Search results

This chapter describes how conversational computer agents have been used in collaborative problem-solving environments. These agent-based systems are designed to (a) assess the students’ knowledge, skills, actions, and various other psychological states on the basis of the students’ actions and the conversational interactions, (b) generate discourse moves that are sensitive to the psychological states and the problem states, and (c) advance a solution to the problem. We describe how this was accomplished in the Programme for International Student Assessment (PISA) for Collaborative Problem Solving (CPS) in 2015. In the PISA CPS 2015 assessment, a single human test taker (15-year-old student) interacts with one, two, or three agents that stage a series of assessment episodes. This chapter proposes that this PISA framework could be extended to accommodate more open-ended natural language interaction for those languages that have developed technologies for automated computational linguistics and discourse. Two examples support this suggestion, with associated relevant empirical support. First, there is AutoTutor, an agent that collaboratively helps the student answer difficult questions and solve problems. Second, there is CPS in the context of a multi-party simulation called Land Science in which the system tracks progress and knowledge states of small groups of 3–4 students. Human mentors or computer agents prompt them to perform actions and exchange open-ended chat in a collaborative learning and problem-solving environment.

This study aims to explore US hotel guests’ privacy concerns with a twofold aim as follows: to investigate the privacy categories, themes and attributes most commonly discussed by guests in their reviews and to examine the influence of cultural proximity on privacy concerns.

This study combined automated text analytics with content analysis. The database consisted of 68,000 hotel reviews written by US guests lodged in different types of hotels in five European cities. Linguistic Inquiry Word Count, Leximancer and SPSS software were used for data analysis. Automated text analytics and a validated privacy dictionary were used to investigate the reviews by exploring the categories, themes and attributes of privacy concerns. Content analysis was used to analyze the narratives and select representative snippets.

The findings revealed various categories, themes and concepts related to privacy concerns. The two most commonly discussed categories were privacy restriction and outcome state. The main themes discussed in association with privacy were “room,” “hotel,” “breakfast” and several concepts within each of these themes were identified. Furthermore, US guests showed the lowest levels of privacy concerns when staying at American hotel chains as opposed to non-American chains or independent hotels, highlighting the role of cultural proximity in privacy concerns.

Hotel managers can benefit from the results by improving their understanding of hotel and service attributes mostly associated with privacy concerns. Specific suggestions are provided to hoteliers on how to increase guests’ privacy and on how to manage issues related to cultural distance with guests.

This study contributes to the hospitality literature by investigating a neglected issue: on-site hotel guests’ privacy concerns. Using an unobtrusive method of data collection and text analytics, this study offers valuable insights into the categories of privacy, the most recurrent themes in hotel guests’ reviews and the potential relationship between cultural proximity and privacy concerns.

This paper aims to develop a novel and generalizable machine-learning based method of measuring established marketing constructs through passive analysis of consumer-generated textual data. The authors term this method scale-directed text analysis.

The method first develops a dictionary of words related to specific dimensions of a construct that is used to assess textual data from any source for a specific meaning. The method explicitly recognizes both specific words and the strength of their underlying sentiment.

Results calculated using this new approach are statistically equivalent to responses to traditional marketing scale items. These results demonstrate the validity of the authors’ methodology and show its potential to complement traditional survey approaches to assessing marketing constructs.

The method we outline relies on machine learning and thus requires either large volumes of text or a large number of cases. Results are reliable only at the aggregate level.

The method detail provides a means of less intrusive data collection such as through scraped social media postings. Alternatively, it also provides a means of analyzing data collected through more naturalistic methods such as open-response forms or even spoken language, both likely to increase response rates.

Scale-directed text analysis goes beyond traditional methods of conducting simple sentiment analysis and word frequency or percentage counts. It combines the richness of traditional textual and sentiment analysis with the theoretical structure and analytical rigor provided by traditional marketing scales, all in an automatic process.

