Disentangling User Samples: A Supervised Machine Learning Approach to Proxy-population Mismatch in Twitter Research

ABSTRACT

This study addresses the issue of sampling biases in social media data-driven communication research. The authors demonstrate how supervised machine learning could reduce Twitter sampling bias induced from “proxy-population mismatch”. Particularly, this study used the Random Forest (RF) classifier to disentangle tweet samples representative of general publics’ activities from non-general—or institutional—activities. By applying RF classifier models to Twitter data sets relevant to four news events and a randomly pooled dataset, the study finds systematic differences between general user samples and institutional user samples in their messaging patterns. This article calls for disentangling Twitter user samples when ordinary user behaviors are the focus of research. It also builds on the development of machine learning modeling in the context of communication research.     

Personality structure of brands on social networking sites and its effects on brand affect and trust: evidence of brand anthropomorphization

The current study explored the personality dimensions of brands on Social Networking Sites (SNSs) using a South Korean sample. Through a series of exploratory and confirmatory factor analyses, six dimensions (i.e. Openness, Conscientiousness, Agreeableness, Extraversion, Warmth, and Contentiousness) were yielded for brands with SNS presence. The data also produced six dimensions (i.e. Openness, Sincerity, Extraversion, Competence, Agreeableness, and Ruggedness) for brands without SNS presence. The results show that brands with which consumers interact on SNSs have more human-like personalities than brands absent from SNSs, and the personality dimensions affected brand affect and trust to varying degrees.     

Effects of different types of instructor comments in online discussions

The current study examined the effects of different types of instructor comments on student engagement in an online discussion. In particular, this study examined three comment types: (1) praise-oriented comments, agreeing with students’ initial messages and recapping their ideas, (2) elaboration-encouraging comments, requesting elaboration on the initial messages, and (3) perspective-widening comments, suggesting different or thought-provoking opinions responding to the initial messages. Seventeen graduate students enrolled in an online course were recruited. A content analysis was employed to evaluate the levels of knowledge construction. Interactivity was measured by the number of posts and idea units. The findings revealed that perspective-widening comments facilitated students’ evaluation on the peer postings and brought up diverse perspectives, contributing to knowledge construction. While elaboration-encouraging comments led to high interactivity among students, praise-oriented comments had no significant effect on the frequency of interactivity.     

Effects of discussion representation: comparisons between social and cognitive diagrams

An online discussion facilitates students’ higher order thinking in online classes, especially when adopted with the instructor’s guidance. The current experimental study examined the effects of two different discussion representation tools (social and cognitive diagrams) on students’ discussion behaviors. The social diagram emphasized interactivity of participants by illustrating who posted messages and who replied. The cognitive diagram described how discussion topic evolved by summarizing main topics discussed. Thirteen graduate students enrolled in an online graduate course participated in the study. While analyzing five instructional cases, students were asked to discuss each case in a group that the instructor assigned randomly. For each case, one group was provided with the social diagram whereas the other with the cognitive diagram. Major findings revealed both tools facilitated online discussion activities as the instructor intended: the social diagram turned out to promote socially desirable responses while the cognitive diagram produced more cognitively desirable responses. Further studies on how the two types of discussion diagrams can be integrated in online discussions will be needed.     

Development of an instrument for measuring cognitive conflict in secondary‐level science classes

Based on conceptual change theory, cognitive conflict is known as an important factor in conceptual change even though there are still questions about its positive and negative effects on science learning. However, there is no reliable method by which to assess the cognitive conflict students experience in their learning. The purpose of this research was to develop an instrument for measuring secondary students' cognitive conflict levels as they learned science. The results of this study indicate that our instrument is a valid and reliable tool for measuring cognitive conflict levels. Factor analysis supported the model that cognitive conflict consists of four constructs: recognition of an anomalous situation, interest, anxiety, and cognitive reappraisal of the conflict situation. Implications for instruction and possibilities for future research are discussed. © 2003 Wiley Periodicals, Inc. J Res Sci Teach 40: 585–603, 2003     

Does high-quality professional development make a difference? Evidence from TIMSS

Abstract

While professional development has been emphasised in many countries, few comparative studies have examined the degree of teachers’ participation in professional development and its effect on teachers. Using the TIMSS 2007 data, this study analyses to what extent teachers participated in professional development in 50 TIMSS countries and examines the relationship among high-quality teachers’ professional development, teachers’ beliefs, attitudes and job satisfaction. The study found that the extent to which teachers have participated in professional developmental activities was surprisingly low. In particular, teachers in 47 countries have rarely participated in active learning. The study also found that high-quality teacher professional development was significantly associated with greater job satisfaction, higher expectations for student performance and more positive attitudes in teachers concerning their implementing the school’s curriculum. Policy implications for professional development are discussed.     

Roles of Abductive Reasoning and Prior Belief in Children’s Generation of Hypotheses about Pendulum Motion

The purpose of the present study was to test the hypothesis that student’s abductive reasoning skills play an important role in the generation of hypotheses on pendulum motion tasks. To test the hypothesis, a hypothesis-generating test on pendulum motion, and a prior-belief test about pendulum motion were developed and administered to a sample of 5th grade children. A significant number of subjects who have prior belief about the length to alter pendulum motion failed to apply their prior belief to generate a hypothesis on a swing task. These results suggest that students’ failure in hypothesis generation was related to abductive reasoning ability, rather than simple lack of prior belief. This study, then, supports the notion that abductive reasoning ability beyond prior belief plays an important role in the process of hypothesis generation. This study suggests that science education should provide teaching about abductive reasoning as well as scientific declarative knowledge for developing children’s hypothesis-generation skills.     

Responses to the Negative Emotions of Others by Autistic, Mentally Retarded, and Normal Children

Attention, facial affect, and behavioral responses to adults showing distress, fear, and discomfort were compared for autistic, mentally retarded, and normal children. The normal and mentally retarded children were very attentive to adults in all 3 situations. In contrast, many of the autistic children appeared to ignore or not notice the adults showing these negative affects. As a group, the autistic children looked at the adults less and were much more engaged in toy play than the other children during periods when an adult pretended to be hurt. The autistic children were also less attentive to adults showing fear, although their behavior was not different from the normal children. Few of the children in any group showed much facial affect in response to these situations. The results are discussed in terms of the importance of affect in the social learning experiences of the young child.     

“This Is How You Hetero:” Sexual Minorities in Heteronormative Sex Education

The efficacy of sex education has been questioned, as students participate in high rates of unsafe sex after completion. Without exploring various sexual minority (SM) identities (e.g., gay, lesbian, and bisexual) and forms of sex, sex education may be especially unhelpful for SMs by perpetuating the heteronormative (i.e., assuming heterosexuality is normal) environment they typically experience. In the present qualitative study, semistructured interviews were conducted with 12 SMs regarding their experiences with sex education using a grounded theory approach. Participants described sex education as being heteronormative and exclusive of their identities, making them feel invisible, sexually unprepared, and shameful. Sex education also reportedly contributed to sexual hesitance, sexual violence, and risky sexual behaviors. Participants endorsed histories of depression, anxiety, and suicidality, often associated with their identity. To become more informed and sexually prepared, they sought information through conversations with others, online searching, college courses, and trial and error sexual experiences. Participants also advocated for inclusive sex education, which would incorporate all identities and associated safe sex practices. They concluded that inclusivity in curricula could lead to various improved outcomes for SMs, such as safe sex, a sense of community, identity confidence, healthy relationships, and resilience.