Mahtab Ghazizadeh

Extending the Technology Acceptance Model to assess automation

Often joint human–automation performance depends on the factors influencing the operator’s tendency to rely on and comply with automation. Although cognitive engineering (CE) researchers have studied automation acceptance as related to task–technology compatibility and human–technology coagency, information system (IS) researchers have evaluated user acceptance of technology, using the Technology Acceptance Model (TAM). The parallels between the two views suggest that the user acceptance perspective from the IS community can complement the human–automation interaction perspective from the CE community. TAM defines constructs that govern acceptance and provides a framework for evaluating a broad range of factors influencing technology acceptance and reliance. TAM is extensively used by IS researchers in various applications and it can be applied to assess the effect of trust and other factors on automation acceptance. Likewise, extensions to the TAM framework use the constructs of task–technology compatibility and past experience to extend its description of the role of human–automation interaction in automation adoption. We propose the Automation Acceptance Model (AAM) to draw upon the IS and CE perspectives and take into account the dynamic and multi-level nature of automation use, highlighting the influence of use on attitudes that complements the more common view that attitudes influence use.

Augmenting the Technology Acceptance Model with Trust: Commercial Drivers’ Attitudes towards Monitoring and Feedback

This study evaluates truck drivers’ attitudes toward an on-board monitoring system (OBMS), using an extended version of the Technology Acceptance Model (TAM) that accounts for drivers’ trust in OBMS. Crashes that involve trucks incur a high cost to society and driver-related factors contribute to about one third of all large truck fatal crashes in the US. Therefore, safety initiatives that can increase drivers’ awareness of their risky behaviors are highly desirable. In-vehicle feedback systems are designed to serve this purpose; however, their benefits will not be realized unless their information can positively influence safe driving. Acceptance constructs for the proposed model were measured using a survey administered after the monitoring system was introduced to the drivers but before the system was actually installed in their trucks. In line with the TAM, the results demonstrated that perceived usefulness is the most important determinant of intention to use the OBMS. Trust was also a major determinant of intention to use, suggesting that the acceptance model can be usefully augmented by this construct.

Text Mining to Decipher Free-Response Consumer Complaints: Insights From the NHTSA Vehicle Owner’s Complaint Database

Objective: This study applies text mining to extract clusters of vehicle problems and associated trends from free-response data in the National Highway Traffic Safety Administration’s vehicle owner’s complaint database. Background: As the automotive industry adopts new technologies, it is important to systematically assess the effect of these changes on traffic safety. Driving simulators, naturalistic driving data, and crash databases all contribute to a better understanding of how drivers respond to changing vehicle technology, but other approaches, such as automated analysis of incident reports, are needed. Method: Free-response data from incidents representing two severity levels (fatal incidents and incidents involving injury) were analyzed using a text mining approach: latent semantic analysis (LSA). LSA and hierarchical clustering identified clusters of complaints for each severity level, which were compared and analyzed across time. Results: Cluster analysis identified eight clusters of fatal incidents and six clusters of incidents involving injury. Comparisons showed that although the airbag clusters across the two severity levels have the same most frequent terms, the circumstances around the incidents differ. The time trends show clear increases in complaints surrounding the Ford/Firestone tire recall and the Toyota unintended acceleration recall. Increases in complaints may be partially driven by these recall announcements and the associated media attention. Conclusion: Text mining can reveal useful information from free-response databases that would otherwise be prohibitively time-consuming and difficult to summarize manually. Application: Text mining can extend human analysis capabilities for large free-response databases to support earlier detection of problems and more timely safety interventions.

Perspectives on the Training of Human Factors Students for the User Experience Industry

We surveyed 140 HFES student members and found that nearly 80% of students were considering a future career in UX. In contrast, only 12% felt that their training has prepared them extremely well for a UX career. An analysis of 40 UX job postings revealed that while these positions required some human factors related skills, 37% of their job requirements emphasized design familiarity and programming skills. Students indicated that gaps in their education and preparation represent the largest challenge they face in entering the UX field. They further identified the broad definition of UX and lack of access to industry positions as other challenges in transitioning to UX professionals. It is recommended to focus on increasing HFES’s relevance to students interested in future UX careers.

Commercial Drivers’ Initial Attitudes toward an On-Board Monitoring System

Several studies have shown the effectiveness of on-board monitoring for improving commercial driver safety but little has been done to examine truck drivers’ attitudes toward such systems. The purpose of the current study is to examine these drivers’ initial attitudes toward an on-board monitoring system (OBMS) as influenced by driver characteristics and driving experiences. Commercial drivers’ attitudes and demographics were collected via a questionnaire distributed after a brief introduction of an OBMS. The results of a cluster analysis revealed three subgroups of commercial drivers who had negative, moderately positive, and extremely positive attitudes toward OBMS. Those with extremely positive attitudes, named the Fanatics cluster indicated that they would highly trust the system and felt they would adjust their driving based on system feedback. The cluster with more negative attitudes (named Opponents) showed a higher proportion of non-married drivers than married drivers and reported that the system would be an invasion of privacy. This study provides some insights on commercial drivers’ attitudes toward feedback from technology and can help designers and researchers understand differences in drivers’ willingness to accept and use feedback systems.

Consumer Complaints and Traffic Fatalities: Insights from the NHTSA Vehicle Owner’s Complaint Database

Driving simulators, crash databases, and more recently, naturalistic studies all help understand how changes to vehicle design affect driving safety. The rapid computerization of cars makes it increasingly important to capitalize on these sources and exploit others. The present study explores a rarely analyzed data source on traffic fatalities: National Highway Traffic Safety Administration’s vehicle owner’s complaint database. The textual data within the event description field of each complaint is extracted and analyzed using a text mining approach that involves the use of latent semantic analysis (LSA) for reducing the dimensionality of the problem. Hierarchical clustering is then employed to identify clusters of complaints that share content. Clusters are described in terms of the most frequent terms and the time trends of the complaints within them. The analysis highlights how text mining analysis can help unlock the wealth of information contained in consumer complaint databases.

A memetic heuristic for the co-clustering problem

Co-clustering partitions two different kinds of objects simultaneously. Bregman co-clustering is a well-studied fast iterative algorithm to perform co-clustering. However, this method is very prone to local optima. We propose a memetic algorithm to solve the co-clustering problem. Experimental results show that this method outperforms the multi-start Bregman co-clustering in both accuracy and time.

Text Readability and Drivers' Reading Time: Insights from the Visual Occlusion Method

Internet access, vehicle diagnostics, and real-time communication are becoming increasingly common in all vehicles and the complexity of the information displayed is a major concern. Current guidelines focus on the number of characters in a text message, but other measures of text readability might be more sensitive. This study examined how well different metrics of text readability predict the time it takes a driver to read a message on an in-vehicle display. Participants completed reading tasks while seated in the driver’s seat of a driving simulator. Occlusion goggles were used to mimic the timesharing between the driving task and the secondary reading tasks, according to the ISO 16673 (2007) guidelines. The results showed that message length (number of characters) predicts the total time spent on the task (Total Shutter Open Time [TSOT]) and that the combination of number of words in the message and Shannon entropy of the message predicts TSOT only as well as number of characters alone. Applying the model human processor calculations of reading rates (Card, Moran, & Newell, 1983) showed that participants likely read the messages word-byword in successive saccades, instead of letter-by-letter or phrase-by-phrase. Findings provide direction for more in-depth lexical analyses of text readability related to in-vehicle displays.