Search

The paper has three major objectives. The first objective of the paper is to synthesize and define common categories of meta-analysis. The second objective is to propose a way to comprehend these common categories of meta-analysis through learning from their respective generic conceptual frameworks. The third objective is to point out which R packages to use to conduct a meta-analysis for each category presented in this paper. In practice, research novices may not know where to begin to learn a topic of meta-analysis for their research. Furthermore, they may not know how the topic of meta-analysis progresses from basics to advanced. This makes it almost next to impossible to use meta-analysis in their research. Certainly, this is problematic. Thus, this paper intends to help research novices overcome this problem. To achieve this, the three objectives mentioned at the beginning were implemented. To propose a synthesis of common categories of meta-analysis along with their respective definitions, a literature review was used as a method to synthesize and define the common categories of meta-analysis presented in this paper. Here, synthesizing meta-analysis categories in one figure would make it much easier for research novices to see the breadth of the subject in question. To propose an easy way to understand the concepts of meta-analysis through learning from their respective generic conceptual frameworks, personal experience was used as a method to make that offering. Here, it is encouraged that research novices should go beyond the definitions of categories of meta-analysis. They can try to frame their research by using example conceptual frameworks presented in the paper as a starting point. To point out which R packages to use with what categories of meta-analysis, a literature review was used as a method to identify those appropriate R packages available. Here, it would be useful for research novices to know which R packages they may need for their meta-analysis research. In summary, the contributions of the paper here would make research novices see and understand the common categories of meta-analysis from the bottom to the top of the pyramid, also offers an easy way to understand the common categories of meta-analysis by illustrating how to develop conceptual frameworks for the common categories of meta-analysis, and finally recommending the R packages to be used with the categories of meta-analysis. Ultimately, we hope that research novices will be able to use meta-analysis successfully in their research.

A common diagnosis in the literature is that the writing of Spanish speakers is generally a structural problem. The writing of 81 university students was analysed by classifying the teacher's comments based on 11 variables that were recorded once during a continuous evaluation that supported the assessment. The techniques used were Content Analysis, Cluster Analysis and Self-Study. Four Clusters were identified in which the dominant conception of communication is broken down, while one Cluster was identified with the alternative conception of communication. This expresses compliance with the task instructions, while the remaining Clusters show progressive non-compliance in various measures. The five Clusters form a typology and, with the theoretical assumption of the identification of the respective conceptions, correspond to a situated communication model. It could be applied in the planning of teaching workloads, as well as in the semi-automatic checking of written productions through the implementation of AI solutions.

One important line of inquiry in educational psychology involves the study of change of individuals' cognitive-motivational processes. The conjunctive use of longitudinal data with latent growth curve modeling procedures has, for example, allowed researchers to identify initial levels and to trace trajectories of theoretical variables such as self-efficacy over time. The study reported in this article proposed a conceptual model that depicted relations between a deep-learning approach, mastery goals and self-efficacy over time. A final sample of 195 second-year university students (100 females, 95 males) took part in this three-wave panel study. We used various inventories to test the initial states and rates of change of the three aforementioned constructs. As an a posteriori analysis, we included prior academic achievement as a possible predictor of change. The results ascertained from our analyses indicate an increase in growth of a deep-learning approach, mastery goals and self-efficacy across the two-year period. Importantly, a posteriori results accentuated the role of prior academic achievement as a predictor of the initial level of personal self-efficacy.

Three quantitative methods are compared for their ability to understand different COVID-19 fatality ratios in 33 OECD countries. Linear regression provides a limited overview without sensitivity to the diversity of cases. Cluster Analysis and Dynamic Patterns Synthesis (DPS) gives scrutiny to the granularity of case similarities and differences, and reveals case exceptions. Qualitative Comparative Analysis (QCA) develops causal theory about what conditions are sufficient for explaining outcomes by using robust and transparent conventions. Configurational case-based methods offer important advantages over inferential statistics when there is a need to focus on diversity in small n. These techniques can be combined as multi-methods. DPS and QCA can be used concurrently to aid research insights. These methods are also strengthened by additional qualitative evidence about the cases.

