Search

Grit, the passion for achieving long-term goals, has been conceived as a two-dimensional construct (Consistency of interest and Perseverance of effort). The construct is well known for its easy measurement and its relationship with performance, including academic performance. However, there have been different criticisms, such as the overlap of grit with other personality characteristics, the variability in the psychometric structure of the Grit-O and Grit-S tests, and the fact that some work reports a weak influence of grit on academic performance. Within this framework, this study contrasts different psychometric structures of the Grit-S scale and its relationship with the academic performance of higher education students. So, with this purpose, a dichotomous model of high and low grit, a K-medias clustering model, and three structural equation models have been tested. The results indicate that (a) there is a statistically significant relationship between grit--mainly determined by the consistency dimension--and academic performance, although it decreases when controlling for contextual variables, and (b) Consistency mediates the relationship between Perseverance and academic performance. The instability of grit to predict performance is discussed, and the thesis of a two-sub-dimensional structure is supported.

Data transfer between isolated clusters is imperative for cybersecurity education, research, and testing. Such techniques facilitate hands-on cybersecurity learning in isolated clusters, allow cybersecurity students to practice with various hacking tools, and develop professional cybersecurity technical skills. Educators often use these remote learning environments for research as well. Researchers and students use these isolated environments to test sophisticated hardware, software, and procedures using full-fledged operating systems, networks, and applications. Virus and malware researchers may wish to release suspected malicious software in a controlled environment to observe their behavior better or gain the information needed to assist their reverse engineering processes. The isolation prevents harm to networked systems. However, there are times when the data is required to move in such quantities or speeds that it makes downloading onto an intermediate device untenable. This study proposes a novel turnstile model, a mechanism for one-way file transfer from one enterprise system to another without allowing data leakage. This system protects data integrity and security by connecting the isolated environment to the external network via a locked-down interconnection. Using medium-security isolated clusters, the researchers successfully developed a unidirectional file transfer system that acts as a one-way "turnstile" for secure file transfer between systems not connected to the internet or other external networks. The Turnstile system (source code available at github.com/monnin/turnstile) provides unidirectional file transfer between two computer systems. The solution enabled data to be transferred from a source system to a destination system without allowing any data to be transferred back in the opposite direction. The researchers found an automated process of transferring external files to isolated clusters optimized the transfer speed of external files to isolated clusters using Linux distributions and commands.

Autism is heterogeneous, which complicates providing tailored support and future prospects. We aim to identify subgroups in autistic adults with average to high intelligence, to clarify if certain subgroups might need support. We included 14 questionnaire variables related to aging and/or autism (e.g., demographic, psychological, and lifestyle). Community detection analysis was used for subgroup identification in an original sample of 114 autistic adults with an adulthood diagnosis (autism) and 58 non-autistic adults as comparison group (COMP), and a replication sample (N[subscript Autism] = 261; N[subscript COMP] = 287), both aged 30-89 years. Next, we identified subgroups and assessed external validity (for cognitive and psychological difficulties, and quality of life [QoL]) in the autism samples. To test specificity, we repeated the analysis after adding 123 adults with ADHD, aged 30-80 years. As expected, the autism and COMP groups formed distinct subgroups. Among autistic adults, we identified three subgroups of which two were replicated. One of these subgroups seemed most vulnerable on the cluster variables; this subgroup also reported the most cognitive and psychological difficulties, and lowest QoL. Adding the ADHD group did not alter results. Within autistic adults, one subgroup could especially benefit from support and specialized care, although this must be tested in future studies.

Automated image analysis has received increasing attention in social scientific research, yet existing scholarship has mostly covered the application of supervised learning to classify images into predefined categories. This study focuses on the task of unsupervised image clustering, which aims to automatically discover categories from unlabelled image data. We first review the steps to perform image clustering and then focus on one key challenge in this task--finding intermediate representations of images. We present several methods of extracting intermediate image representations, including the bag-of-visual-words model, self-supervised learning, and transfer learning (in particular, feature extraction with pretrained models). We compare these methods using various visual datasets, including images related to protests in China from Weibo, images about climate change on Instagram, and profile images of the Russian Internet Research Agency on Twitter. In addition, we propose a systematic way to interpret and validate clustering solutions. Results show that transfer learning significantly outperforms the other methods. The dataset used in the pretrained model critically determines what categories the algorithms can discover.

