Testing for independence between pairs of autocorrelated binomial data sequences

A problem arises in analyzing the existence of interdependence between the behavioral sequences of two individuals: tests involving a statistic such as chi-square assume independent observations within each behavioral sequence, a condition which may not exist in actual practice. Using Monte Carlo simulations of binomial data sequences, we found that the use of a chi-square test frequently results in unacceptable Type I error rates when the data sequences are autocorrelated. We compared these results to those from two other methods designed specifically for testing for intersequence independence in the presence of intrasequence autocorrelation. The first method directly tests the intersequence correlation using an approximation of the variance of the intersequence correlation estimated from the sample autocorrelations. The second method uses tables of critical values of the intersequence correlation computed by Nakamuraet al. (J. Am. Stat. Assoc., 1976,71, 214–222). Although these methods were originally designed for normally distributed data, we found that both methods produced much better results than the uncorrected chi-square test when applied to binomial autocorrelated sequences. The superior method appears to be the variance approximation method, which resulted in Type I error rates that were generally less than or equal to 5% when the level of significance was set at .05.     

An extended application ‘Brain Q’ processing EEG and MEG data of finger stimulation extended from ‘Zeffiro’ based on machine learning and signal processing

GoalTo apply signal processing and machine learning skills and knowledge in processing the EEG and MEG signal and further localize and evaluate the source of the finger stimulation.MethodsCognitive control is usually applied in information processing and behavioral response. In the preprocessing, baseline correction is implemented to analyze the pre-stimuli, combining ERP to mark the event related potential, studying the time-locked only behavior. Z-score transform, coherence and spec trum are calculated and analyzed in the functional connectivity analysis.In addition to the functional analysis, Bayes Optimizer evaluates the neuro imaging according to the hierarchical Bayes. The introduction of the application is described from both user and developer’s prospects. Results: Introduction of both user and developers aspects, on its modules from pre-processing, functional analysis and results visualization and evaluation is conducted with one specific clinical data case, including the correlation is higher especially on gamma band and the MVAR coherence on the whole source space depicting the relation between different regions, especially on somatosensory (compared by thalamus) when stimulated by finger activity, phase-lock property of the E/MEG signal and etc. Compared to a manual selection, the scaling parameter prediction can be improved with support vector machine (SVM). The evaluation results with Bayes Optimization, location prediction is superior in the somatosensory area and in the thalamus, the total reconstructed source space is larger, one of the realization of cognitive system comparing different kernels and classifiers. The SVM and discriminant classifier gives similar results evaluating the dipole localization and the parameter choice related as well to the shape parameter, noise level, hyperprior and etc.ConclusionApproaches of Brain Q are found to be suitable for pre-processing for the EEG and MEG data. The system is capable of functional analysis including coherence and spectral related computation. Machine learning techniques are conducted as well to analyze and evaluate the result of the dipole reconstruction and help to predict the better model parameters and the localization of the origin dipoles. A case on finger stimulation clinical data is conducted and the results of the analysis temporarily and spatially manifests its functionality for users and potential extensions for developers.     

A latent topic model with Markov transition for process data

We propose a latent topic model with a Markov transition for process data, which consists of time-stamped events recorded in a log file. Such data are becoming more widely available in computer-based educational assessment with complex problem-solving items. The proposed model can be viewed as an extension of the hierarchical Bayesian topic model with a hidden Markov structure to accommodate the underlying evolution of an examinee's latent state. Using topic transition probabilities along with response times enables us to capture examinees' learning trajectories, making clustering/classification more efficient. A forward-backward variational expectation-maximization (FB-VEM) algorithm is developed to tackle the challenging computational problem. Useful theoretical properties are established under certain asymptotic regimes. The proposed method is applied to a complex problem-solving item in the 2012 version of the Programme for International Student Assessment (PISA).     

Investigating cycling kinematics and braking maneuvers in the real world: e-bikes make cyclists move faster,brake harder,and experience new conflicts

Pedelecs (e-bikes), which facilitate higher speeds with less effort in comparison to traditional bicycles (t-bikes), have grown considerably in popularity in recent years. Despite the large expansion of this new transportation mode, little is known about the behavior of e-cyclists, or whether cycling an e-bike increases crash risk and the likelihood of conflicts with other road users, compared to cycling on t-bikes. In order to support the design of safety measures and to maximize the benefits of e-bike use, it is critical to investigate the real-world behavior of riders as a result of switching from t-bikes to e-bikes.Naturalistic studies provide an unequaled method for investigating rider cycling behavior and bicycle kinematics in the real world in which the cyclist regularly experiences traffic conflicts and may need to perform avoidance maneuvers, such as hard braking, to avoid crashing. In this paper we investigate cycling kinematics and braking events from naturalistic data to determine the extent to which cyclist behavior changes as a result of transferring from t-bikes to e-bikes, and whether such change influences cycling safety.Data from the BikeSAFE and E-bikeSAFE naturalistic studies were used in this investigation to evaluate possible changes in the behavior of six cyclists riding t-bikes in the first study and e-bikes in the second one. Individual cyclists’ kinematics were compared between bicycle types. In addition, a total of 5092 braking events were automatically extracted after identification of dynamic triggers. The 286 harshest braking events (136 cases for t-bike and 150 for e-bike) were then validated and coded via video inspection.Results revealed that each of the cyclists rode faster on the e-bike than on the t-bike, increasing his/her average speed by 2.9–5.0 km/h. Riding an e-bike also increased the probability to unexpectedly have to brake hard (odds ratio = 1.72). In addition, the risk of confronting abrupt braking and sharp deceleration were higher when riding an e-bike than when riding a t-bike.Our findings provide evidence that cyclists’ behavior and the way cyclists interact with other road users change when cyclists switch from t-bikes to e-bikes. Because of the higher velocity, when on e-bikes, cyclists appear to have harder time predicting movements within the traffic environment and, as a result, they need to brake abruptly more often to avoid collisions, compared with cycling on t-bikes. This study provides new insights into the potential impact on safety that a cycling society moving to e-bikes may have, indicating that e-cycling requires more reactive maneuvers than does cycling traditional bicycles and suggesting that any distractive activity may be more critical when riding e-bikes compared to traditional bikes.     

