Refine
Year of publication
Document Type
- Article (78)
- Postprint (16)
- Review (3)
- Monograph/Edited Volume (2)
- Part of Periodical (2)
- Other (1)
Keywords
- German (10)
- eye-tracking (9)
- interference (8)
- locality (8)
- Eye movements (7)
- Bayesian data analysis (6)
- Reading (6)
- individual differences (6)
- self-paced reading (6)
- sentence processing (6)
Institute
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, Levy, Scheepers & Tily, 2013), presumably to keep Type I error down to the nominal a 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. (C) 2017 The Authors. Published by Elsevier Inc.
We present the fundamental ideas underlying statistical hypothesis testing using the frequentist framework. We start with a simple example that builds up the one-sample t-test from the beginning, explaining important concepts such as the sampling distribution of the sample mean, and the iid assumption. Then, we examine the meaning of the p-value in detail and discuss several important misconceptions about what a p-value does and does not tell us. This leads to a discussion of Type I, II error and power, and Type S and M error. An important conclusion from this discussion is that one should aim to carry out appropriately powered studies. Next, we discuss two common issues that we have encountered in psycholinguistics and linguistics: running experiments until significance is reached and the ‘garden-of-forking-paths’ problem discussed by Gelman and others. The best way to use frequentist methods is to run appropriately powered studies, check model assumptions, clearly separate exploratory data analysis from planned comparisons decided upon before the study was run, and always attempt to replicate results.
Argument-head distance and processing complexity: Explaining both locality and antilocality effects
(2006)
Although proximity between arguments and verbs (locality) is a relatively robust determinant of sentence-processing difficulty (Hawkins 1998, 2001, Gibson 2000), increasing argument-verb distance can also facilitate processing (Konieczny 2000). We present two self-paced reading (SPR) experiments involving Hindi that provide further evidence of antilocality, and a third SPR experiment which suggests that similarity-based interference can attenuate this distance-based facilitation. A unified explanation of interference, locality, and antilocality effects is proposed via an independently motivated theory of activation decay and retrieval interference (Anderson et al. 2004).*
With the arrival of the R packages nlme and lme4, linear mixed models (LMMs) have come to be widely used in experimentally-driven areas like psychology, linguistics, and cognitive science. This tutorial provides a practical introduction to fitting LMMs in a Bayesian framework using the probabilistic programming language Stan. We choose Stan (rather than WinBUGS or JAGS) because it provides an elegant and scalable framework for fitting models in most of the standard applications of LMMs. We ease the reader into fitting increasingly complex LMMs, using a two-condition repeated measures self-paced reading study.
Sentence comprehension requires that the comprehender work out who did what to whom. This process has been characterized as retrieval from memory. This review summarizes the quantitative predictions and empirical coverage of the two existing computational models of retrieval and shows how the predictive performance of these two competing models can be tested against a benchmark data-set. We also show how computational modeling can help us better understand sources of variability in both unimpaired and impaired sentence comprehension.
This tutorial analyzes voice onset time (VOT) data from Dongbei (Northeastern) Mandarin Chinese and North American English to demonstrate how Bayesian linear mixed models can be fit using the programming language Stan via the R package brms. Through this case study, we demonstrate some of the advantages of the Bayesian framework: researchers can (i) flexibly define the underlying process that they believe to have generated the data; (ii) obtain direct information regarding the uncertainty about the parameter that relates the data to the theoretical question being studied; and (iii) incorporate prior knowledge into the analysis. Getting started with Bayesian modeling can be challenging, especially when one is trying to model one’s own (often unique) data. It is difficult to see how one can apply general principles described in textbooks to one’s own specific research problem. We address this barrier to using Bayesian methods by providing three detailed examples, with source code to allow easy reproducibility. The examples presented are intended to give the reader a flavor of the process of model-fitting; suggestions for further study are also provided. All data and code are available from: https://osf.io/g4zpv.
Within quantitative phonetics, it is common practice to draw conclusions based on statistical significance alone Using incomplete neutralization of final devoicing in German as a case study, we illustrate the problems with this approach. If researchers find a significant acoustic difference between voiceless and devoiced obstruents, they conclude that neutralization is incomplete, and if they find no significant difference, they conclude that neutralization is complete. However, such strong claims regarding the existence or absence of an effect based on significant results alone can be misleading. Instead, the totality of available evidence should be brought to bear on the question. Towards this end, we synthesize the evidence from 14 studies on incomplete neutralization in German using a Bayesian random-effects meta-analysis. Our meta-analysis provides evidence in favor of incomplete neutralization. We conclude with some suggestions for improving the quality of future research on phonetic phenomena: ensure that sample sizes allow for high-precision estimates of the effect; avoid the temptation to deploy researcher degrees of freedom when analyzing data; focus on estimates of the parameter of interest and the uncertainty about that parameter; attempt to replicate effects found; and, whenever possible, make both the data and analysis available publicly. (c) 2018 Elsevier Ltd. All rights reserved.
