Computerized Text Analysis Research Articles

Lost in transmission? Self- and other-annotation of emotional words

It is generally assumed that someone's affective state can be correctly detected and interpreted by other people, nowadays even by computer algorithms, in their writing. However, it is unclear whether these perceptions match the actual experience and communicative intention of the author. Therefore, we investigated the relation between affect expression and perception in text in a two-part study. In Part 1, participants (authors) wrote about emotional experiences according to four combinations of two appraisals (High/Low Pleasantness, High/Low Control), rated the valence of each text, and annotated words using 22 emotions. In Part 2, another group of participants (readers) rated and annotated the same texts. We also compare the human evaluations to those provided by computerized text analysis. Results show that valence differed across conditions and that authors rated and annotated their texts differently than readers. Although the automatic analysis detected levels of positivity and negativity across conditions similar to human valence ratings, it relied on fewer and different words to do so. We discuss implications for affective science and automatic sentiment analysis.

Framing Requests for Help: Language and Gender in Parental Advice-Seeking Letters

Confirming the importance of language to not just relay thoughts but also construct hierarchies, 251 parental advice-seeking letters written at the dawn of the parent education movement in the 1920s and 1930s were coded with the Linguistic Inquiry and Word Count computerized text analysis program. An examination of language use patterns shows how parents framed their requests and how they manifested signs of “doing gender” in the process. The fathers’ and mothers’ letters differed in length, common verb use, negation word use, pronoun use, preposition use, and locus on three summary variables (analytic, authenticity, and clout). They did not differ in the use of words associated with general affect, positive emotion, negative emotion, anger, and anxiety, thus exemplifying how the practice of “doing gender” can be circumstantial. The findings underscore the importance of studying the rhetorical aspects of parental advice-seeking communications and illustrate how language and gender influence these communications.

Contrasting compulsive behaviour: Computerized text analysis of compulsion narratives.

Those who gamble compulsively, and those who shop or buy in a compulsive manner share a number of common characteristics, stemming from similar impulse-control issues. As such, it is predicted that a lexical analysis of personal narratives of compulsion would share similarities. Using secondary data from an online mental health forum, Psychforums, the research analyzed narratives of compulsive gambling (n = 199) and compulsive buying (n = 196) using the automated text analysis tool, LIWC. The results indicated that compulsive buying narratives rated significantly higher in clout and emotional tone and significantly lower in authenticity, with no significant differences noted in analytical thinking between the two compulsion narratives. Recommendations for future research include that demographic variables be incorporated and that narratives sourced from different online platforms should be contrasted.

Latent Linguistic Motifs in Social Media Postings Resisting COVID-19 Misinformation.

Social media has become a predominant source of information for many health care consumers. However, false and misleading information is a pervasive problem in this context. Specifically, health-related misinformation has become a significant public health challenge, impeding the effectiveness of public health awareness campaigns and resulting in suboptimal responsiveness to the communication of legitimate risk-related information. Little is known about the mechanisms driving the seeding and spreading of such information. In this paper, we specifically examine COVID-19 tweets which attempt to correct misinformation. We employ a mixed-methods approach comprising qualitative coding, deep learning classification, and computerized text analysis to understand the manifestation of speech acts and other linguistic variables. Results indicate significant differences in linguistic variables (e.g., positive emotion, tone, authenticity) of corrective tweets and their dissemination level. Our deep learning classifier has a macro average performance of 0.82. Implications for effective and persuasive misinformation correction efforts are discussed.