Testing autocorrelation and partial autocorrelation: Asymptotic methods versus resampling techniques

Autocorrelation and partial autocorrelation, which provide a mathematical tool to understand repeating patterns in time series data, are often used to facilitate the identification of model orders of time series models (e.g., moving average and autoregressive models). Asymptotic methods for testing autocorrelation and partial autocorrelation such as the 1/T approximation method and the Bartlett's formula method may fail in finite samples and are vulnerable to non-normality. Resampling techniques such as the moving block bootstrap and the surrogate data method are competitive alternatives. In this study, we use a Monte Carlo simulation study and a real data example to compare asymptotic methods with the aforementioned resampling techniques. For each resampling technique, we consider both the percentile method and the bias-corrected and accelerated method for interval construction. Simulation results show that the surrogate data method with percentile intervals yields better performance than the other methods. An R package pautocorr is used to carry out tests evaluated in this study.     

Équivalence de la version en français pour le Canada et de la version originale états-unienne de l’Adult Self-Report

The aim of the present study was to evaluate the equivalence of the French-Canadian version of the Adult Self-Report (ASR; Achenbach & Rescorla, 2003) to its original English (USA) version. Both the original English version and the newly translated French-Canadian version were administered to 251 bilingual students from four Canadian provinces (Manitoba, New-Brunswick, Ontario, and Quebec), with a one to two weeks interval. Correlations varying from 0.72 to 0.87 indicated a high degree of correspondence between the scales of the two versions. Cohen's d calculated using Morris and DeShon's (2002) formula for within-subjects designs indicate that differences between mean scores varied from near-zero (d = 0.01) to small (d = 0.19), and suggest that they are not clinically significant. Ordinal alpha coefficients for the scales of the French-Canadian version varied from 0.66 to 0.96 but are inferior to those of the English version in seven of the eleven scales. Globally, results indicate that the equivalence of the French-Canadian version to the original English version of the ASR is sufficient to recommend its use.     

Panning for the gold in health research: Incorporating studies’ methodological quality in meta-analysis

Background: Systematic reviews now routinely assess methodological quality to gauge the validity of the included studies and of the synthesis as a whole. Although trends from higher quality studies should be clearer, it is uncertain how often meta-analyses incorporate methodological quality in models of study results either as predictors, or, more interestingly, in interactions with theoretical moderators.

Objective and methods: We survey 200 meta-analyses in three health promotion domains to examine when and how meta-analyses incorporate methodological quality.

Results: Although methodological quality assessments commonly appear in contemporary meta-analyses (usually as scales), they are rarely incorporated in analyses, and still more rarely analysed in interaction with theoretical determinants of the success of health promotions. The few meta-analyses (2.5%) that did include such an interaction analysis showed that moderator results remained significant in higher quality studies or were present only among higher quality studies. We describe how to model quality interactively with theoretically derived moderators and discuss strengths and weaknesses of this approach and in relation to current meta-analytic practice.

Conclusions: In large literatures exhibiting heterogeneous effects, meta-analyses can incorporate methodological quality and generate conclusions that enable greater confidence not only about the substantive phenomenon but also about the role that methodological quality itself plays.     

Do Bettors Correctly Perceive Odds? Three Studies of How Bettors Interpret Betting Odds as Probabilistic Information

This paper reports on three studies investigating how accurately bettors (=people who regularly bet on sports events) interpret the probabilistic information implied by betting odds. All studies were based on data collected by web surveys prompting a total of 186 experienced bettors to convert sets of representative odds into frequency judgments. Bayesian statistical methods were used to analyze the data. From the results, the following conclusions were made: (i) On the whole, the bettors produced well‐calibrated judgments, indicating that they have realistic perceptions of odds. (ii) Bettors were unable to consciously adjust judgments for different margins. (iii) Although their interval judgments often covered the estimates implied by the odds, the bettors tended to overestimate the variation of expected profitable bets between months. The results are consistent with prior research showing that people tend to make accurate probability judgments when faced with tasks characterized by constant and clear feedback. Copyright © 2015 John Wiley & Sons, Ltd.