Question Response Time Research Articles

Facilitating question identification through natural intensity eyebrow movements in virtual avatars

In conversation, recognizing social actions (similar to ‘speech acts’) early is important to quickly understand the speaker’s intended message and to provide a fast response. Fast turns are typical for fundamental social actions like questions, since a long gap can indicate a dispreferred response. In multimodal face-to-face interaction, visual signals may contribute to this fast dynamic. The face is an important source of visual signalling, and previous research found that prevalent facial signals such as eyebrow movements facilitate the rapid recognition of questions. We aimed to investigate whether early eyebrow movements with natural movement intensities facilitate question identification, and whether specific intensities are more helpful in detecting questions. Participants were instructed to view videos of avatars where the presence of eyebrow movements (eyebrow frown or raise vs. no eyebrow movement) was manipulated, and to indicate whether the utterance in the video was a question or statement. Results showed higher accuracies for questions with eyebrow frowns, and faster response times for questions with eyebrow frowns and eyebrow raises. No additional effect was observed for the specific movement intensity. This suggests that eyebrow movements that are representative of naturalistic multimodal behaviour facilitate question recognition.

Conversational Eyebrow Frowns Facilitate Question Identification: An Online Study Using Virtual Avatars.

Conversation is a time-pressured environment. Recognizing a social action (the ''speech act,'' such as a question requesting information) early is crucial in conversation to quickly understand the intended message and plan a timely response. Fast turns between interlocutors are especially relevant for responses to questions since a long gap may be meaningful by itself. Human language is multimodal, involving speech as well as visual signals from the body, including the face. But little is known about how conversational facial signals contribute to the communication of social actions. Some of the most prominent facial signals in conversation are eyebrow movements. Previous studies found links between eyebrow movements and questions, suggesting that these facial signals could contribute to the rapid recognition of questions. Therefore, we aimed to investigate whether early eyebrow movements (eyebrow frown or raise vs. no eyebrow movement) facilitate question identification. Participants were instructed to view videos of avatars where the presence of eyebrow movements accompanying questions was manipulated. Their task was to indicate whether the utterance was a question or a statement as accurately and quickly as possible. Data were collected using the online testing platform Gorilla. Results showed higher accuracies and faster response times for questions with eyebrow frowns, suggesting a facilitative role of eyebrow frowns for question identification. This means that facial signals can critically contribute to the communication of social actions in conversation by signaling social action-specific visual information and providing visual cues to speakers' intentions.

Predicting Web Survey Breakoffs Using Machine Learning Models

Web surveys are becoming increasingly popular but tend to have more breakoffs compared to the interviewer-administered surveys. Survey breakoffs occur when respondents quit the survey partway through. The Cox survival model is commonly used to understand patterns of breakoffs. Nevertheless, there is a trend to using more data-driven models when the purpose is prediction, such as classification machine learning models. It is unclear in the breakoff literature what are the best statistical models for predicting question-level breakoffs. Additionally, there is no consensus about the treatment of time-varying question-level predictors, such as question response time and question word count. While some researchers use the current values, others aggregate the value from the beginning of the survey. This study develops and compares both survival models and classification models along with different treatments of time-varying variables. Based on the level of agreement between the predicted and actual breakoff, we find that the Cox model and gradient boosting outperform other survival models and classification models respectively. We also find that using the values of time-varying predictors concurrent to the breakoff status is more predictive of breakoff, compared to aggregating their values from the beginning of the survey, implying that respondents’ breakoff behaviour is more driven by the current response burden.

Design and Evaluation of a Cyber Protection Team Planner Work Aid

Cyber Protection Teams (CPTs) are responsible for providing mission assurance by ensuring that mission-relevant cyber terrain is secure and protected for operations. In order to do this effectively, CPT planners must correlate and integrate information regarding customers’ missions with information about the systems and networks required to perform those missions. To support this challenge, a CPT Planner work aid was designed and evaluated to assess its effectiveness. The evaluation compared the work aid with current CPT planning methods. Ten participants with varying planning expertise performed four planning tasks (managing approvals, assigning resources, managing accesses, and scheduling meetings). To test the efficacy of the proposed work aid, participants were tasked with planning a fictional mission utilizing both the traditional planning method and the work aid method. Data was collected from objective metrics to include time and task performance score, as well as subjective perception of workload. In addition, participants were asked a series of situation awareness (SA) questions after each task to measure their understanding of the mission-relevant information. Score and time to answer SA questions were also recorded. While CPT planning experience did not have a significant effect on performance, results revealed significant improvements in task time, SA question response time, and perceived workload when using the work aid versus the baseline condition.