Can We Algorithmize Politics? The Promise and Perils of Computerized Text Analysis in Political Research

ABSTRACTIn recent years, political scientists increasingly have used data-science tools to research political processes, positions, and behaviors. Because both domestic and international politics are grounded in oral and written texts, computerized text analysis (CTA)—typically based on natural-language processing—has become one of the most notable applications of data-science tools in political research. This article explores the promises and perils of using CTA methods in political research and, specifically, the study of international relations. We highlight fundamental analytical and methodological gaps that hinder application and review processes. Whereas we acknowledge the significant contribution of CTA to political research, we identify a dual “engagement deficit” that may distance those without prior background in data science: (1) the tendency to prioritize methodological innovation over analytical and theoretical insights; and (2) the scholarly and political costs of requiring high proficiency levels and training to comprehend, assess, and use advanced research models.

A Mixed-Method Comparison of Therapist and Client Language across Four Therapeutic Approaches

This paper illustrates a methodological approach that combines computerized text analysis, quantitative analysis, and qualitative discourse analysis in comparing large bodies of therapeutic language. More specifically, it explores how language use in psychotherapy is associated with different therapeutic approaches and therapeutic roles (i.e., therapists and clients). The dataset consisted of 155 therapeutic sessions (over 1,057,000 words) that are illustrative of four approaches, i.e., psychoanalysis, humanistic therapy, Cognitive Behavioral Therapy (CBT), and eclectic therapy. The transcripts were divided according to therapeutic approaches and therapeutic roles and processed using Linguistic Inquiry Word Count (LIWC) in terms of four summary variables, i.e., analytical thinking, clout, emotional tones, and authenticity. A series of mixed-effects models with session as the random effect was fitted, and the statistical patterns were illustrated using linguistic examples and discussed from a discourse analytic perspective. The approach demonstrates methodological strengths in exploring large-scale data and expanding the research scope permitted by traditional discourse analysis. The findings underline professional knowledge and institutionalized roles as key factors influencing the use of therapeutic language, providing meaningful insights for the clinical understanding and future research into therapeutic language.

Identifying the linguistic markers of intuition in human resource (HR) practice

AbstractIntuitions are judgements that arise automatically and non‐consciously. Recognising when intuitive judgements are being used and whether doing so is appropriate is an important skill both for HR practitioners and managers involved in human resource (HR) processes. Because ‘intuiting’ is involuntary and unconscious it is difficult to access, monitor and control, however people can access and articulate their intuitive judgements through the spoken/written word. Identifying ‘linguistic markers’ of intuitions in spoken/written communications could help identify when/how intuition is being used in HR and improve HR practices in areas such as selection (e.g., obviating implicit/unconscious biases) and creativity (e.g., as a source of insights). This research used computerised text analysis (CTA) to identify linguistic markers of intuition based on HR practitioners' descriptions of what happens when they ‘intuit’. We outline implications of these findings for improving HR decision‐making processes and practices and their potential applications in data analytics, AI and machine learning in HR.

Do gender and racial differences exist in letters of recommendation for obstetrics and gynecology residency applicants?

Studies in several fields demonstrate gender and racial differences in descriptions of applicants in letters of recommendation. However, gender and racial biases in obstetrics and gynecology residency letters of recommendation are understudied. This study aimed to determine whether there are differences in the letters of recommendation written for medical students applying for obstetrics and gynecology residency based on applicant gender and underrepresented in medicine status. We analyzed all letters of recommendation submitted to a single obstetrics and gynecology residency program during the 2019-2020 application cycle using a computerized text analysis software to evaluate the presence of 25 word categories. Multivariable regression was used to compare differences in the frequency of word categories by gender and underrepresented in medicine status. Concurrently, we performed a qualitative content analysis of letters from a random sample of applicants balanced for gender and underrepresented in medicine status and explored emerging themes until thematic saturation was reached. We analyzed 3060 letters written for 834 applicants, 721 (87%) of whom were women and 198 (24%) of whom were underrepresented in medicine. Men authored 1605 letters (53%). There was no difference in word categories used in letters written for men and women applicants on multivariable analysis. Similarly, there was no word category difference between letters written for applicants who were underrepresented in medicine compared with White and Asian applicants. However, women letter writers used more communal (relationship-oriented) words compared with men letter writers (P<.001). Moreover, on the qualitative analysis of letters from a random sample of 30 applicants, comments about personality traits were frequent in all letters. Comments on surgical skills, work ethic, and leadership were found more in letters for White and Asian applicants, and phrases that were doubt raisers were found more in letters written for underrepresented medicine applicants. On linguistic analysis, letters of recommendation written for obstetrics and gynecology residency applicants were similar overall compared with applicant race and gender.

