Computational Linguistic Analysis Research Articles

How Many Latours Is Too Many? Measuring Brand Name Congestion in Bordeaux Wine

AbstractFirms rely on brand names to market goods to consumers, and consumers rely on brand names to locate goods that satisfy their preferences. If multiple firms are using the same or similar names, consumers may be confused about which product to buy, and firms may not obtain the benefits of their investments in quality. Recently, both firms and scholars in a number of industries have expressed concern about brand name congestion—too many firms clustering around too few terms. This paper applies computational linguistic analysis to chateau names in the Bordeaux wine region to study the degree of brand congestion within a mature, traditional, and high-value market. We find that Bordeaux producers have highly similar names to one another, far more than in comparable wine regions such as California and Alsace. More than a quarter of all Bordeaux producers have a name that is identical or nearly so to at least one other producer, and many terms are claimed by dozens of different producers. Interestingly, however, we find that the most famous and renowned producers have names that tend to be more distinctive than their less famous brethren. (JEL Classifications: C88, D83, L66, O34)

Regional personality assessment through social media language.

We explore the personality of counties as assessed through linguistic patterns on social media. Such studies were previously limited by the cost and feasibility of large-scale surveys; however, language-based computational models applied to large social media datasets now allow for large-scale personality assessment. We applied a language-based assessment of the five factor model of personality to 6,064,267 U.S. Twitter users. We aggregated the Twitter-based personality scores to 2,041 counties and compared to political, economic, social, and health outcomes measured through surveys and by government agencies. There was significant personality variation across counties. Openness to experience was higher on the coasts, conscientiousness was uniformly spread, extraversion was higher in southern states, agreeableness was higher in western states, and emotional stability was highest in the south. Across 13 outcomes, language-based personality estimates replicated patterns that have been observed in individual-level and geographic studies. This includes higher Republican vote share in less agreeable counties and increased life satisfaction in more conscientious counties. Results suggest that regions vary in their personality and that these differences can be studied through computational linguistic analysis of social media. Furthermore, these methods may be used to explore other psychological constructs across geographies.

Linguistic complexity in teachers' assessment of German essays in high stakes testing

In national educational standards, linguistic complexity is one of the core aspects characterizing the demands of reading material and student writing quality. In Germany, teachers are supposed to consider linguistic complexity in their assessment of language arts Abitur examinations, the high stakes test qualifying for higher education admission in Germany. In the present study, we investigated if and how teachers consider linguistic complexity in their assessment of writing quality. To systematically identify a range of linguistic complexity aspects, we conducted computational linguistic analyses of student essays (N = 344) written during Abitur examinations in the subject German. Experienced teachers (N = 33) then regraded a subsample of the essays (n = 16) and rated their linguistic complexity. We used mixed-effects models to study the relationships between computational measures of linguistic complexity, teachers’ assessment of writing quality (language, content, and overall grade), and their linguistic complexity rating. Results confirm that teachers’ complexity rating and their assessment of the language aspect of writing quality is related to the computed linguistic complexity features, primarily measures of syntactic complexity and language use (word frequency measures). But findings also show that teachers have difficulties in separating linguistic complexity from aspects of content and language accuracy.

Illusory social agents within and beyond voices: Acomputational linguistics analysis of the experience of psychosis.

ObjectivesPsychosis has a strong social component and often involves the experience of being affected by ‘illusory social agents’. However, this experience remains under‐characterized, particularly for social agents in delusions and non‐vocal hallucinations. One useful approach is a form of computational linguistics called corpus linguistics that studies texts to identify patterns of meaning encoded in both the semantics and linguistic structure of the text.MethodsTwenty people living with psychosis were recruited from community and inpatient services. They participated in open‐ended interviews on their experiences of social agents in psychosis and completed a measure of psychotic symptoms. Corpus linguistics analysis was used to identify key phenomenological features of vocal and non‐vocal social agents in psychosis.ResultsSocial agents i) are represented with varying levels of richness in participants’ experiences, ii) are attributed with different kinds of identities including physical characteristics and names, iii) are perceived to have internal states and motivations that are different from those of the participants, and iv) interact with participants in various ways including through communicative speech acts, affecting participants’ bodies, and moving through space. These representations were equally rich for agents associated with hallucinated voices and those associated with non‐vocal hallucinations and delusions.ConclusionsWe show that the experience of illusory social agents is a rich and complex social experience reflecting many aspects of genuine social interaction and is not solely present in auditory hallucinations, but also in delusions and non‐vocal hallucinations.Practitioner Points The experience of being affected by illusory social agents in psychosis extends beyond hallucinated voices.They are a rich and complex social experience reflecting many aspects of genuine social interaction.These are also likely to be a source of significant distress and disability.