When randomized control trials are not possible, quasi-experimental methods often represent the gold standard. One quasi-experimental method is difference-in-difference (DiD), which compares changes in outcomes before and after treatment across groups to estimate a causal effect. DiD researchers often use fairly exhaustive robustness checks to make sure the assumptions of the DiD are met. However, less thought is often put into the approach to score item responses from the outcome measure used. For example, surveys are often scored by adding up the item responses to produce sum scores, and achievement tests often rely on scores produced by test vendors, which frequently employ a unidimensional item response theory (IRT) scoring model that implicitly assumes control and treatment participants are exchangeable (i.e., that they come from the same distribution). In this study, several IRT models that parallel the DiD design in terms of groups and timepoints are presented, and their performance is examined. Results indicate that using a scoring approach that parallels the DiD study design reduces bias and improves power, though these approaches can also lead to increased Type I error rates.

Support vector classifiers are one of the most popular linear classification techniques for binary classification. Different from some commonly seen model fitting criteria in statistics, such as the ordinary least squares criterion and the maximum likelihood method, its algorithm depends on an optimization problem under constraints, which is unconventional to many students in a second or third course in statistics or data science. As a result, this topic is often not as intuitive to students as some of the more traditional statistical modeling tools. In order to facilitate students' mastery of the topic and promote active learning, we developed some in-class activities and their accompanying Shiny applications for teaching support vector classifiers. The designed course materials aim at engaging students through group work and solidifying students' understanding of the algorithm via hands-on explorations. The Shiny applications offer interactive demonstration of the changes of the components of a support vector classifier when altering its determining parameters. With the goal of benefiting the broader statistics and data science education community, we have made the developed Shiny applications publicly available. In addition, a detailed in-class activity worksheet and a real data example are also provided in the online supplementary materials.

The purpose of this study is to demonstrate clustering methods within a functional data analysis (FDA) framework for identifying subgroups of individuals that may be exhibiting categories of misfit. Person response functions (PRFs) estimated within a FDA framework (FDA-PRFs) provide graphical displays that can aid in the identification of persons that have responded unexpectedly to items comprising an achievement test. Typical person fit statistics are also useful for detecting unexpected response patterns, but they do not provide insight into the underlying behaviors responsible for those responses. However, different responding behaviors tend to produce FDA-PRFs of different shapes, and may provide additional information regarding the reasons for misfit. Functional clustering methods are useful for categorizing respondents into subgroups based on the shapes of their FDA-PRFs. In this study, a small simulation illustrates the potential of clustering FDA-PRFs for identifying persons displaying common types of responding behaviors. The methodology is also applied to data from a high school biology assessment and a mathematics achievement test. Clustering FDA-PRFs offers a promising methodology for operationalizing person fit evaluations in large-scale assessments, and may be a valuable step in person fit assessment when used in conjunction with traditional indices of psychometric quality.

As an excellent means of improving students' effective learning, knowledge tracking can assess the level of knowledge mastery and discover latent learning patterns based on students' historical learning evaluation of related questions. The advantage of knowledge tracking is that it can better organize and adjust students' learning plans, provide personalized guidance, and thus, achieve the purpose of artificial intelligence-assisted education. However, existing methods, for instance, convolutional knowledge tracing, lacking consideration of graph structure and multivariate time-series prediction, result in poor prediction accuracy. Inspired by recent successes of the graph neural network (GNN), we present a novel multivariate graph knowledge tracking (MVGKT) framework to address these limitations. Specifically, a multivariate time-series knowledge tracking system based on spatio-temporal GNN is designed to model student learning trajectories in different spatial and temporal dimensions and capture both temporal dependencies and interstudent correlations. MVGKT incorporates a gate recurrent unit attentive mechanism and graph Fourier transform, discrete Fourier transform, and graph convolution network to increase the predictive accuracy of the student performances. In addition, we design a question difficulty extraction system to obtain information on the difficulty of the questions, and thus, enhance the data features. Numerous experiments on the ASSISTments dataset have demonstrated that MVGKT is superior to existing knowledge-tracking methods on four metrics and has shown that our approach can enhance the predictive accuracy of student performance.

The National Assessment of Educational Progress (NAEP) program only reports state-level subgroup results if it samples at least 62 students identifying with the subgroup. Since some subgroups constitute small proportions of many states' general student populations, these minority subgroups are seldom sufficiently sampled to meet this sample size requirement. Consequently, education researchers and policymakers are regularly left without a comprehensive understanding of how states are supporting the learning and achievement of different subgroups of students, including underserved subgroups. Using grade 8 mathematics results in 2015, this study compares the performance of two separate techniques in predicting mean subgroup achievement on NAEP. Results demonstrate that a small area estimation technique, known as the Fay-Herriot (FH) model, represents a promising approach. Follow-up research should employ the FH technique with math achievement data from other testing years, as well as data from NAEP reading to evaluate the technique's generalizability.