This study aimed to reveal the trend of research on e-assessment in the field of educational sciences through scientific mapping and bibliometric analyses. For this purpose, the numerical distribution of research on e-assessment, citation analysis, research themes and the change of trend topics were examined. The publications to be examined were selected from WoS database according to PRISMA model, and 911 studies were included in the analysis. VOSviewer, Biblioshiny, Smart Bibliometrics and Leximancer software were used in data analysis. Apparently, there has been a significant increase in the number of research since 2005, and publications have been mostly produced in form of articles and papers. The most cited and the most productive countries are the USA, the United Kingdom and Australia, while the most cited journals are Assessment & Evaluation in Higher Education and British Journal of Educational Technology. An analysis of the keyword map revealed that the themes of technology and motivation, blended learning and collaboration, interaction and innovative approaches, validity and reliability, higher education, quality, basic disciplines and COVID-19 were frequently emphasized in the studies on e-assessment. An analysis of trend topics by years showed that, between 2010 and 2021, the trend topic distribution changed to include topics such as COVID-19, academic integrity, engagement, cheating, case study, and higher education. All these findings reveal that e-evaluation activities have displayed a development and transformation over time with the effect of developing technology, the pandemic, the spread of e-learning, the expansion of communication opportunities and many other factors.

Currently, the majority of e-learning lessons created and disseminated advocate a "one-size-fits-all" teaching philosophy. The e-learning environment, however, includes slow learners in a noticeable way, just like in traditional classroom settings. Learning analytics of educational data from a learning management system (LMS) have been considered by the researchers as a potential means of identifying slow e-learners and supporting, contesting, and altering present educational practices in e-learning. We used the students' rates of learning and grade points along with the total learning time, which is calculated from the time series log data, to cluster the learners. The rate at which a student learns determines whether he or she is a slow learner, an average learner, or a gifted learner. For classifying learners, we followed a step-by-step procedure that included instructional design to create a dataset, learning analytics of the dataset, and a machine learning strategy to cluster e-learners. The system has been adequately integrated with the methods for measuring student learning. A strategy based on the revised Bloom's Taxonomy is offered for the assessment of learners. The K-Means clustering approach is used to group learners who have similar performance without collecting a learner's previous academic records or demographic information. In the experimental evaluation, 7.7% of e-learners are grouped as slow learners, while advanced learners make up 61.3 percent of the student body and average learners make up 31 percent. According to the study, there is a correlation between learning rate and academic success, with fast learners having a lower learning rate.

The level of information technology is increasing, and technology is developed. University English teaching has also changed under its influence. Different from the traditional teaching in the past, more and more students adopt the mode of "Internet + Smartphone" to learn English. This paper proposes a teaching mode evaluation method in the context of big data. Through the K-means algorithm based on it, the data clustering of business English teaching mode is completed. According to the data obtained from clustering, combined with the standardization of positive and negative indicators, the business English teaching mode system is constructed. Experiments show that when the clustering K value is 16, the clustering accuracy of this method is 99.2%, and it can effectively get the index weights, and the evaluation time is short, which is better than the other two comparison methods.

This study presents the Academic Investment Model (AIM) as a novel approach to predicting student academic performance by incorporating learning styles as a predictive feature. Utilizing data from 138 Marketing students across China, the research employs a combination of machine learning clustering methods and manual feature engineering through a four-quadrant clustering technique. The AIM model delineates student investment into four quadrants based on their time and energy commitment to academic pursuits, distinguishing between result-oriented and process-oriented investments. The findings reveal that the four-quadrant method surpasses machine learning clustering in predictive accuracy, highlighting the robustness of manual feature engineering. The study's significance lies in its potential to guide educators in designing targeted interventions and personalized learning strategies, emphasizing the importance of process-oriented assessment in education. Future research is recommended to expand the sample size and explore the integration of deep learning models for validation.

