ABSTRACT The availability of individual-level digital trace data offers exciting new ways to study media uses and effects based on the actual content that people encountered. In this article, we argue that to really reap the benefits of this data, we need to update our methodology for automated text analysis. We review challenges for the automatic identification of theoretically relevant concepts in texts along three dimensions: format/style, language, and modality. These dimensions unveil a significantly higher level of diversity and complexity in individual-level digital trace data, as opposed to the content traditionally examined through automated text analysis in our field. Consequently, they provide a valuable perspective for exploring the limitations of traditional approaches. We argue that recent developments within the field of Natural Language Processing, in particular, transfer learning using transformer-based models, have the potential to aid the development, application, and performance of various computational tools. These tools can contribute to the meaningful categorization of the content of social (and other) media.

ABSTRACT Linkage analysis, i.e. linking media exposure, content, and surveys, has been a powerful tool to assess media effects. However, the development of online communication and the advent of social media brings about many challenges for traditional linkage designs. In this paper, we explain the three steps of linkage designs for online communication effects and the usage of computational approaches to capture communication exposure and content. We then review recent designs and studies that use different forms of digital trace data to link digital communication exposure, content, and surveys: Tracking data, data donations, and screenshots/screen recordings. We describe (practical) challenges and opportunities when linking digital communication traces with self-reports and show how these data could be analyzed to establish different media effects.

ABSTRACT The need for privacy has been gaining importance in various disciplines and areas of communication, including computer-mediated, interpersonal, and health communication. These disciplines require reliable measurement of the need for privacy across different contexts. We propose a theoretical concept of the need for privacy as a personality trait and develop a multi-dimensional scale. In Study 1, we developed and tested the Need for Privacy Scale (NFP-S) using exploratory and confirmatory factor analysis (German internet users, n 1 = 3,278; n 2 = 1,226). The results support a second-order model with three first-order factors, i.e., informational, psychological, and physical need for privacy; social need for privacy was identified to be inherent to the three other dimensions and as such not a dimension on its own. The 12-item scale was validated with regard to loneliness, online privacy concerns, sex, online data protection and online information disclosure. In Study 2, we confirmed the factorial structure with a representative sample of N = 1,000 German internet users. We validated the scale with a self-assessment of the need for privacy, offline and online privacy concerns, online privacy literacy, social media usage, online data protection and social media self-disclosure, and sociodemographics.

ABSTRACT Women of Color (WoC) in science, technology, engineering, and math (STEM) leave doctoral programs at disproportionately high rates. Supportive mentorship is key to increasing belonging and rates of retention, yet little is known about how conversations between mentees and their mentors on academic and personal stress topics unfold in real-time. Applying the lens of Social Cognitive Career Theory to communication dynamics between mentees and mentors, the present study utilized a dynamic dyadic systems (DDS) perspective to examine observationally coded data from six mentee-mentor dyads. First, hierarchical clustering analysis was applied to identify speaking turn types. Then, sequence analysis was used to identify common multi-turn patterns or conversation motifs (CM). Results showed five predominant CMs: (CM1) support provision through listening; (CM2) focus on mentor’s experience; (CM3) support provision through advice; (CM4) mentee’s making a bid for support; and (CM5) mentor dominated conversations. This study demonstrates methods for identifying potentially meaningful patterns of support in stress conversations between mentees and mentors. The application of such methods with larger samples may aid in understanding ways to increase retention among WoC in STEM through mentor support provision.

ABSTRACT Conversations between people are where, among other things, stressors are amplified and attenuated, conflicts are entrenched and resolved, and goals are advanced and thwarted. What happens in dyads’ back-and-forth exchanges to produce such consequential and varied outcomes? Although numerous theories in communication and in social psychology address this question, empirical tests of these theories often operationalize conversational behavior using either discrete messages or overall features of the conversation. Dynamic systems theories and methods provide opportunities to examine the interdependency, self-stabilization, and self-organization processes that manifest in conversations over time. The dynamic dyadic systems perspective exemplified by the articles in this special issue (a) focuses inquiry on the turn-to-turn, asynchronous exchange of messages between two partners, (b) emphasizes behavioral patterns within and the structural and temporal organization of conversations, and (c) adapts techniques used in analysis of intensive longitudinal data to identify and operationalize those dynamic patterns. As an introduction to the special issue, this paper describes a dynamic dyadic systems perspective on conversation and discusses directions for future research, such as applications to human-computer interaction, family communication patterns, health care interventions, and group deliberation.