The use of personality inventories has been limited because of their cost and the length. To overcome these limitations, this study created the Personality Identity Estimator (PIE), an easy-to-use inventory to estimate personality types that can be used at no cost. PIE is a categorical inventory containing 12 items with 3 items for each of the 4 personality-type dimensions in Jungian theory. A sample of 1,104 was used to create PIE. Validity was established through multivariate analyses using data from the administration of the Myers-Briggs Type Indicator to 553 of the sample. Reliability was established by test-retest. Consequently, PIE is a new short-form instrument to estimate personality type following Jung's concept of personality types. PIE can be completed quickly, is easy to score and interpret, and can be used for self-assessment of personality types. PIE can be used to address individual differences, to increase self-awareness, and as an interactive instructional tool. Thus, PIE is a new valid and reliable tool that can be used in both instructional and research settings confidently and at no charge. Permission is granted to use the Personality Identity Estimator in practice and research. A printable copy of PIE is appended.

The main objective of this research was to study the factors influencing the science and environment of education program for blind students at the elementary level. This research used mixed methods (quantitative and qualitative approaches), specifically a questionnaire survey was conducted to better understand the current situation on Science and Environmental Education Learning among 192 blind students in 14 primary schools for the blind in Thailand and in-depth interviews with the directors, science and Environmental teachers, and students' parents, 30 interviewees in total. The data were collected and analyzed into frequency and percentage (and K-means clustering) using SPSS software. The research findings from data collection from 192 blind students illustrated that the importance of facilitator that involve teaching materials (43%) was support appropriate to Science and Environmental Education teaching materials for blind students, 34% students had a good relationship with their classmates and teachers. A total of (32%) were provided more Braille textbooks for Science and Environmental Education teaching. K-means cluster analysis showed four clusters of science and environmental education learning blind students. The study concluded that the factors influencing effectiveness of science and environmental education learning of blind students consisted of facilitators, the creation of scientific learning processes, media and technology as the medium for communicating scientific knowledge, appropriate curriculum for blind students, and cognitive abilities of blind students in physical, mental, intelligence, and emotion terms.

Are students who prefer online education different from those who prefer on-ground education, and how? This is an important question because educational institutions need to better understand student segmentations. This research examined 251 survey responses from students enrolled in Computer Information Systems courses at three universities over five years (2016-2021) and reviewed student attitudes, perceived skills, and their sociological characteristics. Through two-step cluster analysis unsupervised machine learning, two distinct clusters of students emerged, namely Onliners and On-grounders. The top nine out of the eleven group characteristics for Onliners are: select more online courses, regard online instruction as effective, work better without supervision, rely less on classroom interactions in learning, value convenience, can prioritize, are better organized, better prepared, and older. By understanding these group characteristics, educational institutions can make better decisions in policy making, resources allocation, and student recruitment/retention.

MOOCs (massive open online courses) have attracted considerable attention from researchers. Fueled by constant change and developments in educational technology, the trends of MOOCs have varied greatly over the years. To detect and visualize the developments and changes in MOOC research, 4,652 articles published between 2009 and 2021 were retrieved from Web of Science and Scopus with the aid of CiteSpace. This study sought to explore the number of publications, co-citation network, cluster analysis, timeline analysis, burstness analysis, and dual-map overlays based on co-citation relationships. The first finding was that the number of publications on MOOCs had increased consistently, and grew especially quickly between 2013 and 2015. Second, the main topic of the top 10 co-cited studies revolved around the problem of learner continuance. Third, blended programs, task-technology fit, and comparative analysis have emerged as popular subjects. Fourth, the development of MOOC research has followed distinct phases, with 2009 to 2012 the starting phase, 2013 to 2015 the high growth phase, 2016 to 2018 the plateau phase, and 2019 to 2021 another peak phase. Lastly, both cluster analysis and dual-map overlays provided empirical evidence of cross-disciplinary research. Our findings provided an in-depth and dynamic understanding of the development and evolution of MOOC research and also proposed novel ideas for future studies.