In education settings, treatments are often non-randomly assigned to clusters, such as schools or classrooms, while outcomes are measured for students. This research design is called the clustered observational study (COS). We examine the consequences of common support violations in the COS context. Common support violations occur when the covariate distributions of treated and control units have substantial areas of non-overlap. Such violations are likely to occur in a COS, especially with a small number of treated clusters. We provide a comprehensive review of methods for overlap violations in the context of COS designs. We provide an overview of diagnostic tests and trimming methods to ensure overlap holds for the distributions of treated and control covariates. We then outline how trimming changes the estimand and how profiling can be used to understand the causal quantity for which overlap holds. Finally, we demonstrate how steps to achieve adequate overlap can result in very narrowly defined causal effects that may have little policy relevance. We use data on Catholic schools to illustrate concepts throughout.

Team project-based learning has become increasingly common in higher education. This study aimed to characterise and understand students' team learning experiences in team project-based learning by considering various aspects, such as individual qualities, teamwork, task, and instructor support. K-means clustering analysis was performed using eleven variables: team learning preference, co-regulation, team efficacy, team climate, team trust, psychological safety, team interaction, perceived team loafing, task authenticity, the frequency of instructor feedback, and usefulness of instructor feedback. Data were collected from students who were involved in team project-based learning at a South Korean university in the 2nd semester of 2021, and 782 responses were used in the analysis. Consequently, six clusters were identified, each showing different patterns of individual team learning attitudes and skills, teamwork, task authenticity, and instructor feedback. Following the cluster analysis, one-way ANOVA was performed, finding significant differences in team performance and team learning satisfaction between clusters. Based on these findings, important factors in team learning are discussed, and implications of enhancing team project-based learning are presented.

The imbalance in student-teacher ratio and the diversity of student population pose challenges to MOOC's quality of instructor support. An understanding of student profiles, such as who they are and how they behave, is critical to improving personalized support of MOOC learning environments. While past studies have explored different types of student profiles, few have been done to investigate which student profiles lead to successful performance and what behavior patterns are exhibited by successful and unsuccessful performance groups. To address this research gap, we employed both bottom-up and top-down strategies, to gain useful insights into student learning in the context of MOOCs. From learning behavior records of 26,862 students in six MOOCs, we identified and validated three behavior attributes: effort regulation, self-assessment, and learner participation. Our results revealed that effort regulation emerged as the foremost important factor that positively contributes to students' academic performance in MOOCs. Particularly, online persistence was the strongest positive predictor impacting student success. Based on the behavior attributes ascertained, we demonstrated five student sub-profiles with different behavior patterns: Persistence Achievers and Social Collaborators in the successful group; Dabblers, Disengagers, and Slackers in the unsuccessful group. Our analysis revealed that successful performers engaged with the course in quite different ways. We also investigated how effort regulation differed significantly between successful and unsuccessful performers. Unexpectedly, we also noticed that Persistence Achievers, despite their success, exhibited a high degree of procrastination. This work offers novel insights into instructional interventions for supporting MOOC learning.

The purpose of this study was to examine clusters of less-skilled college readers. College students with below average reading comprehension skills (N = 77) read and thought aloud about four texts, recalled the texts, and completed standardized assessments of reading skills. Based on the findings of cluster analyses of the cognitive processes relevant to creating coherent mental representations of the texts, two clusters of readers were noted. One cluster had more paraphrases and inaccurate or irrelevant connections to background knowledge (invalid inferences) as well as fewer recalled ideas from the texts than the second cluster, which had more accurate and relevant connections to background knowledge (valid inferences). The findings inform developing and personalizing reading interventions for college students.

Analysis of the differential treatment effects across targeted subgroups and contexts is a critical objective in many evaluations because it delineates for whom and under what conditions particular programs, therapies or treatments are effective. Unfortunately, it is unclear how to plan efficient and effective evaluations that include these moderated effects when the design includes partial nesting (i.e., disparate grouping structures across treatment conditions). In this study, we develop statistical power formulas to identify requisite sample sizes and guide the planning of evaluations probing moderation under two-level partially nested designs. The results suggest that the power to detect moderation effects in partially nested designs is substantially influenced by sample size, moderation effect size, and moderator variance structure (i.e., varies within groups only or within and between groups). We implement the power formulas in the R-Shiny application PowerUpRShiny and demonstrate their use to plan evaluations.