How may I help you? Driving brand engagement through the warmth of an initial chatbot message

Despite the growing number of brands that rely on chatbots to address customer service inquiries that once required human intervention, academics and practitioners are only beginning to acknowledge the role of chatbots in brand-building activities. Chatbots can initiate online conversations, thereby often serving as a consumer’s first brand impression. However, little is known about how managers can strategically tailor a chatbot’s initial message to foster consumer–brand connections and, ultimately, engagement. Three studies demonstrate that when chatbots initiate a conversation using a warm (vs. competent) message, brand engagement increases, as assessed using both computerized text analysis and traditional scale measures. Brand–self distance mediates this effect, such that a warm (vs. competent) initial chatbot message makes consumers feel closer to the brand. Further, the authors identify brand affiliation as a theoretically relevant moderator. This research thus offers managers insight into how initial chatbot messages can attract and engage consumers.

Sexual Dimorphism in Language, and the Gender Shift Hypothesis of Homosexuality.

Psychological sex differences have been studied scientifically for more than a century, yet linguists still debate about the existence, magnitude, and causes of such differences in language use. Advances in psychology and cognitive neuroscience have shown the importance of sex and sexual orientation for various psychobehavioural traits, but the extent to which such differences manifest in language use is largely unexplored. Using computerised text analysis (Linguistic Inquiry and Word Count: LIWC 2015), this study found substantial psycholinguistic sexual dimorphism in a large corpus of English-language novels (n = 304) by heterosexual authors. The psycholinguistic sex differences largely aligned with known psychological sex differences, such as empathising–systemising, people–things orientation, and men’s more pronounced spatial cognitive styles and abilities. Furthermore, consistent with predictions from cognitive neuroscience, novels (n = 158) by lesbian authors showed minor signs of psycholinguistic masculinisation, while novels (n = 167) by homosexual men had a female-typical psycholinguistic pattern, supporting the gender shift hypothesis of homosexuality. The findings on this large corpus of 66.9 million words indicate how psychological group differences based on sex and sexual orientation manifest in language use in two centuries of literary art.

Analysis of Online Peripartum Depression Communities: Application of Multilabel Text Classification Techniques to Inform Digitally-Mediated Prevention and Management.

Peripartum depression (PPD) is a significant public health problem, yet many women who experience PPD do not receive adequate treatment. In many cases, this is due to social stigmas surrounding PPD that prevent women from disclosing their symptoms to their providers. Examples of these are fear of being labeled a “bad mother,” or having misinformed expectations regarding motherhood. Online forums dedicated to PPD can provide a practical setting where women can better manage their mental health in the peripartum period. Data from such forums can be systematically analyzed to understand the technology and information needs of women experiencing PPD. However, deeper insights are needed on how best to translate information derived from online forum data into digital health features. In this study, we aim to adapt a digital health development framework, Digilego, toward translation of our results from social media analysis to inform digital features of a mobile intervention that promotes PPD prevention and self-management. The first step in our adaption was to conduct a user need analysis through semi-automated analysis of peer interactions in two highly popular PPD online forums: What to Expect and BabyCenter. This included the development of a machine learning pipeline that allowed us to automatically classify user post content according to major communication themes that manifested in the forums. This was followed by mapping the results of our user needs analysis to existing behavior change and engagement optimization models. Our analysis has revealed major themes being discussed by users of these online forums- family and friends, medications, symptom disclosure, breastfeeding, and social support in the peripartum period. Our results indicate that Random Forest was the best performing model in automatic text classification of user posts, when compared to Support Vector Machine, and Logistic Regression models. Computerized text analysis revealed that posts had an average length of 94 words, and had a balance between positive and negative emotions. Our Digilego-powered theory mapping also indicated that digital platforms dedicated to PPD prevention and management should contain features ranging from educational content on practical aspects of the peripartum period to inclusion of collaborative care processes that support shared decision making, as well as forum moderation strategies to address issues with cyberbullying.