In seeking to close equity gaps within a first-year student seminar course, course designers leveraged emerging research on intrapersonal competency cultivation, known to significantly predict student success across diverse students (NAS, 2018). After re-designing the course to intentionally cultivate specific intrapersonal competencies, researchers set out to explore how well the course closed historical institutional equity gaps as measured by end-of-term GPA. Over four years of data collection and course refinement, traditional regression analysis were useful for informing course improvements that resulted in the closing of some equity gaps. However, students were still being placed on academic probation and certain identities of students were over-represented in academic probation numbers. As such, the team utilized random forest, cluster analysis, and then regression analysis that allowed them to focus improvement efforts on a cluster of students that would have otherwise remained unidentified through traditional analysis measures.

Purpose of the study is to substantiate the feasibility of using the model of benchmarking of educational services in the socially responsible educational-innovative cluster in the context of COVID-19 pandemic. Specifically, the authors focus on justify a "nuclear" approach to the cluster formation. The process-oriented benchmarking model was used to apply best practices of providing higher educational service in the context of COVID-19 pandemic. The method of factor analysis was used to determine the impact of each of the 4P subsystems of benchmarking. Using the method of benchmarking makes it possible to develop a final competitive product in the context of the COVID-19 pandemic -- an educational service for all industry and territorial stakeholders. In the light of COVID-19 pandemic, the formation of socially responsible educational-innovative clusters in Ukraine is one of the most promising and effective trends to modernize the provision of educational service. The uniqueness of the educational service is created by strengthening and synergizing the competitive advantages and competencies of all participants in the educational cluster: teachers, students, employers, research staff, the local community.

A deep understanding of the learning level of online learners is a critical factor in promoting the success of online learning. Using knowledge structures as a way to understand learning can help analyze online students' learning levels. The study used concept maps and clustering analysis to investigate online learners' knowledge structures in a flipped classroom's online learning environment. Concept maps (n = 359) constructed by 36 students during one semester (11 weeks) through the online learning platform were collected as analysis objects of learners' knowledge structures. Clustering analysis was used to identify online learners' knowledge structure patterns and learner types, and a non-parametric test was used to analyze the differences in learning achievement among learner types. The results showed that (1) there were three online learners' knowledge structure patterns of increasing complexity, namely, spoke, small-network, and large-network patterns. Moreover, online learners with novice status mostly had spoke patterns in the context of flipped classrooms' online learning. (2) Two types of online learners were found to have different distributions of knowledge structure patterns, and the complex knowledge structure type of learners exhibited better learning achievement. The study explored a new way for educators to analyze knowledge structures by data mining automatically. The findings provide evidence in the online learning context for the relationship between complex knowledge structures and better learning achievement while suggesting the existence of inadequate knowledge preparedness for flipped classroom learners without a special instructional design.

Background: Children and adolescents in residential care have more adverse childhood experiences (ACEs) than those living in any other type of care. However, there is scant research on risk factors accounting for this accumulation. Objective: To establish the prevalence of ACEs in child residential care and to explore predictors of accumulation and clustering patterns. Method: A cluster-randomized sample of 1275 children ages 5-18 years was drawn from 78 residential care facilities across Romania. Clinical psychologists from Child Protective Services (CPS) aided in the collection of ACEs, family, health and history data. Network analysis was used to explore ACE clustering patterns, while a Zero-Inflated Negative Binomial regression provided insights into covariate risk factors. Results: Within the network, influential ACEs were found to be consistent with previous research. However, using expanded ACEs on a high exposure sample, two new clusters emerged: (1) ACEs reported by children at the highest mean ages (abused by a romantic partner, arrested or incarcerated, sexually abused), (2) ACEs pertaining to social exclusion (seriously medically ill, discriminated, bullied). Several risk factors significantly predicted ACE accumulation: craniocerebral trauma, prenatal exposure to drugs, cognitive delay, conduct disorder and frequent change of homes. Institutional placement and special education were found as protective factors. Conclusions: A child in residential care that was institutionalized for reasons other than neglect or abuse had 80% probability of not being ever exposed to any ACEs. Nevertheless, CPS can utilize important risk factors and influential past ACEs to detect children at higher risks of further adversity.

