User Search Performance Research Articles

Users Meet Clarifying Questions: Toward a Better Understanding of User Interactions for Search Clarification

The use of clarifying questions (CQs) is a fairly new and useful technique to aid systems in recognizing the intent, context, and preferences behind user queries. Yet, understanding the extent of the effect of CQs on user behavior and the ability to identify relevant information remains relatively unexplored. In this work, we conduct a large user study to understand the interaction of users with CQs in various quality categories, and the effect of CQ quality on user search performance in terms of finding relevant information, search behavior, and user satisfaction. Analysis of implicit interaction data and explicit user feedback demonstrates that high-quality CQs improve user performance and satisfaction. By contrast, low- and mid-quality CQs are harmful, and thus allowing the users to complete their tasks without CQ support may be preferred in this case. We also observe that user engagement, and therefore the need for CQ support, is affected by several factors, such as search result quality or perceived task difficulty. The findings of this study can help researchers and system designers realize why, when, and how users interact with CQs, leading to a better understanding and design of search clarification systems.

Effects of design features of train electronic guidance interface on visual search, behavior, and usability

In urban rail trains, traditional printed guide information interfaces are being replaced by digital interfaces. An electronic guide screen is an important device for passengers to obtain information about the ride because the train is running underground. In particular, when the train is crowded, passengers must obtain information in advance and make appropriate decisions. Therefore, it is important to efficiently search for guidance interfaces. In this study, use scenarios of the guidance interface were simulated. The typical information layout and display mode in the current electronic guidance interface were the experimental variables. Through eye tracking and subjective evaluation, the effects of design features on user search performance, visual behavior, and usability satisfaction were discussed. The experimental results showed that the information layout and display mode of the guidance interface had an interactive effect on the task completion time, fixation counts, and usability score; they also had a significant effect on the recall accuracy and saccade ratio. Task completion time had a strong positive correlation with usability satisfaction and was correlated with visual behavior. The study confirmed that the split-screen display mode combined with interlaced layout could improve search performance and subjective satisfaction. Relevance to industryPrinted guide signs in subway cars are being replaced by electronic guide screens of different styles. Passengers must search for the required information within a limited period. In this study, the effects of different interface layouts and display modes on search performance, visual behavior, and usability were studied. This had a positive significance in improving user experience.

Effects of App Icon Border Form and Interface Background Color Saturation on User Visual Experience and Search Performance

The purpose of this study was to investigate the icon border form and interface background color saturation’s individual and connected impacts on interface interaction. They were two important factors affecting user visual experience and search performance, but previously they have been studied individually. An experiment using eye-tracking, behavior measurement, and subjective evaluation was carried out to determine the influence of app icon border form and interface background color saturation on user visual experience and search performance. The statistical results showed that during the human-computer interaction process, compared with the circle and rounded square border forms, the square icons had the longest search completion time and first fixation time and the lowest subjective evaluation score. In addition, rounded square icons had the best user visual experience. Compared with high-saturation background color, less time was required to complete search tasks in the low-saturation background color condition. Higher perceptual usability and aesthetic scores were also obtained under this condition compared to the high-saturation background color condition. Hence, this study confirmed that the two factors of icon border form and interface background color saturation had the basic properties of the interaction effect. The combination of rounded square icons and low-saturation background color can be beneficial to improve user search performance and visual experience. The findings can help in the design of human-computer interaction interfaces.

The effects of dynamic prompt and background transparency of hover feedback design on the user interface of shopping websites

PurposeThe implementation of interaction design concepts to facilitate communication between users and shopping websites has gained increasing attention in recent years. Mouse hover is a vital interaction method for users to access shopping sites and significantly impacts their decision-making experience. A well-designed mouse hover function can effectively enhance the user's search performance and improve the user experience. The purpose of this study is to investigate whether the dynamic prompt designs at the hover position and the degree of feedback transparency may affect the user's task performance and personal feelings when operating the hover function on the shopping website.Design/methodology/approachThe study employed two independent variables in the experiments: dynamic prompt and background transparency of hover feedback. A between-subject design of 2 (single flicker and continuous flicker) × 3 (transparency at 0%, 25%, and 50%) was adopted in the experiment. A total of 60 participants were invited to participate in the experiment using the purposive sampling method. Participants were asked to complete four operation tasks, and the time of each was recorded. They were then asked to complete the system usability scale (SUS) questionnaire and conduct subjective evaluations before they were briefly interviewed.FindingsThe generated results revealed that: (1) In the interaction with a shopping website, the degree of background transparency of hover feedback affected the participants' task performance. (2) The effect of a single flicker had significantly higher subjective evaluation results regarding the degree of rationality compared with a continuous flicker. (3) The participants' perceptions of the information clarity were affected when the background transparency of hover feedback was 50%. (4) The effect of a continuous flicker was better than that of a single flicker when the background transparency of hover feedback was 25%. The participants' attraction to a continuous flicker was significantly higher than that of a single flicker. Nonetheless, when the degree of background transparency was at 0% or 50%, the results were the opposite.Originality/valueThe findings generated from the research can be a reference for the development of hover operation in the user interface design for shopping websites.