ABSTRACT The present research examined how value-free and value-driven measures of argument strength (MAS) can be computationally extracted using a theory-driven approach at scale in a naturalistic setting by analyzing a total of 7,961 real-world debates and 42,716 judgments in rhetorical quality. In the first study, value-free MAS was significantly related to the rhetorical quality of arguments (i.e. their persuasiveness). The results indicate that the side that provides more information-source citation, less quantitative specificity, more unique words, and more abstract language is more likely to be perceived as convincing in dialectical argumentation, where two people are exchanging opposing arguments. In the second study, the added influence of value-driven MAS is investigated. The results show that the similarity between the moral values represented in arguments and those that are salient to argument receivers predicts the rhetorical quality. The research demonstrates how rhetorical quality can be measured and predicted at scale, and how naturally generated arguments can be used for scientific progress in persuasion research.

ABSTRACT The empirical identification of frames drawing on automated text analysis has been discussed intensely with regard to the validity of measurements. Adding to an evolving discussion on automated frame identification, we systematically contrast different machine-learning approaches with a manually coded gold standard to shed light on the implications of using one or the other: (1) topic modeling, (2) keyword-assisted topic modeling (keyATM), and (3) supervised machine learning as three popular and/or promising approaches. Manual coding is based on the Policy Frames codebook, providing an established base that allows future research to dovetail our contribution. Analysing a large dataset of 12 Austrian newspapers’ EU coverage over 11 years (2009–2019), we contribute to addressing the methodological challenges that have emerged for social scientists interested in employing automated tools for frame analysis. While results confirm the superiority of supervised machine-learning, the semi-supervised approach (keyATM) seems unfit for frame analysis, whereas the topic model covers the middle ground. Results are extensively discussed regarding their implications for the validity of approaches.

ABSTRACT Understanding the implications of aggressive political debate style amid corrosive modes of campaign politics requires fine-grained analyses of political performance, attending to multiple communication modalities. Politicians’ facial expressions, emotional tone, and speech content can all independently convey aggression and dominance, and often work in combination for purposes of emphasis. Yet micro-coding individual visual, tonal, and verbal features across more than a handful of debate segments becomes extremely labor intensive, hampering research, especially historical, longitudinal, and cross-cultural work. To address this limitation, we develop a novel multimodal classifier using an Interaction Canonical Correlation Network (ICCN) that incorporates video and audio features with speech coding of candidate debate performance, trained on a 20% sample of 10-second segments from each of the first televised U.S presidential debates between 1980 and 2020. In the analysis, we demonstrate this classifier can accurately detect aggression by political candidates in U.S. debates. We sharpen its performance by distinguishing between debate eras characterized by lower and higher levels of aggression and validate the approach by comparing the performance of unimodal with multimodal classification. This classifier opens new avenues for computational social science research, including explaining candidate behavior within debates at a larger scale and across different eras.

ABSTRACT Informed by the dynamic dyadic systems perspective on analyzing conversational data, this study uses sequence analysis to illuminate turn patterns that characterize parents’ emotion coaching/dismissing communication and adolescents’ emotion regulation/dysregulation in conversations about an emotionally evocative event. This study analyzed two conversations from 60 parent-adolescent dyads, 30 with harmful parental alcohol use and 30 without, where the adolescent described recent events that elicited positive and negative feelings. Using configural frequency analysis, we identified turn sequences that were over- and under-represented in the data, and we compared the prevalence of different turn sequences between families with harmful and non-harmful alcohol use. Then, using sequence analysis, we identified conversational motifs that reflect different patterns of parental communication and adolescent emotion regulation in conversation. Three conversational motifs emerged in conversations about negative emotions: (a) information sharing, (b) emotion regulating, and (c) reactive. Conversations about positive emotions revealed two conversational motifs: (a) emotion regulating and (b) information sharing. In addition, the conversational motifs were examined as predictors of adolescents’ post- interaction appraisals of their own emotion regulation and their parent’s responsiveness and control during the conversation. Findings point to the utility of sequence analysis for documenting patterns of interaction in parent-adolescent emotion-focused conversations.

ABSTRACT Dispositional communication competencies can be assessed in (a) a generic form that does not include any reference to a particular medium of interaction or in (b) a communication medium-specific version. To date, little is known about the specific media that individuals use as a reference and the weights they assign to them when responding to generic communication items – an important research gap because the use of diverse communication media has risen considerably during the COVID-19 pandemic. Based on media theories, two hypotheses were derived: Generic ratings contain a “hidden” face-to-face (FtF) communication framing that is dominant in the cognitive processing (media naturalness perspective) versus media are equally weighted in the mental aggregate of respondents (adaptation perspective). According to a preregistered study plan, generic and medium-specific communication items were assessed to investigate these hypotheses (referencing FtF, videoconferencing, chat, and e-mail interaction contexts). Training (n = 200) and test (n = 389) datasets were analyzed using latent variable modeling. Results indicated that generic ratings have a strong hidden FtF framing. These hidden framings impact the predictive power of the competencies to explain communication criteria (i.e. communication satisfaction). Exploratory analyses indicated that individual differences in media experience may affect the framings.