Multiple imputation (MI) is a popular method for handling missing data. In education research, it can be challenging to use MI because the data often have a clustered structure that need to be accommodated during MI. Although much research has considered applications of MI in hierarchical data, little is known about its use in cross-classified data, in which observations are clustered in multiple higher-level units simultaneously (e.g., schools and neighborhoods, transitions from primary to secondary schools). In this article, we consider several approaches to MI for cross-classified data (CC-MI), including a novel fully conditional specification approach, a joint modeling approach, and other approaches that are based on single- and two-level MI. In this context, we clarify the conditions that CC-MI methods need to fulfill to provide a suitable treatment of missing data, and we compare the approaches both from a theoretical perspective and in a simulation study. Finally, we illustrate the use of CC-MI in real data and discuss the implications of our findings for research practice.

In order to evaluate the effect of a policy or treatment with pre- and post-treatment outcomes, we propose an approach based on a transition model, which may be applied with multivariate outcomes and accounts for unobserved heterogeneity. This model is based on potential versions of discrete latent variables representing the individual characteristic of interest and may be cast in the hidden (latent) Markov literature for panel data. Therefore, it can be estimated by maximum likelihood in a relatively simple way. The approach extends the difference-in-difference method as it is possible to deal with multivariate outcomes. Moreover, causal effects may be expressed with respect to transition probabilities. The proposal is validated through a simulation study, and it is applied to evaluate educational programs administered to pupils in the sixth and seventh grades during their middle school period. These programs are carried out in an Italian region to improve non-cognitive skills (CSs). We study if they impact also on students' CSs in Italian and Mathematics in the eighth grade, exploiting the pretreatment test scores available in the fifth grade. The main conclusion is that the educational programs aimed to develop noncognitive abilities help the best students to maintain their higher cognitive abilities over time.

A Rasch analysis was conducted to assess a rubric that measures the quality of school-level variables that support the implementation of a social and emotional learning (SEL) program. SEL planning teams from 84 schools rated the quality of school-level variables (e.g., principal support, available funds, and resources) that support the implementation of an SEL program in their schools. A Rasch analysis suggested that the rubric was a two-dimensional solution, Planning and Activity. Ratings were collected for 3 years. Cluster analysis demonstrated that there were five groups of schools with different trends in their implementation process. The analysis demonstrated that the rubric can be improved by revising items and writing additional items to measure the full range of implementation quality.

Despite several theoretical, structural, and statistical issues reported against Five Facets Mindfulness Questionnaire (FFMQ), most studies in the educational sector of Pakistan use it without analyzing its factor structure. Since culture might change the structure, this first systematic study filled the gap and explored the factor structure of the least researched FFMQ (15 items) among nonmeditator university students in collectivist society Pakistan while employing optimal statistical techniques. We interpreted our results culturally and also conducted a cross-cultural examination with the Spanish sample. We employed polychoric data to conduct exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). The EFA extracted a four-factor (10 items) solution without the Describe facet and showed good inter-item correlation. The CFA exhibited a theory-based two-factor hierarchical model (self-regulation, orient to experience) with good fit indices, acceptable factor loadings, correct directions, and strong reliabilities (coefficient H > 0.8). The multivariate analysis showed no difference in mindfulness between males and females, except for Nonreact. The first multigroup CFA (MGCFA) showed metric invariance while examining the polarity of items across gender. In cross-cultural multivariate analysis, we found no difference in mindfulness between Pakistan and Spanish samples, except for Nonjudge, but the second MGCFA showed metric invariance in the correlated model only. Our data suggest that the two-factor hierarchical structure, without Describe facet, provides a more parsimonious and culturally accepted structure among nonmeditators in Pakistan. However, due to the independent nature of the two factors, the total score cannot be computed. The FFMQ (15 items) appears individualistic, and including culturally acceptable items will improve its universality.

Although educational research and evaluation generally occur in multilevel settings, many analyses ignore cluster effects. Neglecting the nature of data from educational settings, especially in non-randomized experiments, can result in biased estimates with long-term consequences. Our manuscript improves the availability and understanding of artificial neural networks, an underutilized method trending in other disciplines. This method also shows promise for dealing with challenges faced by educational researchers, such as analyzing clustered data. Therefore, we simulated data to generalize the potential benefits of artificial neural networks to different data types. We also compared artificial neural networks to more familiar methods and investigated the time it demanded to perform each technique. Hence, readers can decide when it may be more appropriate to use one method instead of another.