The explosion of internet-generated content, coupled with methodologies such as sentiment analysis, present exciting opportunities for marketers to generate market intelligence on consumer attitudes and brand opinions. The purpose of this paper is to review the marketing literature on online sentiment analysis and examines the application of sentiment analysis from three main perspectives: the unit of analysis, sampling design and methods used in sentiment detection and statistical analysis.

The paper reviews the prior literature on the application of online sentiment analysis published in marketing journals over the period 2008-2016.

The findings highlight the uniqueness of online sentiment analysis in action-oriented marketing research and examine the technical, practical and ethical challenges faced by researchers.

The paper discusses the application of sentiment analysis in marketing research and offers recommendations to address the challenges researchers confront in using this technique.

This study provides academics and practitioners with a comprehensive review of the application of online sentiment analysis within the marketing discipline. The paper focuses attention on the limitations surrounding the utilization of this technique and provides suggestions for mitigating these challenges.

Purpose: To outline recent developments in digital service delivery in order to encourage researchers to pursue collaborations with computer science, operations research, and data science colleagues and to show how such collaborations can expand the scope of research on emotion in service delivery.

Design/methodology/approach: Uses archived resources available at http://LivePerson.com to extract data based in genuine service conversations between agents and customers. We refer to these as “digital traces” and analyze them using computational science models.

Findings: Although we do not test significance or causality, the data presented in this chapter provide a unique lens into the dynamics of emotions in service; results that are not obtainable using traditional research methods.

Research limitations/implications: This is a descriptive study where findings unravel new dynamics that should be followed up with more research, both research using traditional experimental methods, and digital traces research that allows inferences of causality.

Practical implications: The digital data and newly developed tools for sentiment analyses allow exploration of emotions in large samples of genuine customer service interactions. The research provides objective, unobtrusive views of customer emotions that draw directly from customer expressions, with no self-report intervention and biases.

Originality/value: This is the first objective and detailed depiction of the actual emotional encounters that customers express, and the first to analyze in detail the nature and content of customer service work.

This study aims to investigate motivations and human values of everyday consumers who participate in the annual day of consumption restraint known as Buy Nothing Day (BND). In addition, this study demonstrates a hybrid content analysis method in which artificial intelligence and human contributions are used in the data analysis.

This research uses a hybrid method of content analysis of a large Twitter data set spanning three years.

Consumer motivations are categorized as relating to consumerism, personal welfare, wastefulness, environment, inequality, anti-capitalism, financial responsibility, financial necessity, health, ethics and resistance to American culture. Of these, consumerism and personal welfare are the most common. Moreover, human values related to “openness to change” and “self-transcendence” were prominent in the BND tweets.

This research demonstrates the effectiveness of a hybrid content analysis methodology and uncovers the motivations and human values that average consumers (as opposed to consumer activists) have to restrain their consumption. This research also provides insight for firms wishing to better understand and respond to consumption restraint.

This research provides insight for firms wishing to better understand and respond to consumption restraint.

The question of why everyday consumers engage in consumption restraint has received little attention in the scholarly discourse; this research provides insight into “everyday” consumer motivations for engaging in restraint using a hybrid content analysis of a large data set spanning over three years.

This paper develops a generalizable, machine-learning-based method for measuring established marketing constructs using passive analysis of consumer-generated textual data from service reviews. The method is demonstrated using topic and sentiment analysis along dimensions of an existing scale: lodging quality index (LQI).

The method induces numerical scale ratings from text-based data such as consumer reviews. This is accomplished by automatically developing a dictionary from words within a set of existing scale items, rather a more manual process. This dictionary is used to analyze textual consumer review data, inducing topic and sentiment along various dimensions. Data produced is equivalent with Likert scores.

Paired t-tests reveal that the text analysis technique the authors develop produces data that is equivalent to Likert data from the same individual. Results from the authors’ second study apply the method to real-world consumer hotel reviews.

Results demonstrate a novel means of using natural language processing in a way to complement or replace traditional survey methods. The approach the authors outline unlocks the ability to rapidly and efficiently analyze text in terms of any existing scale without the need to first manually develop a dictionary.