Reading Between the Lines: Understanding Customer Experience With Disruptive Technology Through Online Reviews

A customer’s experience with a brand, as evidenced in online customer reviews, has attracted multidisciplinary scholarly attention. Customer experience plays an important role as an antecedent to brand engagement, brand adoption, and eventual brand loyalty. Thus, it is important for businesses to understand their customers’ experiences so that they can make changes as necessary. The COVID-19 pandemic has brought unprecedented changes to the business landscape, forcing businesses to move online, with many utilizing enterprise video conferencing (EVC) to maintain daily operations. To ensure efficient digitization, many turned to the online reviews of others’ experiences with EVC before engaging with it themselves. This research examined how the customer experience is portrayed through emotional tone and word choice in online reviews for the EVC platform Zoom. Using computerized text analysis, key differences were found in the emotional tone and word choice for low- and high-rated reviews. The complexity and emotionality expressed in reviews have implications on the usability of the review for others. The results from this study suggest that online customer reviews with a high rating express a higher level of expertise and confidence than low-rated reviews. Given the potential dissemination and impact, digital marketers may be well advised to first and foremost respond to online reviews that are high in emotional tone.

Referential Activity Differentially Mediates Expression of Positive and Negative Emotions in Borderline Personality Disorder.

Borderline Personality Disorder (BPD) is a high risk and prevalent personality disorder that is associated with increased negative emotions, decreased positive emotions, and impairments in symbolization and impaired reflective functioning. These dimensions, while they may impact one another, have not been investigated concurrently from qualitative, linguistic narratives. We hypothesized a BPD group would have lower expression of positive emotions and greater expression of negative emotions and less reflective function than healthy controls. Additionally, we explored the role of referential activity (an index of symbolic capacity) between BPD and healthy controls in the context of valenced emotional expression. An adult, female BPD group (n = 13) and a demographically matched healthy control group (n = 14) were recruited and administered the Adult Attachment Interview and/or the Object Relations Inventory. Computerized text analyses were used to assess positive emotion and negative emotion, the Weighted Referential Activity Dictionaries to assess referential activity, and the Computerized Reflective Function dictionary. On the Object Relations Inventory, the BPD group expressed more frequent negative emotions and less frequent positive emotions; on the Adult Attachment Interview, the BPD group exhibited less expression of positive emotions. There were no differences between BPD and controls on referential activity or reflective functioning on either interview. However, BPD status fully mediated the significant relationship between referential activity and negative emotion expression. The BPD group utilized more referential activity when expressing negative emotions than controls. Conversely, the control group utilized more referential activity when expressing positive emotions than controls. Referential activity seems to play an important role in explaining the BPD versus control difference in valenced linguistic emotional expression. Furthermore, these results suggest the object relations inventory elicits more robust linguistic features relevant to BPD.

Computerized Text Analysis on Self-Description Text to Get Student’s Prevailing, Confidence, and Drives

Using a sample of 26 participants, we asked the students to prepare a text in written self-description at the end of the final path presentation in the sixth semester. One interesting thing is the relationship between personality differences and the types of words that people use. Recent studies covers an observing the relationship between personal qualities and the use of language in creating self-description. Text files are organized for transcription LIWC analysis to calculate the percentage of clout, authentic, and drives that student’s use. The results supported this possibility: LIWC’s clout, authentic, and drives dictionaries can be used as indicators of the student’s prevailing, confidence, motives and needs, we found that when calculated on written text, LIWC scores were connected with language dimension and psychology issues. These findings recommend that LIWC clout, authentic, motives, and needs dictionaries possibly will catch self-reported subjective emotion experience when applied to everyday activities. To enrich students’ confidence, power, motives, and needs toward the psychology issues faced by the students, students should be stimulated to discover the field of psychological processes and linguistic patterns by detecting their emotional and perceptual experiences.