Impact of interface design on drivers’ behavior in partially automated cars: An on-road study

In partially automated vehicles, the driver and the automated system share control of the vehicle. Consequently, the driver may have to switch between driving and monitoring activities. This can critically impact the driver’s situational awareness. The human–machine interface (HMI) is responsible for efficient collaboration between driver and system. It must keep the driver informed about the status and capabilities of the automated system, so that he or she knows who or what is in charge of the driving. The present study was designed to compare the ability of two HMIs with different information displays to inform the driver about the system’s status and capabilities: a driving-centered HMI that displayed information in a multimodal way, with an exocentric representation of the road scene, and a vehicle-centered HMI that displayed information in a more traditional visual way. The impact of these HMIs on drivers was compared in an on-road study. Drivers’ eye movements and response times for questions asked while driving were measured. Their verbalizations during the test were also transcribed and coded. Results revealed shorter response times for questions on speed with the exocentric and multimodal HMI. The duration and number of fixations on the speedometer were also greater with the driving-centered HMI. The exocentric and multimodal HMI helped drivers understand the functioning of the system, but was more visually distracting than the traditional HMI. Both HMIs caused mode confusions. The use of a multimodal HMI can be beneficial and should be prioritized by designers. The use of auditory feedback to provide information about the level of automation needs to be explored in longitudinal studies.

Multi-feature based Question–Answerer Model Matching for predicting response time in CQA

Users of Community Question Answering (CQA) could not manage their time conveniently because their questions are often not answered quickly enough. To address this problem, we try to provide a function for CQA sites to inform users when their questions will be answered. In this paper, we propose a Question–Answerer Model Matching based answerer’s response time prediction named (QAM2), which consists of two parts: the construction of the Multi-feature based Question–Answerer Model (MQAM, including the answerer model and the question model) and the prediction of question response time based on MQAM Matching Strategy (QAMMS). Firstly, the MQAM is built according to some extracted deep features (e.g., answerer’s interest, professional level, activity, question category and difficulty), which are neglected in most existing methods on the prediction of question response time. Herein, the Label Cluster Latent Dirichlet Allocation (LC-LDA) model was proposed to overcome the compulsive allocation behaviors caused by traditional topic models (e.g. LDA), which treats the words that are irrelevant or weakly related to the subject as the topic of short texts when extracting the feature of answerer’s interest and question category. Meanwhile, an improved PageRank algorithm-topic sensitive weighted PageRank (TSWPR) is used to eliminate the impact of “indiscriminate” users who have answered many questions with low quality of answers. Secondly, we use the model matching strategy based on multiple classifier for matching MQAM and calculating the question response time of each answerer. Experiments conducted on two real data sets of Stack Overflow show that the proposed method can improve significantly the accuracy of question response time prediction in CQA.

The Optimization of Humanoid Robot’s Dialog in Improving Communication between Humanoid Robot and Older Adults

Older adults with dementia are increasing in Japan. Because of this, “communication robots” are being introduced into nursing settings to substitute for the shortage of nurses and other care workers. Our research group is currently developing a humanoid nursing robot with caring function (HNR), specifically for the functional and practical use of older adults with dementia. The purpose of this study was to clarify improvement points of the current function of humanoid robot’s (Pepper) (SoftBank Robotics) dialog pattern for improving optimal communication between humanoid robot and older adults. Dialog programs were installed in the humanoid robot Pepper, including the application program Care Prevention Gymnastics Exercises for Pepper (Pepper-CPGE) that was made by Xing Company, Japan. Dialogues between older adults and Pepper-CPGE were recorded by video camera. Data from transcriptions of the conversations captured from video and from field notes were analyzed focusing on human-robot interaction. From the recorded scenes and conversations, the following were points to be improved: 1) Intonation of the words vocalized by Pepper-CPGE was different from that expected by the older adult resulted in inappropriate responses by the older adult; 2) The timing between Pepper’s questions and the responses of the older adult were not timely and did not match (differences in the question-response time), which lead to confusion among the older adults; and 3) Other surroundings older adults were interested in the dialogue with Pepper-CPGE. However, Pepper-CPGE cannot communicate with multiple older adults. ImprovingPepper-CPGE’s ability to communicate with multiple older adults as an older adult’s dialogue with Pepper-CPGE can cause other older adults to also interact not only with the older adult that Pepper-CPGE is interacting with but also with Pepper-CPGE. This study shows that transactive relations among humanoid robots and older adults can be facilitated. Improving communication between humanoid robots and the older adults by optimizing a structured dialogue is needed to enhance appropriate engagement and participation. It is necessary to create a “Caring dialogue Database” for HNRs in order to know the patient/client and to share the aesthetic experiences of human-robot interactions. Also, it is important to develop a dialog pattern that enables humanoid robots to sympathize with older adults.

Effect of Access to an Electronic Medical Resource on Performance Characteristics of a Certification Examination: A Randomized Controlled Trial.