We used the timing of serial recall in several situations to reveal important aspects of recall groupings that participants construct and the reasons those groupings occur. We examined the timing of responses in the recall of digit strings within two published experiments. Cowan, Saults, Elliott, and Moreno (2002) examined memory for nine-item lists in a way that deconfounded the presentation modality, input versus output serial position (using a varied starting point of recall), memory load from items not yet recalled (using whole vs. partial recall), and the presence or absence of the temporal grouping of the lists into triads. Accuracy was strikingly different in the two modalities, with grouping drastically changing recall of acoustic lists but with little difference between grouped versus ungrouped visual lists. Despite the modality difference in recall, the present timing results show that participants imposed a similar grouped structure in recall for both modalities. When errors were made, the original grouping structure still was maintained in the output timing. Cowan et al. (2005, Experiment 1) presented ungrouped acoustic lists of varying lengths in a span task. In this case, there was never a grouping structure presented to the participant and the accuracy data did not reveal evidence of grouping, but we show clear grouping in the timing structure. This grouping structure varied with both the list length and an individual's span. Timing in both experiments shows that a grouping structure is imposed on responses in a manner that is dependent on several factors not present in the stimuli.

This study examined Korean learners' production of intervocalic English biconsonantal clusters consisting of /p/, /t/, /k/, /m/, /n/, or /[eng] / followed by /l/ or /[voiced alveolar approximant]/. The results of the production experiment showed several factors influencing Korean learners' production. First, Korean learners' production was influenced by their L1 sound inventory. The mispronunciation of a liquid (i.e., /l/ as [[voiced alveolar approximant]] and /[voiced alveolar approximant] / as [l]) was observed and this will be due to the fact that [l] and [[voiced alveolar approximant]] are not contrastive in Korean. Second, L1 phonological rules affecting a post-consonantal onset liquid influenced Korean learners' production of English consonant clusters containing a liquid. The cluster /tl/ was mostly pronounced as [ll] and /nl/ as [ll] or [nn]. In addition, the nasalization of a stop consonant before a liquid was attested. Next, Korean learners' production also exhibited the influence of phonetic similarity between the two members of a cluster by illustrating relatively lower production accuracy for /tl/, /t[voiced alveolar approximant]/, and /nl/ consisting of two alveolar consonants. Finally, familiarity with the test words seemed to be a factor influencing Korean learners' production. Korean learners illustrated relatively lower production accuracy for the words "outlet" and "online" which are frequently used as loanwords in Korea.

Purpose: Nonnative consonant cluster learning has become a useful experimental approach for learning about speech motor learning, and we sought to enhance our understanding of this area and to establish best practices for this type of research. Method: One hundred twenty individuals completed a nonnative consonant cluster learning task within a speech motor learning paradigm. Following a brief prepractice, participants then practiced the production of eight word-initial nonnative consonant clusters embedded in bisyllabic nonwords (e.g., GD in /gdivu/). The clusters ranged in difficulty according to linguistic typology and sonority sequencing. Acquisition was operationalized as the change across the practice section and learning was assessed with two retention sessions (R1: 30 min after practice; R2: 2 days after practice). We evaluated changes in accuracy as well as in the acoustic details of the cluster production at each time point. Results: Overall, participants improved in their production of the consonant clusters. Accuracy increased, and duration measures decreased in specific measures associated with cluster production. The change in coordination measured in the acoustics changed both for clusters that were incorrectly produced and for those that were correctly produced, indicating continued motor learning even in accurate tokens. Conclusions: These results aid our understanding of the complexity of nonnative consonant cluster learning. In particular, both factors related to both phonological and speech motor control properties affect the learning of novel speech sequences.

Higher education institutions in the UK have organised into mission groups for the advocacy of shared interests and ideologies. Although research productivity is claimed as a key point of difference between these groups, this claim has received relatively little empirical scrutiny. The current study examined the clustering of UK universities based on the research productivity of academic psychologists. It found evidence for the Russell Group's (RG's) claim that it represents leading, research-intensive universities, at least with respects to the discipline of psychology. Productivity metrics of a representative sample of 1339 academic psychologists were extracted from Scopus and Scimago database and were averaged to derive department level productivity indicators. Results from cluster analysis provided evidence in favour of RG's research superiority claim. Cluster level averages of the cluster comprising RG universities were approximately 50-300% higher than those of the cluster comprising non-RG universities. As anticipated, the universities of Oxford and Cambridge surpassed all others to form separate clusters representing an 'elite' within the RG.