Student’s Prevailing, Confidence, and Drives: Computerized Text Analysis on Self-Description Text

Using a sample of 26 participants, we asked the students to prepare a text in written self-description at the end of the final path presentation in the sixth semester. One interesting thing is the relationship between personality differences and the types of words that people use. Recent studies covers an observing the relationship between personal qualities and the use of language in creating self-description. Text files are organized for transcription LIWC analysis to calculate the percentage of clout, authentic, and drives that student’s use. The results supported this possibility: LIWC’s clout, authentic, and drives dictionaries can be used as indicators of the student’s prevailing, confidence, motives and needs, we found that when calculated on written text, LIWC scores were connected with language dimension and psychology issues. These findings recommend that LIWC clout, authentic, motives, and needs dictionaries possibly will catch self-reported subjective emotion experience when applied to everyday activities. To enrich students’ confidence, power, motives, and needs toward the psychology issues faced by the students, students should be stimulated to discover the field of psychological processes and linguistic patterns by detecting their emotional and perceptual experiences.

Innovations in Text Analysis: Review of the International Conference MAXDAYS2

The article analyzes the conference held online by the German company VERBI Software GmbH - a developer of the well-known MAXQDA software designed for computerized content analysis of texts. Over the past decade Russian undergraduate and graduate students have been actively using this program in their research work as evidenced by a number of publications including in the "Istoricheskaya Informatika" journal. The author focuses on the dynamics of topics of the annual 2019-2021 conferences and characterizes the latest online conference. The author draws attention to the fact that the online status of the conference resulted in a sharp increase in the number of participants in comparison with previous traditional conferences. The novelty of this study is the fact that the author compares the reports of three conferences focusing on the range of methodological and methodical issues of computerized text analysis, constant update of research tools, new trends in the provision of educational materials to program users including open access to Maxqda Press publications. The article provides useful links to conference materials posted both on the developer's website and on YouTube. The author pays attention to those tools (platforms for online conferences) which are used by software developers to attract the maximum number of participants: free conference registration, user-friendly interface and international aware work (9 languages).

A conceptual framework to develop a patient-reported experience measure of the quality of mental health care: a qualitative study of the PREMIUM project in France

ABSTRACT Background: The objective of this study was to develop a conceptual framework to define a domain map describing the experience of patients with severe mental illnesses (SMIs) on the quality of mental health care. Methods: This study used an exploratory qualitative approach to examine the subjective experience of adult patients (18–65 years old) with SMIs, including schizophrenia (SZ), bipolar disorder (BD) and major depressive disorder (MDD). Participants were selected using a purposeful sampling method. Semistructured interviews were conducted with 37 psychiatric inpatients and outpatients recruited from the largest public hospital in southeastern France. Transcripts were subjected to an inductive analysis by using two complementary approaches (thematic analysis and computerized text analysis) to identify themes and subthemes. Results: Our analysis generated a conceptual model composed of 7 main themes, ranked from most important to least important as follows: interpersonal relationships, care environment, drug therapy, access and care coordination, respect and dignity, information and psychological care. The interpersonal relationships theme was divided into 3 subthemes: patient-staff relationships, relations with other patients and involvement of family and friends. All themes were spontaneously raised by respondents. Conclusion: This work provides a conceptual framework that will inform the subsequent development of a patient-reported experience measure to monitor and improve the performance of the mental health care system in France. The findings showed that patients with SMIs place an emphasis on the interpersonal component, which is one of the important predictors of therapeutic alliance. Trial registration: NCT02491866