The technique makes a methodological contribution by outlining a new means of generating scale-equivalent data from text alone. The method has the potential to both unlock entirely new sources of data and potentially change how service satisfaction is assessed and opens the door for analysis of text in terms of a wider range of constructs.

The purpose of this paper is to make two main contributions: first, showcase the diversity of service research in terms of the variety of used theories and methods, and second, explain (post-publication) success of articles operationalized as interest in an article (downloads), usage (citations), and awards (best paper nomination). From there, three sub-contributions are derived: stimulate a dialogue about existing norms and practices in the service field, enable and encourage openness amongst service scholars, and motivate scholars to join the field.

A mixed method approach is used in combining quantitative and qualitative research methods while analyzing 158 Journal of Service Management (JOSM) articles on several criteria such as their theory, methodology, and main descriptive elements (e.g. number of authors or references) and then using automated text analysis (e.g. investigating the readability of articles, etc.).

The results show that the JOSM publishes a large variety of articles with regard to theories, methods of data collection, and types of data analysis. For example, JOSM has published a mixture of qualitative and quantitative articles and papers containing firm-level and customer-level data. Further, the results show that even though conceptual articles create the same amount of interest (downloads), they are used more (citations).

This paper presents many descriptive results which do not allow for making inferences toward the entire service research discipline. Further, it is only based on one service research journal (JOSM) through a five-year span of publication.

The results have a number of implications for the discipline that are presented and discussed. Amongst them are that: the discipline should be more open toward conceptual articles, service research shows an imbalance toward theory testing, there is more potential to work with transactional data, and writing style should be more accessible (i.e. readable).

This paper is the first to conduct an in-depth analysis of service research articles to stimulate dialogue about common publishing practices in the JOSM and to increase the openness of the field.

The purpose of this study is to show how non-random groupings of YouTube videos can be combined with automated text analysis (ATA) of user comments to conduct quasi-experiments on consumer sentiment towards different types of brands in a naturalistic setting.

NCapture extracted thousands of comments on multiple videos representing different experimental treatments and Leximancer revealed differences in the lexical patterns of user comments for different types of brands.

User comments consistently revealed hypothesized relationships between brand types, based on existing theory regarding motivations for nostalgia and the relationship between consumer preferences, online product ratings and purchases. These results demonstrate the viability of conducting quasi-experimental research in naturalistic settings via non-random groupings of YT videos and ATA of user comments.

This research adopts a single quasi-experimental design: the non-equivalent group, after-only design. However, the same basic approach can be used with other quasi-experimental designs to examine different kinds of research questions.

Overall, this research points to the potential for ATA of comments on different categories of YT videos as a relatively straightforward approach for conducting field experiments that establish the ecological validity of laboratory findings. The method is easy to use and does not require the participation and cooperation of private, third party social media research companies.

Using composite style measures of the letter to shareholders, the purpose of this paper is to elaborate dominant rhetorical profiles and qualify them from an impression management (IM) perspective. In addition, the paper examines how institutional differences affect rhetorical profiles by comparing intensity and contingencies of rhetorical profiles of UK and US companies.

The authors use automated text analysis to capture linguistic style characteristics of a panel of UK and US companies and employ factor analysis to determine rhetorical profiles. Next, the authors investigate company-level and country-level determinants of a company’s rhetorical stance.

The authors document three prominent rhetorical profiles: an emphatic acclaiming stance, a cautious plausibility-based framing position, and a logic-based rationalizing orientation. The profiles represent distinct self-presentational logics and have different readability effects. Rhetorical IM is stronger in US companies, but higher expected scrutiny in the US institutional environment affects sensitivity of rhetorical postures to message credibility and litigation risk, while marginally increasing the less litigation-sensitive defensive framing style in US letters.

The authors develop replicable archival-based measures of prominent rhetorical IM traits of the shareholder letter, based on composite style features. The authors argue that they are qualitatively different from content-based IM proxies. The authors investigate their institutional and organizational relevance by examining how company features and country-level differences affect incentives and constraints for style-based rhetorical IM.