The study of learning behaviors with multi features is of great significance for interactive cooperation. The data prediction and decision are to realize the comprehensive analysis and value mining. In this study, hierarchical learning behavior based on feature cluster is proposed. Based on the massive data in interactive learning environment, the descriptive model and learning algorithm suitable for feature clustering are designed, and sufficient experiments obtain the optimal performance indexes. The data analysis results are reliable. On this basis, the hierarchical learning behaviors based on feature clusters are visualized, the rules of different learning behaviors are summarized, then we propose the practical scheme of interactive cooperation. The hierarchical learning behaviors can be realized by feature clusters, which can effectively improve the modes of interactive cooperation, and help to improve the learning effectiveness.

This experimental investigation seeks to corroborate a knowledge structure sorting task approach as a measure to more fully account for prior knowledge when reading. A latent semantic analysis (LSA) network derived from thousands of texts typically read by first year college students was used to create a prototypical referent network model of the global collective knowledge structure of the key terms in the text. Bilingual Chinese-English participants (n = 205) were randomly assigned to four treatments to sort terms in both languages, then to read an English expository text of an unfamiliar topic, then sort in both languages again, and lastly complete a comprehension posttest. All pre- and post- sorting tasks data were converted to Pathfinder networks as measures of knowledge structure. Multiword clusters in the LSA network were present in the initial pre-reading group-average sorting networks of both languages, but especially in Chinese (their L1), and these clusters tended to persist after reading. Reading had only a small influence on the post-reading group-average networks. Sorting in Chinese had a stronger influence downstream than did sorting in English (L1 > L2 influence). For researchers, these innovative approaches to establish local and global collective knowledge networks show promise as complementary measures to explain learning in terms of knowledge structure alignment and transitions, and pragmatically, sorting tasks are relatively easy to implement and interpret in real classrooms as formative diagnostic measures of conceptual understanding.

The main challenge in higher education is student retention. While many methods have been proposed to overcome this challenge, early and continuous feedback can be very effective. In this article, we propose a method for predicting student final grades in a course using only their performance data in the current semester. It assists students in analyzing how much effort they need to put into the course to obtain the desired grades while helping course instructors identify student types at the early stages of the course to provide better support for them. Our method, initially clusters students into several groups based on experience points (XP) that they obtain during a semester. Then, we estimate cluster size and balance clusters by generating and adding virtual students to the smaller clusters. Finally, we drop unimportant student attributes using a feature selection technique. We then predict their final grades via three different algorithms. We have compared the performance of our method with other approaches using data collected from a course for nine years, using data collected from 679 students. The results indicate that our method outperformed the others while achieving 78.02% average accuracy only four weeks after starting the course. It shows we can effectively predict final grades, which will potentially enhance students' learning outcomes.

Over the last decade, a wealth of empirical evidence has accumulated describing the merits of team-based, collaborative, and interdisciplinary research, including: increased productivity among researchers, greater citation impact, increased multi-sector engagement, and the generation of novel solutions to grand challenges. Funding agencies have accordingly increased the frequency of large-scale collaborative and partnered grant opportunities. However, institutional structures and processes can inadvertently limit team-based interdisciplinary research at universities. Research Clusters (which we define as interdisciplinary networks of researchers who organize to solve key challenges facing society) provide a flexible and adaptable mechanism to enable collaborative research across internal and external institutional boundaries. Versions of research clusters are now commonly a central theme in research strategic plans at universities, but there remain very few resources available to research administrators and leadership to support the development of their internal strategies and processes to support research clusters. Here, we describe our experiences developing and implementing initiatives to catalyze clusters of research excellence at the institutional level and reflect on early successes and challenges. We share our framework for identifying, evaluating, and catalyzing research clusters and provide specific examples of internal processes and analytical tools that we have developed.

