that arise due to similarity-based interference in reading. To our knowledge, this is the first-ever integration of a complete process model of eye-movement control with linguistic dependency completion processes in sentence comprehension. In future work, this proof of concept model will need to be evaluated using a comprehensive set of benchmark data.

q-bio.NC↗

arXiv2023

Data aggregation can lead to biased inferences in Bayesian linear mixed models and Bayesian ANOVA: A simulation study

Bayesian linear mixed-effects models and Bayesian ANOVA are increasingly being used in the cognitive sciences to perform null hypothesis tests, where a null hypothesis that an effect is zero is compared with an alternative hypothesis that the effect exists and is different from zero. While software tools for Bayes factor null hypothesis tests are easily accessible, how to specify the data and the model correctly is often not clear. In Bayesian approaches, many authors use data aggregation at the by-subject level and estimate Bayes factors on aggregated data. Here, we use simulation-based calibration for model inference applied to several example experimental designs to demonstrate that, as with frequentist analysis, such null hypothesis tests on aggregated data can be problematic in Bayesian analysis. Specifically, when random slope variances differ (i.e., violated sphericity assumption), Bayes factors are too conservative for contrasts where the variance is small and they are too liberal for contrasts where the variance is large. Running Bayesian ANOVA on aggregated data can - if the sphericity assumption is violated - likewise lead to biased Bayes factor results. Moreover, Bayes factors for by-subject aggregated data are biased (too liberal) when random item slope variance is present but ignored in the analysis. These problems can be circumvented or reduced by running Bayesian linear mixed-effects models on non-aggregated data such as on individual trials, and by explicitly modeling the full random effects structure. Reproducible code is available from \url{https://osf.io/mjf47/}.

stat.ME↗

arXiv2023

When words collide: Bayesian meta-analyses of distractor and target properties in the picture-word interference paradigm

In the picture-word interference paradigm, participants name pictures while ignoring a written or spoken distractor word. Naming times to the pictures are slowed down by the presence of the distractor word. Various properties of the distractor modulate this slow down, for example naming times are shorter with frequent vs. infrequent distractors. Building on this line of research, the present study investigates in more detail the impact of distractor and target word properties on picture naming times. We report the results of several Bayesian meta-analyses, based on 35 datasets. The aim of the first analysis was to obtain an estimation of the size of the distractor frequency effect, and of its precision, in typical picture-word interference experiments where this variable is not manipulated. The analysis shows that a one-unit increase in log frequency results in response times to the pictures decreasing by about 4ms (95% Credible Interval: [-6, -2]). With the second and third analyses, we show that after accounting for the effect of frequency, two variables known to influence processing times in visual word processing tasks also influence picture naming times: distractor length and orthographic neighborhood. Finally, we found that distractor word frequency and target word frequency interact; the effect of distractor frequency decreases as the frequency of the target word increases. We discuss the theoretical and methodological implications of these findings, as well as the importance of obtaining high-precision estimates of experimental effects.

q-bio.NC↗

arXiv2022

The posterior probability of a null hypothesis given a statistically significant result

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.

stat.AP↗

arXiv2021

Workflow Techniques for the Robust Use of Bayes Factors

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. Moreover it's not clear how straightforwardly this approach can be implemented in practice, and in particular how sensitive it is to the details of the computational implementation. Here, we investigate these questions for Bayes factor analyses in the cognitive sciences. We explain the statistics underlying Bayes factors as a tool for Bayesian inferences and discuss that utility functions are needed for principled decisions on hypotheses. Next, we study how Bayes factors misbehave under different conditions. This includes a study of errors in the estimation of Bayes factors. Importantly, it is unknown whether Bayes factor estimates based on bridge sampling are unbiased for complex analyses. We are the first to use simulation-based calibration as a tool to test the accuracy of Bayes factor estimates. Moreover, we study how stable Bayes factors are against different MCMC draws. We moreover study how Bayes factors depend on variation in the data. We also look at variability of decisions based on Bayes factors and how to optimize decisions 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, and we illustrate this workflow using an example from the cognitive sciences. We hope that this study will provide a workflow to test the strengths and limitations of Bayes factors as a way to quantify evidence in support of scientific hypotheses. Reproducible code is available from https://osf.io/y354c/.

stat.ME↗

arXiv2020

Toward a principled Bayesian workflow in cognitive science

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. All data and code accompanying this paper are available from https://osf.io/b2vx9/.

stat.ME↗

arXiv2019

How to capitalize on a priori contrasts in linear (mixed) models: A tutorial

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-test. Reproducible code is available from https://osf.io/7ukf6/.