Culture Clash in Organizations

This symposium is designed to advance research on cultural clash, the dissonant interactions of members from separate cultures, within and beyond organizations. The literature has long established that culture clash degrades organizational performance. The papers in this symposium prod this assumption, examining its mechanisms and scope conditions, establishing when and how culture clash occurs. The first presentation examines how status disagreement emerges from diversity in collectivist orientation, undermining group productivity. The second presentation studies group assimilation and power asymmetries in a post-merger organization using computational linguistic analysis of email communication. The third presentation looks at culture clash among founding teams in start-ups, finding that founding teams with high interpersonal cultural diversity adopt industry-typical cultural scripts while founding teams with high intrapersonal cultural diversity cultivate distinctive cultures. Finally, the fourth presentation theorizes how extreme groups build strong cultural beliefs and attract new members by studying the proliferation of Covid-19 conspiracy theories on Twitter. Taken together, the papers in this symposium revisit old assumptions and theories of culture clash, updating them with innovative experiments, modern computational linguistic techniques, and novel data sets. Perceiving Status through Different Cultural Lenses Presenter: Sujin Jang; INSEAD Presenter: Catarina Fernandes; Emory U., Goizueta Business School A Language-Based Method for Measuring Inter-Group Boundaries and Assimilation Presenter: Anjali M. Bhatt; Harvard Business School Presenter: Amir Goldberg; Stanford U. Agreeing to Be Different: Founding Team Cultural Diversity and Startup Cultural Atypicality Presenter: Yeonsin Ahn; INSEAD The Other Pandemic: #Coronahoax Conspiracy Theories Presenter: Henrich Greve; INSEAD Presenter: Hayagreeva Rao; Stanford U. Presenter: Paul Vicinanza; Stanford Graduate School of Business Presenter: Yan Echo Zhou; -

Main topics in EIST during its first decade: A computational-linguistic analysis

We analyse 465 articles published in EIST from June 2011 until June 2021 to identify topics addressed in the journal. We find eight main topics and assess how their shares changed over time as well as how many citations they received. The topics with the largest shares in all publications are “Theory of socio-technical transitions” and “Urban regimes and niches”. The two most cited topics, “Theory of socio-technical transitions” and “Geography and diffusion of eco-innovations”, showed a rising share over time, while the share of topic “Finance, investment and growth” declined. We further assess the geographical coverage of topics, through affiliations of the corresponding authors. The resulting map indicates dominant topics for the 34 countries that contributed to publications in EIST.

Research on a voice changed by distortion

The concept and essence of distortion are considered from a technical stand-point and criminal law perspective. The most common ways of distorting a voice and a speech are provided, as well as certain methods of detecting an intentional change in a voice and human speech by computer tools and linguistic analysis.
 Some software and hardware tools changing a speech signal both in real time and in a pre-prepared recording are analyzed.
 To solve diagnostic and identification tasks, a pressing issue in forensic video and audio analysis is studied which is addressed to forensic experts in this field more often.

Computational linguistic analysis applied to a semantic fluency task: A replication among first-episode psychosis patients with and without derailment and tangentiality

Automated tools do not yet exist to measure formal thought disorder, including derailment and tangentiality, both of which can be subjectively rated using the Scale for the Assessment of Positive Symptoms after a clinical research interview. CoVec, a new automated tool, measures the semantic similarity among words averaged in a five- and ten-word window (Coherence-5 and Coherence-10, respectively). One prior report demonstrated that this tool was able to differentiate between patients with those types of thought disorder and patients without them (and controls). Here, we attempted a replication of the initial findings using data from a different sample of patients hospitalized for initial evaluation of first-episode psychosis. Participants were administered a semantic fluency task and the animal lists were analyzed with CoVec. In this study, we partially replicated the prior findings, showing that first-episode patients with derailment had significantly lower Coherence-5 and Coherence-10 compared with patients without derailment. Further research is warranted on this and other highly reliable and objective methods of detecting formal thought disorder through simple assessments such as semantic fluency tasks.