Educational assessment data are often collected from a set of test centers across various geographic regions, and therefore the data samples contain clusters. Such cluster-based data may result in clustering effects in variance estimation. However, in many grouped jackknife variance estimation applications, jackknife groups are often formed by a random grouping method that ignores the cluster structures of the data. In this study, we constructed both random and cluster-based jackknife groups for data known to have cluster structures and compared the jackknifed standard errors, yielded by two different grouping methods, of item response theory (IRT) scaling coefficient estimates and equated scores. Three independent data samples from an international test of English were used for the study. The cluster-based jackknife group results showed relatively larger jackknifed standard errors of scaling coefficient estimates and scale scores than the results of the random jackknife groups for all three data samples. For cluster-based assessment data, the cluster-based jackknife approach provides a more appropriate way to estimate the standard errors of the parameters of IRT calibration, scaling, and equating analyses. [The content of this manuscript was presented at the 2015 annual meeting of the American Educational Research Association in Chicago, Illinois, April 16-20, 2015.]

The purpose of this study is to perform bibliometric analysis of the scientific researches published in mathematics education subject area between 1980 and 2018, to find out the general layout of the scientific knowledge and communication structure of the field using an objective method, driven from the data. The publications were analyzed according to publication year, type of publication, language of publication, title of publication, name of the author, country of the author and number of citations taken from the sources scanned in Web of Science (WoS) Core Collection database; structural and temporal analysis of various networks, such as collaboration networks and author collaborations were performed. In addition, a word analysis was also conducted in the study to determine the concepts used in the researches published in mathematics education area and covered in the data set. It was found that the publications involving mathematics education increased from 1980 to 2018. The review of the countries according to the number of studies that they have performed in the field revealed that the most productive countries are United States, England and Turkey. Regarding the centrality of the countries in the network, the countries that are located at the center of the network were found to be England, US and Spain, respectively. Regarding the ranking of the countries according to the number of citations taken for the scientific works that they have performed in mathematics education area, it was found that US, Turkey and Malaysia had the highest number of citation bursts. Regarding the ranking of the works according to title, the topics that have been mostly worked on are elementary mathematics classroom, teacher education and achievement gap. The words that are mostly used in mathematics education research area were mathematics, education, student and achievement, respectively.

For decades, cluster initiatives and funding programmes have been used as instruments of industrial and innovation policy--addressing system failures by strengthening linkages among actors, fostering innovation, and developing more effective innovation systems. More recently, a growing segment of these initiatives are also focused on driving system-level transformation and contributing to broader societal benefits. This segment is characterized by larger-scale and longer-term strategic efforts involving a variety of stakeholders across different parts of society, aimed at contributing to addressing societal challenges. These characteristics are shared with the emerging frame of transformative innovation policy, which highlights the importance of embedded practices of learning and reflexivity to enable continuous monitoring of progress and inform and adapt the direction of systemic change processes--requiring new approaches to governance and evaluation. Despite deep experience with implementing cluster programmes and other systemic innovation policy instruments, practitioners still struggle with monitoring and evaluation. Current approaches focus on evidencing strengthened innovation (and economic effects) on the level of firms and research actors, and fail to capture contributions on the level of the broader system. This article presents an evolving approach for tracking system transformation in clusters and collaborative innovation initiatives. Through an interactive, co-development process with initiatives in the Swedish Vinnväxt programme, this research proposes a definition and set of system effect categories for cluster initiatives. It tests a participatory approach for tracking their contribution to system-level change over time, providing an initial case on which to build and apply in other transformative innovation programmes.

Objectives: To determine whether first-year college students cluster in networks based on subjective perceptions of loneliness. Participants: 492 first-year Notre Dame students completed surveys across two semesters and provided communication data used to reconstruct their social networks. Methods: Subjective perceptions of loneliness are measured using the Social and Emotional Loneliness Scale for Adults (SELSA). Correlations between an individual's loneliness and the average loneliness of their alters are compared to associations in random networks created using a rewiring algorithm to determine statistical significance. Results: During their first semester, students are more likely than chance to form ties with other students with similar levels of family and romantic loneliness. In their second semester, students cluster on romantic loneliness but not on family or social loneliness. Conclusions: Students are more likely than chance to form ties with people with similar self-perceived levels of loneliness, but only for certain types of loneliness and during certain periods.