Electronic resources are increasingly used in medical practice. Their use during high-stakes certification examinations has been advocated by many experts, but whether doing so would affect the capacity to differentiate between high and low abilities is unknown. To determine the effect of electronic resources on examination performance characteristics. Randomized controlled trial. Medical certification program. 825 physicians initially certified by the American Board of Internal Medicine (ABIM) who passed the Internal Medicine Certification examination or sat for the Internal Medicine Maintenance of Certification (IM-MOC) examination in 2012 to 2015. Participants were randomly assigned to 1 of 4 conditions: closed book using typical or additional time, or open book (that is, UpToDate [Wolters Kluwer]) using typical or additional time. All participants took the same modified version of the IM-MOC examination. Primary outcomes included item difficulty (how easy or difficult the question was), item discrimination (how well the question differentiated between high and low abilities), and average question response time. Secondary outcomes included examination dimensionality (that is, the number of factors measured) and test-taking strategy. Item response theory was used to calculate question characteristics. Analysis of variance compared differences among conditions. Closed-book conditions took significantly less time than open-book conditions (mean, 79.2 seconds [95% CI, 78.5 to 79.9 seconds] vs. 110.3 seconds [CI, 109.2 to 111.4 seconds] per question). Mean discrimination was statistically significantly higher for open-book conditions (0.34 [CI, 0.32 to 0.35] vs. 0.39 [CI, 0.37 to 0.41] per question). A strong single dimension showed that the examination measured the same factor with or without the resource. Only 1 electronic resource was evaluated. Inclusion of an electronic resource with time constraints did not adversely affect test performance and did not change the specific skill or factor targeted by the examination. Further study on the effect of resource inclusion on other examinations is warranted. ABIM Foundation.

Effects of tag usage on question response time

Given a newly posted question on a Question and Answer (Q&A) site, how long will it take until an answer is received? Does response time relate to factors about how the question asker composes the question? If so, what are those factors? With advances in social media and the Web, Q&A sites have become a major source of information for Internet users. Response time of a question is an important aspect in these sites as it is associated with the users’ satisfaction and engagement, and thus the lifespan of these online communities. In this paper, we study and estimate response time for questions in StackOverflow, a popular online Q&A forum where software developers post and answer questions related to programming. We analyze a long list of factors in the data and identify those that have clear relation with response time. Our key finding is that tag-related factors, such as their “popularity” (how often the tag is used) and the num ber of their “subscribers” (how many users can answer questions containing the tag), provide much stronger evidence than factors not related to tags. Finally, we learn models using the identified evidential features for predicting the response time of questions, which also demonstrate the significance of tags chosen by the question asker.

The feasibility of using SMS as a health survey tool: An exploratory study in patients with rheumatoid arthritis

AimTo explore the feasibility of using SMS to conduct health surveys in patients with rheumatoid arthritis (RA). MethodsUsing a two-way SMS system, we conducted a baseline survey using a short and a long quality of life instrument, with two follow-up administrations of the short survey among patients with RA on follow-up at the Singapore General Hospital (SGH). Our outcome variables were survey response rate, question response time, proportion of replies following the format instructions, and survey administration cost. ResultsAmong 85 participants (of 99 eligible patients approached), the response rate was 67% for the long survey, 80% for the short survey at baseline, and 74% and 70% at the first and second follow-ups, respectively. Survey response rate varied with the employment status of respondents. Approximately 78% of all replies followed the format instructions. All these replies were received within the stipulated 24-h timeline and half of them were received within 29min. Response time for the other 22% of replies that did not follow the format instructions was much longer. The average administration cost per survey was S$0.34 (equivalent of US$0.26). ConclusionsTwo-way SMS is a promising tool to conduct short health surveys. Evidence for implementing long surveys over SMS is still weak. The post survey interviews with survey participants suggested that sequential delivery of survey questions coupled with SMS reminders in the interim, tailoring survey questions to individual health conditions, and providing feedback on individual survey results can potentially improve the response rate of such surveys.

Pace: An Alternative Measure of Student Question Response Time

In some professions, speed and accuracy are as important as acquired requisite knowledge and skills. The availability of computer-based testing now facilitates examination of these two important aspects of student performance. We found that student response times in a conventional non-speeded multiple-choice test, at both the global and individual question levels, closely approximated lognormal distributions. We propose a new measure, pace, which is derived from the survival function of these distributions for analysis of individual person response times. These pace estimates could be used both to rank and compare students; pace also performed maximally compared to other parameterizations in generalizability and dependability studies. While pace was very weakly related to person ability, there was no detectable relationship to question parameters of shift, natural logarithmic mean, or natural logarithmic standard deviation. That is, pace was a person-dependent, question-independent measure. Pace measurements were also successfully used as covariates in models for estimation of person response time to specified questions and person accuracy in response to specified questions. Thus, the analysis of pace can contribute significantly to comprehensive evaluation of student performance in both the speed and ability domains and is a requisite to best practice in testing and assessment.

Performance analysis of a distributed question/answering system

The problem of question/answering (Q/A) is to find answers to open-domain questions by searching large collections of documents. Unlike information retrieval systems very common today in the form of Internet search engines, Q/A systems do not retrieve documents, but instead provide short, relevant answers located in small fragments of text. This enhanced functionality comes with a price: Q/A systems are significantly slower and require more hardware resources than information retrieval systems. This paper proposes a distributed Q/A architecture that enhances the system throughput through the exploitation of interquestion parallelism and dynamic load balancing and reduces the individual question response time through the exploitation of intraquestion parallelism. Inter and intraquestion parallelism are both exploited using several scheduling points: one before the Q/A task is started and two embedded in the Q/A task. An analytical performance model is introduced. The model analyzes both the interquestion parallelism overhead generated by the migration of questions and the intraquestion parallelism overhead generated by the partitioning of the Q/A task. The analytical model indicates that both question migration and partitioning are required for a high-performance system.