Applying LDA Topic Modeling in Communication Research: Toward a Valid and Reliable Methodology

ABSTRACT

Latent Dirichlet allocation (LDA) topic models are increasingly being used in communication research. Yet, questions regarding reliability and validity of the approach have received little attention thus far. In applying LDA to textual data, researchers need to tackle at least four major challenges that affect these criteria: (a) appropriate pre-processing of the text collection; (b) adequate selection of model parameters, including the number of topics to be generated; (c) evaluation of the model’s reliability; and (d) the process of validly interpreting the resulting topics. We review the research literature dealing with these questions and propose a methodology that approaches these challenges. Our overall goal is to make LDA topic modeling more accessible to communication researchers and to ensure compliance with disciplinary standards. Consequently, we develop a brief hands-on user guide for applying LDA topic modeling. We demonstrate the value of our approach with empirical data from an ongoing research project.     

Citizen-Centered Journalism and Contested Boundaries

This article explores the existence of a “citizen-centered journalism” that sees citizens as crucial participants in the construction of news and as co-creators of their own worlds. Through qualitative case studies of three news organizations, the article examines the motivation for using a citizen-centered approach, the news routines that are required to do so, the categories of content produced, and the perceived impact of this approach. The results suggest these news organizations are working in partnership with communities and striving to give a voice to historically marginalized communities. The journalists, however, see citizen participation as complementary to professional journalistic routines that favor verifiable information, rather than assigning inherent value to it for its own sake.     

ACADEMIC SELF‐EFFICACY AMONG JUNIOR HIGH SCHOOL STUDENTS IN GHANA: EVALUATING FACTOR STRUCTURE AND MEASUREMENT INVARIANCE ACROSS GENDER

Self‐efficacy is a universal construct, but few validated measures exist for researchers in developing countries to use in assessing youths’ perceptions of their ability to achieve academic success. This study examined the cross‐cultural suitability and psychometric properties of an academic self‐efficacy scale (ASES) adapted for the Ghanaian context. ASES construct validity was assessed with a sample of 4,289 Ghanaian junior high‐school students and exploratory and confirmatory factor analysis. Invariance testing assessed the scale's measurement equivalence by gender and temporal stability of gender equivalence. The ASES is a valid, reliable one‐dimensional scale for assessing young Ghanaians’ perceptions of their academic capabilities, and it works equally well across genders. As adapted, ASES is a valid scale with utility for researchers examining predictors and effects of academic self‐efficacy. The ASES has important implications for decisions regarding investment in programs aimed at improving academic self‐efficacy of youth, both in sub‐Saharan Africa and the increasingly diverse American public schools.     

Microfluidic cell fragmentation for mechanical phenotyping of cancer cells

Circulating tumor cells (CTCs) shed from the primary tumor undergo significant fragmentation in the microvasculature, and very few escape to instigate metastases. Inspired by this in vivo behavior of CTCs, we report a microfluidic method to phenotype cancer cells based on their ability to arrest and fragment at a micropillar-based bifurcation. We find that in addition to cancer cell size, mechanical properties determine fragmentability. We observe that highly metastatic prostate cancer cells are more resistant to fragmentation than weakly metastatic cells, providing the first indication that metastatic CTCs can escape rupture and potentially initiate secondary tumors. Our method may thus be useful in identifying phenotypes that succumb to or escape mechanical trauma in microcirculation.     

Childhood Physical Fitness and Performance as Predictors of High School Sport Participation

ABSTRACT

Participation in youth sport is positively associated with physical fitness and performance. The purpose of the current study was to examine if physical fitness measures during childhood and early adolescence predicted high school sport participation. Participants included youth in the Michigan State University Motor Performance Study. Measures consisted of the endurance shuttle run, Wells sit-and-reach, jump and reach, standing long jump, agility shuttle run and 30-yard dash. Individual, generalized estimating equation models were used to estimate the association between each measure and the odds of participating in high school sport. For both sexes, the endurance shuttle run and agility shuttle run were related to sport participation. For boys only, the long jump and 30-yard dash were also associated with sport participation. Better physical fitness and performance across several dimensions in childhood, including those related to motor skills, enhanced the likelihood of sport participation in high school.     

Analysis of single comments left for bioRxiv preprints till September 2019

IntroductionWhile early commenting on studies is seen as one of the advantages of preprints, the type of such comments, and the people who post them, have not been systematically explored.Materials and methodsWe analysed comments posted between 21 May 2015 and 9 September 2019 for 1983 bioRxiv preprints that received only one comment on the bioRxiv website. The comment types were classified by three coders independently, with all differences resolved by consensus.ResultsOur analysis showed that 69% of comments were posted by non-authors (N = 1366), and 31% by the preprints’ authors themselves (N = 617). Twelve percent of non-author comments (N = 168) were full review reports traditionally found during journal review, while the rest most commonly contained praises (N = 577, 42%), suggestions (N = 399, 29%), or criticisms (N = 226, 17%). Authors’ comments most commonly contained publication status updates (N = 354, 57%), additional study information (N = 158, 26%), or solicited feedback for the preprints (N = 65, 11%).ConclusionsOur results indicate that comments posted for bioRxiv preprints may have potential benefits for both the public and the scholarly community. Further research is needed to measure the direct impact of these comments on comments made by journal peer reviewers, subsequent preprint versions or journal publications.     

EFFECTS OF TEACHER–STUDENT RELATIONSHIPS ON PEER HARASSMENT: A MULTILEVEL STUDY

Peer harassment is a major social problem affecting children and adolescents internationally. Much research has focused on student‐to‐student harassment from either an individual or a multilevel perspective. There is a paucity of multilevel research on students’ relationships with the classroom teacher. The purpose of this study was to use a socioecological perspective to examine the relationships between individual student‐level characteristics, problematic teacher–student relationships, and student‐reported peer harassment. A total of 1,864 children (50.7% female) aged 8 to 13 years (M = 9.82, SD = 1.24), nested in 27 schools (58.2% public) in Spain, participated in the study. Ninety‐four homeroom teachers reported on teacher–student relationships, and students completed self‐report measures related to peer harassment and teacher–student relationships. Multilevel models showed that relationships between students and teachers exerted a varying degree of influence on classroom levels of peer harassment. Specifically, student‐reported teacher support was associated with diminished student‐reported peer victimization, whereas direct and indirect student‐reported teacher‐to‐student aggression was associated with increased peer victimization. Additionally, student‐reported student‐to‐teacher aggression and teacher‐to‐student aggression were associated with increased student‐reported peer aggression. Teacher‐reported variables at the classroom level, however, contributed little to student‐reported outcomes. Results are discussed in the context of future research on relationships between teachers and students.