stat.ME↗

arXiv2018

Parsimonious Mixed Models

The analysis of experimental data with mixed-effects models requires decisions about the specification of the appropriate random-effects structure. Recently, Barr, Levy, Scheepers, and Tily, 2013 recommended fitting `maximal' models with all possible random effect components included. Estimation of maximal models, however, may not converge. We show that failure to converge typically is not due to a suboptimal estimation algorithm, but is a consequence of attempting to fit a model that is too complex to be properly supported by the data, irrespective of whether estimation is based on maximum likelihood or on Bayesian hierarchical modeling with uninformative or weakly informative priors. Importantly, even under convergence, overparameterization may lead to uninterpretable models. We provide diagnostic tools for detecting overparameterization and guiding model simplification.

stat.ME↗

arXiv2018

Feature overwriting as a finite mixture process: Evidence from comprehension data

The ungrammatical sentence "The key to the cabinets are on the table" is known to lead to an illusion of grammaticality. As discussed in the meta-analysis by Jaeger et al., 2017, faster reading times are observed at the verb are in the agreement-attraction sentence above compared to the equally ungrammatical sentence "The key to the cabinet are on the table". One explanation for this facilitation effect is the feature percolation account: the plural feature on cabinets percolates up to the head noun key, leading to the illusion. An alternative account is in terms of cue-based retrieval (Lewis & Vasishth, 2005), which assumes that the non-subject noun cabinets is misretrieved due to a partial feature-match when a dependency completion process at the auxiliary initiates a memory access for a subject with plural marking. We present evidence for yet another explanation for the observed facilitation. Because the second sentence has two nouns with identical number, it is possible that these are, in some proportion of trials, more difficult to keep distinct, leading to slower reading times at the verb in the first sentence above; this is the feature overwriting account of Nairne, 1990. We show that the feature overwriting proposal can be implemented as a finite mixture process. We reanalysed ten published data-sets, fitting hierarchical Bayesian mixture models to these data assuming a two-mixture distribution. We show that in nine out of the ten studies, a mixture distribution corresponding to feature overwriting furnishes a superior fit over both the feature percolation and the cue-based retrieval accounts.

stat.ML↗

arXiv2017

Models of retrieval in sentence comprehension: A computational evaluation using Bayesian hierarchical modeling

Research on interference has provided evidence that the formation of dependencies between non-adjacent words relies on a cue-based retrieval mechanism. Two different models can account for one of the main predictions of interference, i.e., a slowdown at a retrieval site, when several items share a feature associated with a retrieval cue: Lewis and Vasishth's (2005) activation-based model and McElree's (2000) direct access model. Even though these two models have been used almost interchangeably, they are based on different assumptions and predict differences in the relationship between reading times and response accuracy. The activation-based model follows the assumptions of ACT-R, and its retrieval process behaves as a lognormal race between accumulators of evidence with a single variance. Under this model, accuracy of the retrieval is determined by the winner of the race and retrieval time by its rate of accumulation. In contrast, the direct access model assumes a model of memory where only the probability of retrieval varies between items; in this model, differences in latencies are a by-product of the possibility and repairing incorrect retrievals. We implemented both models in a Bayesian hierarchical framework in order to evaluate them and compare them. We show that some aspects of the data are better fit under the direct access model than under the activation-based model. We suggest that this finding does not rule out the possibility that retrieval may be behaving as a race model with assumptions that follow less closely the ones from the ACT-R framework. We show that by introducing a modification of the activation model, i.e, by assuming that the accumulation of evidence for retrieval of incorrect items is not only slower but noisier (i.e., different variances for the correct and incorrect items), the model can provide a fit as good as the one of the direct access model.

cs.CL↗

arXiv2017

A computational investigation of sources of variability in sentence comprehension difficulty in aphasia

We present a computational evaluation of three hypotheses about sources of deficit in sentence comprehension in aphasia: slowed processing, intermittent deficiency, and resource reduction. The ACT-R based Lewis and Vasishth (2005) model is used to implement these three proposals. Slowed processing is implemented as slowed default production-rule firing time; intermittent deficiency as increased random noise in activation of chunks in memory; and resource reduction as reduced goal activation. As data, we considered subject vs. object rela- tives whose matrix clause contained either an NP or a reflexive, presented in a self-paced listening modality to 56 individuals with aphasia (IWA) and 46 matched controls. The participants heard the sentences and carried out a picture verification task to decide on an interpretation of the sentence. These response accuracies are used to identify the best parameters (for each participant) that correspond to the three hypotheses mentioned above. We show that controls have more tightly clustered (less variable) parameter values than IWA; specifically, compared to controls, among IWA there are more individuals with low goal activations, high noise, and slow default action times. This suggests that (i) individual patients show differential amounts of deficit along the three dimensions of slowed processing, intermittent deficient, and resource reduction, (ii) overall, there is evidence for all three sources of deficit playing a role, and (iii) IWA have a more variable range of parameter values than controls. In sum, this study contributes a proof of concept of a quantitative implementation of, and evidence for, these three accounts of comprehension deficits in aphasia.