Implicit search feature based approach to assist users in exploratory search tasks

Analyzing and modeling users’ online search behaviors when conducting exploratory search tasks could be instrumental in discovering search behavior patterns that can then be leveraged to assist users in reaching their search task goals. We propose a framework for evaluating exploratory search based on implicit features and user search action sequences extracted from the transactional log data to model different aspects of exploratory search namely uncertainty, creativity, exploration, and knowledge discovery. We show the effectiveness of the proposed framework by demonstrating how it can be used to understand and evaluate user search performance and thereby make meaningful recommendations to improve the overall search performance of users. We used data collected from a user study consisting of 18 users conducting an exploratory search task for two sessions with two different topics in the experimental analysis. With this analysis we show that we can effectively model their behavior using implicit features to predict the user’s future performance level with above 70% accuracy in most cases. Further, using simulations we demonstrate that our search process based recommendations improve the search performance of low performing users over time and validate these findings using both qualitative and quantitative approaches.

Toward A Human-Centered Hyperlipidemia Management System: The Interaction between Internal and External Information on Relational Data Search

In a distributed information search task, data representation and cognitive distribution jointly affect user search performance in terms of response time and accuracy. Guided by UFuRT (User, Function, Representation, Task), a human-centered framework, we proposed a search model and task taxonomy. The model defines its application in the context of healthcare setting. The taxonomy clarifies the legitimate operations for each type of search task of relational data. We then developed experimental prototypes of hyperlipidemia data displays. Based on the displays, we tested the search tasks performance through two experiments. The experiments are of a within-subject design with a random sample of 24 participants. The results support our hypotheses and validate the prediction of the model and task taxonomy. In this study, representation dimensions, data scales, and search task types are the main factors in determining search efficiency and effectiveness. Specifically, the more external representations provided on the interface the better search task performance of users. The results also suggest the ideal search performance occurs when the question type and its corresponding data scale representation match. The implications of the study lie in contributing to the effective design of search interface for relational data, especially laboratory results, which could be more effectively designed in electronic medical records.

Searching the Web: operator assistance required

This study examines the effects of the search interface and Boolean logic training on user search performance and satisfaction. We compared the use of an assisted search tool to that of a simple search interface like those typically found at commercial search engine sites. We found that the assisted search tool had a significant positive effect on performance, satisfaction, and confidence. Promoting the use of advanced search features is therefore in the best interest of both Web search providers and users. In the absence of an assisted interface, a simple interface coupled with Boolean logic training was also an effective means for improving user performance. Given that most searchers choose to use the simple search box, it is important to provide them with applicable training to promote the effective use of the search tool. Interestingly, coupling the assisted interface with Boolean training was no more effective than either treatment alone.

Collaborative relevance judgment: A group consensus method for evaluating user search performance

AbstractRelevance judgment has traditionally been considered a personal and subjective matter. A user's search and the search result are treated as an isolated event. To consider the collaborative nature of information retrieval (IR) in a group/organization or even societal context, this article proposes a method that measures relevance based on group/peer consensus. The method can be used in IR experiments. In this method, the relevance of a document is decided by group consensus, or more specifically, by the number of users (or experiment participants) who retrieve it for the same search question. The more users who retrieve it, the more relevant the document will be considered. A user's search performance can be measured by a relevance score based on this notion. The article reports the results of an experiment using this method to compare the search performance of different types of users. Related issues with the method and future directions are also discussed.

Effects of link annotations on search performance in layered and unlayered hierarchically organized information spaces

AbstractThe effects of link annotations on user search performance in hypertext environments having deep (layered) and shallow link structures were investigated in this study. Four environments were tested—layered‐annotated, layered‐unannotated, shallow‐annotated, and shallow‐unannotated. A single document was divided into 48 sections, and layered and unlayered versions were created. Additional versions were created by adding annotations to the links in the layered and unlayered versions. Subjects were given three queries of varying difficulty and then asked to find the answers to the queries that were contained within the hypertext environment to which they were randomly assigned. Correspondence between the wording links and queries was used to define difficulty level. The results of the study confirmed previous research that shallow link structures are better than deep (layered) link structures. Annotations had virtually no effect on the search performance of the subjects. The subjects performed similarly in the annotated and unannotated environments, regardless of whether the link structures were shallow or deep. An analysis of question difficulty suggests that the wording in links has primacy over the wording in annotations in influencing user search behavior.

Book availability as a performance measure of a library: an analysis of the effectiveness of a health sciences library.

User satisfaction as a measure of library effectiveness has been studied by using Kantor's branching technique. Cleveland Health Sciences Library, Cleveland, Ohio, was used as the center of this study. The study measured independently the user bibliographic information performance, collection development policy performance, acquisition policy performance, user catalog performance, circulation performance, library operation/functioning performance, and the user search performance. Each of these categories represented one or more of the factors of library performance that could account for the possibility of users' failure. The sample size consisted of 1000 requests for book titles. The book availability rates were 59.60% without the help of the librarian and 63.50% with the help of the librarian. The results indicate that a dynamic study of library policies at regular intervals by using the Kantor's method, which is easily reproducible, can maximize library resources for the most effective fulfillment of user demand.

Designing Speech Displays for Telephone Information Systems

This study used a computer-driven telephone information system as a real-time human-computer interface to simulate applications where synthetic speech is used to access data. Subjects used a telephone keypad to search though an automated department store database to locate and transcribe specific information messages. Because speech provides a sequential and transient information display, users may have difficulty navigating through auditory databases. One issue investigated in this study was whether the alternate use of male and female voices to code different levels of the database would improve user search performance. Other issues investigated were the basic intelligibility of these male and female voices as influenced by different levels of speech rate. All factors were assessed as functions of search or transcription task performance and user preference. Analysis of transcription accuracy, user search efficiency and time, and subjective ratings revealed an overall significant effect of speech rate on all groups of measures but no significant effects for voice type or coding scheme. Results were used to recommend design guidelines for developing speech displays for telephone information systems.