It is well-known in statistics (e.g., Gelman & Carlin, 2014) that treating a result as publishable just because the p-value is less than 0.05 leads to overoptimistic expectations of replicability. These effects get published, leading to an overconfident belief in replicability. We demonstrate the adverse consequences of this statistical significance filter by conducting seven direct replication attempts (268 participants in total) of a recent paper (Levy & Keller, 2013). We show that the published claims are so noisy that even non-significant results are fully compatible with them. We also demonstrate the contrast between such small-sample studies and a larger-sample study; the latter generally yields a less noisy estimate but also a smaller effect magnitude, which looks less compelling but is more realistic. We reiterate several suggestions from the methodology literature for improving current practices.
Research on similarity-based interference has provided extensive evidence that the formation of dependencies between non-adjacent words relies on a cue-based retrieval mechanism. There are two different models that 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 the ACT-R framework, 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 can be affected, while the retrieval time is drawn from the same distribution; in this model, differences in latencies are a by-product of the possibility of backtracking and repairing incorrect retrievals. We implemented both models in a Bayesian hierarchical framework in order to evaluate them and compare them. The data show that correct retrievals take longer than incorrect ones, and this pattern is better fit under the direct-access model than under the activation-based model. 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. 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. This first ever computational evaluation of alternative accounts of retrieval processes in sentence processing opens the way for a broader investigation of theories of dependency completion.
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 execution time of parse steps; intermittent deficiency as increased random noise in activation of elements in memory; and resource reduction as reduced spreading activation. As data, we considered subject vs. object relative sentences, 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 slow parsing times, high noise, and low spreading activation. We find that (a) individual IWA show differential amounts of deficit along the three dimensions of slowed processing, intermittent deficiency, and resource reduction, (b) overall, there is evidence for all three sources of deficit playing a role, and (c) IWA have a more variable range of parameter values than controls. An important implication is that it may be meaningless to talk about sources of deficit with respect to an abstract verage IWA; the focus should be on the individual's differential degrees of deficit along different dimensions, and on understanding the causes of variability in deficit between participants.
Given the replication crisis in cognitive science, it is important to consider what researchers need to do in order to report results that are reliable. We consider three changes in current practice that have the potential to deliver more realistic and robust claims. First, the planned experiment should be divided into two stages, an exploratory stage and a confirmatory stage. This clear separation allows the researcher to check whether any results found in the exploratory stage are robust. The second change is to carry out adequately powered studies. We show that this is imperative if we want to obtain realistic estimates of effects in psycholinguistics. The third change is to use Bayesian data-analytic methods rather than frequentist ones; the Bayesian framework allows us to focus on the best estimates we can obtain of the effect, rather than rejecting a strawman null. As a case study, we investigate number interference effects in German. Number feature interference is predicted by cue-based retrieval models of sentence processing (Van Dyke & Lewis, 2003; Vasishth & Lewis, 2006), but it has shown inconsistent results. We show that by implementing the three changes mentioned, suggestive evidence emerges that is consistent with the predicted number interference effects.
In this paper we examine the effect of uncertainty on readers’ predictions about meaning. In particular, we were interested in how uncertainty might influence the likelihood of committing to a specific sentence meaning. We conducted two event-related potential (ERP) experiments using particle verbs such as turn down and manipulated uncertainty by constraining the context such that readers could be either highly certain about the identity of a distant verb particle, such as turn the bed […] down, or less certain due to competing particles, such as turn the music […] up/down. The study was conducted in German, where verb particles appear clause-finally and may be separated from the verb by a large amount of material. We hypothesised that this separation would encourage readers to predict the particle, and that high certainty would make prediction of a specific particle more likely than lower certainty. If a specific particle was predicted, this would reflect a strong commitment to sentence meaning that should incur a higher processing cost if the prediction is wrong. If a specific particle was less likely to be predicted, commitment should be weaker and the processing cost of a wrong prediction lower. If true, this could suggest that uncertainty discourages predictions via an unacceptable cost-benefit ratio. However, given the clear predictions made by the literature, it was surprisingly unclear whether the uncertainty manipulation affected the two ERP components studied, the N400 and the PNP. Bayes factor analyses showed that evidence for our a priori hypothesised effect sizes was inconclusive, although there was decisive evidence against a priori hypothesised effect sizes larger than 1μV for the N400 and larger than 3μV for the PNP. We attribute the inconclusive finding to the properties of verb-particle dependencies that differ from the verb-noun dependencies in which the N400 and PNP are often studied.