cs.CL↗

arXiv2017

The statistical significance filter leads to overconfident expectations of replicability

We show that publishing results using the statistical significance filter---publishing only when the p-value is less than 0.05---leads to a vicious cycle of overoptimistic expectation of the replicability of results. First, we show analytically that when true statistical power is relatively low, computing power based on statistically significant results will lead to overestimates of power. Then, we present a case study using 10 experimental comparisons drawn from a recently published meta-analysis in psycholinguistics (Jäger et al., 2017). We show that the statistically significant results yield an illusion of replicability. This illusion holds even if the researcher doesn't conduct any formal power analysis but just uses statistical significance to informally assess robustness (i.e., replicability) of results.

stat.ME↗

arXiv2017

Modelling dependency completion in sentence comprehension as a Bayesian hierarchical mixture process: A case study involving Chinese relative clauses

We present a case-study demonstrating the usefulness of Bayesian hierarchical mixture modelling for investigating cognitive processes. In sentence comprehension, it is widely assumed that the distance between linguistic co-dependents affects the latency of dependency resolution: the longer the distance, the longer the retrieval time (the distance-based account). An alternative theory, direct-access, assumes that retrieval times are a mixture of two distributions: one distribution represents successful retrievals (these are independent of dependency distance) and the other represents an initial failure to retrieve the correct dependent, followed by a reanalysis that leads to successful retrieval. We implement both models as Bayesian hierarchical models and show that the direct-access model explains Chinese relative clause reading time data better than the distance account.

stat.AP↗

arXiv2017

Balancing Type I Error and Power in Linear Mixed Models

Linear mixed-effects models have increasingly replaced mixed-model analyses of variance for statistical inference in factorial psycholinguistic experiments. Although LMMs have many advantages over ANOVA, like ANOVAs, setting them up for data analysis also requires some care. One simple option, when numerically possible, is to fit the full variance-covariance structure of random effects (the maximal model; Barr et al. 2013), presumably to keep Type I error down to the nominal alpha in the presence of random effects. Although it is true that fitting a model with only random intercepts may lead to higher Type I error, fitting a maximal model also has a cost: it can lead to a significant loss of power. We demonstrate this with simulations and suggest that for typical psychological and psycholinguistic data, higher power is achieved without inflating Type I error rate if a model selection criterion is used to select a random effect structure that is supported by the data.

stat.AP↗

arXiv2016

The cave of Shadows. Addressing the human factor with generalized additive mixed models

Generalized additive mixed models are introduced as an extension of the generalized linear mixed model which makes it possible to deal with temporal autocorrelational structure in experimental data. This autocorrelational structure is likely to be a consequence of learning, fatigue, or the ebb and flow of attention within an experiment (the `human factor'). Unlike molecules or plots of barley, subjects in psycholinguistic experiments are intelligent beings that depend for their survival on constant adaptation to their environment, including the environment of an experiment. Three data sets illustrate that the human factor may interact with predictors of interest, both factorial and metric. We also show that, especially within the framework of the generalized additive model, in the nonlinear world, fitting maximally complex models that take every possible contingency into account is ill-advised as a modeling strategy. Alternative modeling strategies are discussed for both confirmatory and exploratory data analysis.

stat.AP↗

arXiv2016

Introduction: Cognitive Issues in Natural Language Processing

This special issue is dedicated to get a better picture of the relationships between computational linguistics and cognitive science. It specifically raises two questions: "what is the potential contribution of computational language modeling to cognitive science?" and conversely: "what is the influence of cognitive science in contemporary computational linguistics?"

cs.CL↗

arXiv2016

Statistical methods for linguistic research: Foundational Ideas - Part II

We provide an introductory review of Bayesian data analytical methods, with a focus on applications for linguistics, psychology, psycholinguistics, and cognitive science. The empirically oriented researcher will benefit from making Bayesian methods part of their statistical toolkit due to the many advantages of this framework, among them easier interpretation of results relative to research hypotheses, and flexible model specification. We present an informal introduction to the foundational ideas behind Bayesian data analysis, using, as an example, a linear mixed models analysis of data from a typical psycholinguistics experiment. We discuss hypothesis testing using the Bayes factor, and model selection using cross-validation. We close with some examples illustrating the flexibility of model specification in the Bayesian framework. Suggestions for further reading are also provided.

stat.AP↗