Narrative Investment-Risk Disclosure & REIT Investment

We study narrative investment-risk disclosure, applying computational linguistic analysis to the qualitative discussions of risk throughout 10-K filings. Our sample covers the real estate investment trust (REIT) sector, where firms rely heavily on external capital sources to fund new investment. Using granular investment data from a source that is independent from annual filings, we document a positive relation between changes to investment-risk disclosure and changes to actual investment. Such linkage previously evades the disclosure-informativeness literature, but is feasible in our study based on data advantages to the REIT sector. Changes to risk disclosure do not appear to increase simply the perception of risk among investors. Unlike evidence from the broader market, we find no significant relation to the volatility of stock returns or the volatility of analyst forecast revisions following the filing. Instead, changes to investment-risk disclosure appear informative as they are followed by a pronounced enhancement to market liquidity. The average increase in post-filing trading volume for firms that increase acquisitions is 8%, conditional on a top-tercile increase to investment-risk disclosure; or 0%, when disclosure is non-increasing.

Educational Research: Why Medical Students Choose Neurology: A Computational Linguistics Analysis of Personal Statements.

To understand medical students' motivations for choosing neurology and how applicants conceptualize the field, as this information can be used to enhance interest in neurology and develop educational programs to help identify, support, and recruit future neurologists. Applicants to neurology residencies submit personal statements describing themselves and their motivations. Textual analysis of personal statements has been performed in internal medicine and general surgery, but never before in neurology. We hypothesized that specific words and themes would be mentioned in residency personal statements with high frequencies indicating students' motivations. We used computational linguistics software to assess key words and thereby study motivations, expectations, and themes present among neurology applicants. A total of 2,405 personal statements submitted over 5 years to our institution were de-identified and compiled into a database for evaluation through 3 computational linguistics software programs. We performed calculations of term frequencies (TF) and TF-inverse document frequencies and performed K-means clustering to identify unique words and common themes. Specific disease states were discussed. For example, stroke (TF 2,178), epilepsy (TF 970), and dementia (TF 944) were referenced more often than amyotrophic lateral sclerosis (TF 220) and carpal tunnel (TF 10). The most common proper names cited were Oliver Sacks (TF 94) and Sherlock Holmes (TF 41). Common themes included fascination with the brain, interest in research, desire to help patients, early interests in neurology, continued pursuit of learning, appreciation for time with patients, family history with neurologic illness, and intellectual curiosity. This first computational linguistic analysis of neurology personal statements provides understanding into medical students' motivations and interests. Ongoing subgroup and thematic analyses may inform educational strategies and enhance recruitment to our field.

Verbal Communication in Robotics: A Study on Salient Terms, Research Fields and Trends in the Last Decades Based on a Computational Linguistic Analysis

Verbal communication is an expanding field in robotics showing a significant increase in both the industrial and research field. The application of verbal communication in robotics aims to reach a natural human-like interaction with robots. In this study, we investigated how salient terms related to verbal communication in robotics have evolved over the years, what are the topics that recur in the related literature, and what are their trends. The study is based on a computational linguistic analysis conducted on a database of 7,435 scientific publications over the last 2 decades. This comprehensive dataset was extracted from the Scopus database using specific key-words. Our results show how relevant terms of verbal communication evolved, which are the main coherent topics and how they have changed over the years. We highlighted positive and negative trends for the most coherent topics and the distribution over the years for the most significant ones. In particular, verbal communication resulted in being highly relevant for social robotics. Potentially, achieving natural verbal communication with a robot can have a great impact on the scientific, societal, and economic role of robotics in the future.

Novel Computational Linguistic Measures, Dialogue System and the Development of SOPHIE: Standardized Online Patient for Healthcare Interaction Education

In this article, we describe the iterative participatory design of SOPHIE, an online virtual patient for feedback-based practice of sensitive patient-physician conversations, and discuss an initial qualitative evaluation of the system by professional end users. The design of SOPHIE was motivated from a computational linguistic analysis of the transcripts of 383 patient-physician conversations from an essential office visit of late stage cancer patients with their oncologists. We developed methods for the automatic detection of two behavioral paradigms, lecturing and positive language usage patterns (sentiment trajectory of conversation), that are shown to be significantly associated with patient prognosis understanding. These automated metrics associated with effective communication were incorporated into SOPHIE, and a pilot user study identified that SOPHIE was favorably reviewed by a user group of practicing physicians.

Measuring Brand Congestion in Bordeaux: Too Many Towers, Hills, and Crosses?