We used Chinese prenominal relative clauses (RCs) to test the predictions of two competing accounts of sentence comprehension difficulty: the experience-based account of Levy () and the Dependency Locality Theory (DLT; Gibson, ). Given that in Chinese RCs, a classifier and/or a passive marker BEI can be added to the sentence-initial position, we manipulated the presence/absence of classifiers and the presence/absence of BEI, such that BEI sentences were passivized subject-extracted RCs, and no-BEI sentences were standard object-extracted RCs. We conducted two self-paced reading experiments, using the same critical stimuli but somewhat different filler items. Reading time patterns from both experiments showed facilitative effects of BEI within and beyond RC regions, and delayed facilitative effects of classifiers, suggesting that cues that occur before a clear signal of an upcoming RC can help Chinese comprehenders to anticipate RC structures. The data patterns are not predicted by the DLT, but they are consistent with the predictions of experience-based theories.
In 2019 the Journal of Memory and Language instituted an open data and code policy; this policy requires that, as a rule, code and data be released at the latest upon publication. How effective is this policy? We compared 59 papers published before, and 59 papers published after, the policy took effect. After the policy was in place, the rate of data sharing increased by more than 50%. We further looked at whether papers published under the open data policy were reproducible, in the sense that the published results should be possible to regenerate given the data, and given the code, when code was provided. For 8 out of the 59 papers, data sets were inaccessible. The reproducibility rate ranged from 34% to 56%, depending on the reproducibility criteria. The strongest predictor of whether an attempt to reproduce would be successful is the presence of the analysis code: it increases the probability of reproducing reported results by almost 40%. We propose two simple steps that can increase the reproducibility of published papers: share the analysis code, and attempt to reproduce one's own analysis using only the shared materials.
Intuitively, strongly constraining contexts should lead to stronger probabilistic representations of sentences in memory. Encountering unexpected words could therefore be expected to trigger costlier shifts in these representations than expected words. However, psycholinguistic measures commonly used to study probabilistic processing, such as the N400 event-related potential (ERP) component, are sensitive to word predictability but not to contextual constraint. Some research suggests that constraint-related processing cost may be measurable via an ERP positivity following the N400, known as the anterior post-N400 positivity (PNP). The PNP is argued to reflect update of a sentence representation and to be distinct from the posterior P600, which reflects conflict detection and reanalysis. However, constraint-related PNP findings are inconsistent. We sought to conceptually replicate Federmeier et al. (2007) and Kuperberg et al. (2020), who observed that the PNP, but not the N400 or the P600, was affected by constraint at unexpected but plausible words. Using a pre-registered design and statistical approach maximising power, we demonstrated a dissociated effect of predictability and constraint: strong evidence for predictability but not constraint in the N400 window, and strong evidence for constraint but not predictability in the later window. However, the constraint effect was consistent with a P600 and not a PNP, suggesting increased conflict between a strong representation and unexpected input rather than greater update of the representation. We conclude that either a simple strong/weak constraint design is not always sufficient to elicit the PNP, or that previous PNP constraint findings could be an artifact of smaller sample size.
What is the processing cost of being garden-pathed by a temporary syntactic ambiguity? We argue that comparing average reading times in garden-path versus non-garden-path sentences is not enough to answer this question. Trial-level contaminants such as inattention, the fact that garden pathing may occur non-deterministically in the ambiguous condition, and "triage" (rejecting the sentence without reanalysis; Fodor & Inoue, 2000) lead to systematic underestimates of the true cost of garden pathing. Furthermore, the "pure" garden-path effect due to encountering an unexpected word needs to be separated from the additional cost of syntactic reanalysis. To get more realistic estimates for the individual processing costs of garden pathing and syntactic reanalysis, we implement a novel computational model that includes trial-level contaminants as probabilistically occurring latent cognitive processes. The model shows a good predictive fit to existing reading time and judgment data. Furthermore, the latent-process approach captures differences between noun phrase/zero complement (NP/Z) garden-path sentences and semantically biased reduced relative clause (RRC) garden-path sentences: The NP/Z garden path occurs nearly deterministically but can be mostly eliminated by adding a comma. By contrast, the RRC garden path occurs with a lower probability, but disambiguation via semantic plausibility is not always effective.
Cue-based retrieval theories in sentence processing predict two classes of interference effect: (i) Inhibitory interference is predicted when multiple items match a retrieval cue: cue-overloading leads to an overall slowdown in reading time; and (ii) Facilitatory interference arises when a retrieval target as well as a distractor only partially match the retrieval cues; this partial matching leads to an overall speedup in retrieval time. Inhibitory interference effects are widely observed, but facilitatory interference apparently has an exception: reflexives have been claimed to show no facilitatory interference effects. Because the claim is based on underpowered studies, we conducted a large-sample experiment that investigated both facilitatory and inhibitory interference. In contrast to previous studies, we find facilitatory interference effects in reflexives. We also present a quantitative evaluation of the cue-based retrieval model of Engelmann, Jager, and Vasishth (2019).