Feedback is an essential component of learning environments. However, providing feedback in populated classes can be challenging for teachers. On the one hand, it is unlikely that a single kind of feedback works for all students considering the heterogeneous nature of their needs. On the other hand, delivering personalized feedback is infeasible and time-consuming. Available automated feedback systems have helped solve the problem to some extent. However, they can provide personalized feedback only after a draft is submitted. To help struggling students "during" the writing process, we can use machine learning to cluster students who benefit the same from feedback using keystroke logs. We can apply the results in automated feedback systems that provide process feedback. In this study, we aim to find homogeneous student profiles based on their writing process indicators. We use fourteen process indicators to find clusters in the data set. We used these measures in a four-stage analysis, including (a) data preprocessing, (b) dimensionality reduction, (c) clustering, and (d) the analysis of the writing quality. Clustering techniques identified five different profiles: Strategic planners, Rapid writers, Emerging planners, Average writers, and Low-performing writers. We further validated the emerged profiles by comparing them concerning students' writing quality. The present work broadens our knowledge of how students interact with writing tasks and addresses how variations in writing behaviors lead to qualitatively different products. We discuss the theoretical underpinnings and potentials of finding profiles of students during writing in higher education.

Purpose: This study analyzed segment differences of student preference for video use in lecture classes and university use of video lecture classes. The authors then conducted novel gap analyses to identify gaps between student segments' preferences for videos versus their level of exposure to in-class videos. Multivariate analysis of variance (MANOVA) was used to identify significant factors that explain the gaps. Design/methodology/approach: Segment differences of student preference for video use in lecture classes and university use of video lecture classes were analyzed. Novel gap analyses were then conducted to identify gaps between student segments' preferences for videos versus their level of exposure to in-class videos. MANOVA was used to identify significant factors that explain the gaps. Findings: Gap analysis of video preference relative to video exposure showed a bimodal distribution, with an approximately even split between students with an overall deficit (44.5%) and surplus (47%) of in-class videos. Deficit means students preferred to see more videos than what the lecturer showed them. Surplus means the lecturer showed students more videos than they preferred to see. Further analyses break down the deficits and surpluses based on the type of videos shown. Practical implications: Results are useful as an effective diagnostic tool for education managers because they are not at the individual student level but rather by course level. One implication for educational managers is that a one-size-fits-all approach for all courses will benefit some students and annoy others. Originality/value: This paper extends Alpert and Hodkinson's (2019) findings by identifying preference clusters and performing segmentation analyses based on finer-grained disaggregated data analysis.

This study sought to discover the challenges facing the implementation of School Cluster System in Oshikoto region. Five Cluster Centre Principals (CCPs) and twenty eight teachers were sampled as participants. Questionnaires, documentary analysis and focus group interviews were used as research instruments for the study. Questionnaires were administered to CCPs and focus group interviews were carried out to teachers. The documentary analysis provided information from the Circuit office on how schools were clustered. The study established that while the implementation of School Cluster System was a positive move for education decentralisation and sharing of resources between schools, the system did not have a legal policy framework to regulate the system operations, and no resources were made available to schools for the implementation and management of School Cluster System activities. The study recommended that educational planners should pursue the legalisation of School Cluster System. In the meantime, it was necessary for principals and teachers to be provided with rewards and adequate resources for implementing School Cluster System activities.

In this study, it was aimed to identify the perceptions of preservice preschool teachers about music lesson through metaphors. The study group of the research includes 137 teacher candidates studying in Trakya University, Faculty of Education, and the Department of Preschool Education in 2015-2016 spring semester and taking the course of Music Education II. With the aim of identifying metaphoric perceptions related to music course, teacher candidates were asked for completing the sentence, "Music course is like …; because…". The data were analyzed with content analysis among qualitatively research methods and 58 metaphors were obtained. These metaphors were gathered under 18 different conceptual categories depending on their common features and the reasons of explaining metaphors by the teacher candidates. As a result of the study, it has been identified that teacher candidates explained the concept of music course with the metaphors different from each other and they use the metaphor of "love" and "rainbow" mostly of these metaphors.