SYSTEM OF AUTOMATIC DETERMINATION OF TEXT TONE

In the work, the methods of sentiment analysis of the text are considered. The system of emotion detection of Ukrainian-language texts based on dictionaries and rules is proposed. The developed software downloads text information in various formats and carries out tokenization and lemmatization procedures using the Python TokenizeUKand pymorphy2 libraries. As a result, an array of words in the basic grammatical form using for determining the tone of the text is formed. The obtained word base was analyzed using a tonal dictionary of the Ukrainian language. A dictionary of synonyms was used to expand the vocabulary. If there is no word in the tonal dictionary, the tonality value of its nearest synonym is used for further calculations. Computer analysis of textual information was performed in five emotional categories namely: very positive, positive, neutral, negative, and very negative tone of words. To increase the accuracy and validity of sentiment analysis, coefficients were used that take into account the various emotional load of words of different speech parts and their dissimilar impact on the overall assessment of the text tone. The proposed system of sentiment analysis assumes a greater emotional influence of adjectives compared to verbs and nouns. Since the tone of textual information and the expression of human emotions are subjective factors, the means of fuzzy modeling are used for sentiment analysis of texts. This approach makes it possible to take into account the contribution of all emotional categories in the final evaluation of the text. As a result of an aggregation of normalized data on different emotion categories and defuzzification by the method of the center of gravity for one-element sets, a quantitative estimate of the emotional tone of texts was obtained. The developed system of sentiment analysis was tested on Ukrainian-language texts from different sources and different emotional tones. The use of additional tools provides the analysis of more words in the text and the degree of their emotional tone, which leads to more correct detection of the tone of the whole text. Key words : computer text analysis, sentiment analysis, tokenization, tonality dictionary, fuzzy modeling.

Inquiring into the Corpus of Empire

This article tests the value of corpus linguistics in analyzing nineteenth-century commissions of inquiry into British colonies. It examines and improves the capacity of a computerized text analysis tool called the Linguistic Inquiry and Word Count to identify word meaning, sentiment, and psycholinguistic constructs in nineteenth-century sources. By augmenting its dictionary with nineteenth-century language and cross-checking meaning, we show that the software can code with 97% accuracy. We then demonstrate the tool's potential to explore genres of colonial writing, and to locate emotive language and language relating to power differentials in commission reports, a function we argue may provide a "way in" to assessing how commissioners treated different kinds of British subjects and their testimony in the reports.

Using computerized text analysis to examine associations between linguistic features and clients' distress during psychotherapy.

Raw linguistic data within psychotherapy sessions may provide important information about clients' progress and well-being. In the current study, computerized text analytic techniques were applied to examine whether linguistic features were associated with clients' experiences of distress within and between clients and whether changes in linguistic features were associated with changes in treatment outcome. Transcripts of 729 psychotherapy sessions from 58 clients treated by 52 therapists were analyzed. Prior to each session, clients reported their distress level. Linguistic features were extracted automatically by using natural language parser for first-person singular identification and using positive and negative emotion words lexicon. The association between linguistic features and levels of distress was examined using multilevel models. At the within-client level, fewer first-person singular words, fewer negative emotional words and more positive emotional words were associated with lower distress in the same session; and fewer negative emotion words were associated with lower next session distress (rather small f2 effect sizes = 0.011 < f2 < 0.022). At the between-client level, only first session use of positive emotion words was associated with first session distress (ηp2 effect size = 0.08). A drop in the use of first-person singular words was associated with improved outcome from pre- to posttreatment (small ηp2 effect size = 0.05). The findings provide preliminary support for the association between clients' linguistic features and their fluctuating experience of distress. They point to the potential value of computerized linguistic measures to track therapeutic outcomes. (PsycInfo Database Record (c) 2021 APA, all rights reserved).