Firms rely on brand names to market goods to consumers, and consumers rely on brand names to locate goods that satisfy their preferences. If multiple firms are using the same or similar names, consumers may be confused about which product to buy, and firms may not obtain the benefits of their investments in quality. Recently, both firms and scholars in a number of industries have expressed concern about brand name congestion—too many firms clustering around too few terms. This paper applies computational linguistic analysis to chateau names in the Bordeaux wine region to study the degree of brand congestion within a mature, traditional, and high-value market. We find that Bordeaux producers have highly similar names to one another, far more than in comparable wine regions such as California and Alsace. More than a quarter of all Bordeaux producers have a name that is identical or nearly so to at least one other producer, and many terms are claimed by dozens of different producers. Interestingly, however, we find that the most famous and renowned producers have names that tend to be more distinctive than their less famous brethren. We also provide access to our dataset containing all of the official classifications of Bordeaux wines since 1855.

Method of syntactic-semantik analysis of unstructured texts

The methodology of computer linguistic analysis of unstructured text information is considered. The basic provisions of various approaches to the automation of automatic processing of unstructured texts in computer systems are presented. A formalized description of the method of syntactic-semantic computer analysis of texts is given.

Free associations of citizens and scientists with economic and green growth: A computational-linguistics analysis

The debate about the relationship between economic growth and environmental sustainability triggers a range of associations. Here we analyze open-ended textual responses of citizens and scientists concerning their associations with the terms “economic growth” and “green growth”. We derive from the responses a number of topics and examine how associations differ across distinct opinion segments of people, namely supporters of Green growth, Agrowth and Degrowth. The results indicate that the general public is more critical of the notion of economic growth than academic researchers. Citizens stress problems of corruption, social inequality, unemployment and poverty, with less variation among the three opinion segments compared to scientists. The latter more strongly emphasize the environmental consequences of economic growth. Concerning associations of scientists with the term “green growth”, we find topics questioning its feasibility to be more likely expressed by Degrowth supporters, while topics stressing the possibility of sustainable economic growth by Green growth supporters. We find that topic polarization is stronger for scientists than citizens. Our results provide further validation for opinion clusters identified in previous studies and uncover additional insights about related views on growth and sustainability.

Public views on carbon taxation and its fairness: a computational-linguistics analysis

Carbon taxes evoke a variety of public responses, often with negative implications for policy support, implementation, and stringency. Here we use topic modeling to analyze associations of Spanish citizens with a policy proposal to introduce a carbon tax. This involves asking two key questions, to elicit (1) citizens’ associations with a carbon tax and (2) their judgment of the fairness of such a policy for distinct uses of tax revenues. We identify 11 topics for the first question and 18 topics for the second. We perform regression analysis to assess how respondents’ associations relate to their carbon tax acceptability, knowledge, and sociodemographic characteristics. The results show that, compared to people accepting the carbon tax, those rejecting it show less trust in politicians, think that the rich should pay more than the poor, consider the tax to be less fair, and stress more a lack of renewable energy or low-carbon transport. Respondents accepting a carbon tax emphasize more the need to solve environmental problems and care about a just society. These insights can help policymakers to improve the design and communication of climate policy with the aim to increase its public acceptability.

Designing a Clinician-Facing Tool for Using Insights From Patients' Social Media Activity: Iterative Co-Design Approach.

BackgroundRecent research has emphasized the need for accessing information about patients to augment mental health patients’ verbal reports in clinical settings. Although it has not been introduced in clinical settings, computational linguistic analysis on social media has proved it can infer mental health attributes, implying a potential use as collateral information at the point of care. To realize this potential and make social media insights actionable to clinical decision making, the gaps between computational linguistic analysis on social media and the current work practices of mental health clinicians must be bridged.ObjectiveThis study aimed to identify information derived from patients’ social media data that can benefit clinicians and to develop a set of design implications, via a series of low-fidelity (lo-fi) prototypes, on how to deliver the information at the point of care.MethodsA team of clinical researchers and human-computer interaction (HCI) researchers conducted a long-term co-design activity for over 6 months. The needs-affordances analysis framework was used to refine the clinicians’ potential needs, which can be supported by patients’ social media data. On the basis of those identified needs, the HCI researchers iteratively created 3 different lo-fi prototypes. The prototypes were shared with both groups of researchers via a videoconferencing software for discussion and feedback. During the remote meetings, potential clinical utility, potential use of the different prototypes in a treatment setting, and areas of improvement were discussed.ResultsOur first prototype was a card-type interface that supported treatment goal tracking. Each card included attribute levels: depression, anxiety, social activities, alcohol, and drug use. This version confirmed what types of information are helpful but revealed the need for a glanceable dashboard that highlights the trends of these information. As a result, we then developed the second prototype, an interface that shows the clinical state and trend. We found that focusing more on the changes since the last visit without visual representation can be more compatible with clinicians’ work practices. In addition, the second phase of needs-affordances analysis identified 3 categories of information relevant to patients with schizophrenia: symptoms related to psychosis, symptoms related to mood and anxiety, and social functioning. Finally, we developed the third prototype, a clinical summary dashboard that showed changes from the last visit in plain texts and contrasting colors.ConclusionsThis exploratory co-design research confirmed that mental health attributes inferred from patients’ social media data can be useful for clinicians, although it also revealed a gap between computational social media analyses and clinicians’ expectations and conceptualizations of patients’ mental health states. In summary, the iterative co-design process crystallized design directions for the future interface, including how we can organize and provide symptom-related information in a way that minimizes the clinicians’ workloads.

A Framework for the Computational Linguistic Analysis of Dehumanization.

Dehumanization is a pernicious psychological process that often leads to extreme intergroup bias, hate speech, and violence aimed at targeted social groups. Despite these serious consequences and the wealth of available data, dehumanization has not yet been computationally studied on a large scale. Drawing upon social psychology research, we create a computational linguistic framework for analyzing dehumanizing language by identifying linguistic correlates of salient components of dehumanization. We then apply this framework to analyze discussions of LGBTQ people in the New York Times from 1986 to 2015. Overall, we find increasingly humanizing descriptions of LGBTQ people over time. However, we find that the label homosexual has emerged to be much more strongly associated with dehumanizing attitudes than other labels, such as gay. Our proposed techniques highlight processes of linguistic variation and change in discourses surrounding marginalized groups. Furthermore, the ability to analyze dehumanizing language at a large scale has implications for automatically detecting and understanding media bias as well as abusive language online.

A human-machine partnered approach for identifying social media signals of elevated traumatic grief in Chicago gang territories.

There is a critical need to improve trauma-informed services in structurally marginalized communities impacted by violence and its associated traumatic grief. For community residents, particularly gang-associated youth, repeated exposure to traumatic grief causes serious adverse effects that may include negative health outcomes, delinquency, and future violent offenses. The recent proliferation of digital social media platforms, such as Twitter, provide a novel and largely underutilized resource for responding to these issues, particularly among these difficult-to-reach communities. In this paper, we explore the potential for using a human-machine partnered approach, wherein qualitative fieldwork and domain expertise is combined with a computational linguistic analysis of Twitter content among 18 gang territories/neighborhoods on Chicago’s South Side. We first employ in-depth interviews and observations to identify common patterns by which residents in gang territories/neighborhoods express traumatic grief on social media. We leverage these qualitative findings, supplemented by domain expertise and computational techniques, to gather both traumatic grief- and gang-related tweets from Twitter. We next utilize supervised machine learning to construct a binary classification algorithm to eliminate irrelevant tweets that may have been gathered by our automated query and extraction techniques. Last, we confirm the validity, or ground truth, of our computational findings by enlisting additional domain expertise and further qualitative analyses of the specific traumatic events discussed in our sample of Twitter content. Using this approach, we find that social media provides useful signals for identifying moments of increased collective traumatic grief among residents in gang territories/neighborhoods. This is the first study to leverage Twitter to systematically ground the collective online articulations of traumatic grief in traumatic offline events occurring in violence-impacted communities. The results of this study will be useful for developing more effective tools—including trauma-informed intervention applications—for community organizations, violence prevention initiatives, and other public health efforts.

Real Talk About Fake News: Identity Language and Disconnected Networks of the US Public’s “Fake News” Discourse on Twitter

This article studies “fake news” beyond the consumption and dissemination of misinformation and disinformation. We uncover how the term “fake news” serves as a discursive device for ordinary citizens to consolidate group identity in everyday political utterances on Twitter. Using computational linguistic and network analyses, we demonstrate that over the period of 2016–2018, there is an uptrend in the use of identity language in US Twitter users’ discussions about “fake news,” manifested by the increased frequency of group pronouns in combination with issues and sentiments that boost one’s ingroup and derogate the outgroup. Furthermore, as opposed to the conventional wisdom that “fake news” is a right-wing term, we uncover two disconnected retweet networks surrounding liberal and conservative opinion leaders. Like-minded individuals selectively amplify ingroup messages to claim the power to define falsehood and make group-serving blame attributions. We discuss the theoretical implications of our findings and offer new directions for future research on “fake news,” misinformation, and disinformation.