The aim of this study is to investigate the presence of DIF over the gender variable with the latent class modeling approach. The data were collected from 953 students who participated in the PISA 2018 8th-grade financial literacy assessment in the USA. Latent Class Analysis (LCA) approach was used to identify the latent classes, and the data fit the three-class model better in line with fit indices. In order to obtain more information about the characteristics of the emerging classes, uniform and non-uniform DIF sources were identified by using the Multiple Indicator Multiple Causes (MIMIC) model. The findings are very important in terms of contributing to the interpretation of latent classes. According to the results, the gender variable was a source of DIF for latent classes. It is important to include direct effects by gathering unbiased estimates for the measurement and structural parameters. Disregarding these effects can lead to incorrect identification of implicit classes. A sample application of MIMIC model was performed in a latent class framework with a stepwise approach in this study.

This research aims to propose a bibliometric map of studies on the use of STEM in education. This study used publication co-citation analysis, author co-citation analysis, and word frequency analysis methods to reveal the structure and transformation of STEM literature. Descriptive data such as the distribution of studies in the field by country, institution, and time were obtained from the Web of Science (WoS) database. We used the RStudio program for bibliometric analysis. The International Journal of STEM Education is the most widely published. Journal of Science Education and Technology received the most citations. Guzey S.S. is the author with the most publications on the subject. Capraro M.M. most cited author. The university that publishes the most is Purdue University. The USA is the country with the highest number of publications. The paper by Blickenstaff titled Gender and Education in 2005 has been cited the most worldwide. The research by Maltese, which was published in 2019, had the highest local citation rate. According to the results of cluster analysis, four clusters were formed. The term "STEM" appears to be present in all clusters. "STEM Education" was also included in three clusters.

This paper investigates the viewpoints and perspectives of 179 undergraduate students engaged in a co-creation project regarding their anticipated progress, exploring, simultaneously, the affecting factors. The students attended the physics course in the Department of Industrial Design and Production Engineering at the University of West Attica, Greece. The investigation is implemented through a questionnaire that appraises the students' co-creation expectations in association with collected demographic data via twenty-two close-ended questions (Q1-Q22). Several statistical data sets are presented, including descriptive and correlation statistics and principal component and exploratory factor analyses. The majority of the replies are provided by male participants in their first year of study, with an average age of between 18 and 19. The correlation coefficient between the questions ranges from -0.04 to 0.73, with the maximum occurring between the questions Q19 and Q20. Factor analysis justified by KMO (0.862) and Bartlett's sphericity (1749.843, p= 0.000) tests indicates five principal components within the following factors determining the undergraduate co-creation activities: responsible behaviour, feedback, helping, tolerance, and personal interaction. The above findings may contribute to the implementation of students' co-creation as they are valuable tools for the design and pre-requisites for a successful implementation.

Concept maps are powerful tools used to reveal challenges in students' learning. However, their use introduces complexities when a large group of students' conceptualizations need to be examined. In this study, concept maps of 344 general chemistry students were analyzed after grouping them based on achievement in chemistry, math proficiency, and gender. The analysis was also expanded with the consideration of eccentricity values and the extended chemistry triplet. Although some similarities exist between the map of high-achieving students in chemistry and that of high-performing students in the Mathematics Placement Test (MPT), the calculated eccentricity values show interesting variations. On the other hand, the analysis of the map of the low-performing students in MPT and that of low-achieving students in chemistry revealed no clear patterns of symbolic, macroscopic, and submicroscopic terms. Practical suggestions were included to increase the use of representative maps in different assessment and teaching scenarios.

This study aimed to utilize sentiment and sentence similarity analyses, two Natural Language Processing techniques, to see if and how well they could predict L2 Writing Performance in integrated and independent task conditions. The data sources were an integrated L2 writing corpus of 185 literary analysis essays and an independent L2 writing corpus of 500 argumentative essays, both of which were compiled in higher education contexts. Both essay groups were scored between 0 and 100. Two Python libraries, TextBlob and SpaCy, were used to generate sentiment and sentence similarity data. Using sentiment (polarity and subjectivity) and sentence similarity variables, regression models were built and 95% prediction intervals were compared for integrated and independent corpora. The results showed that integrated L2 writing performance could be predicted by subjectivity and sentence similarity. However, only subjectivity predicted independent L2 writing performance. The prediction interval of subjectivity for independent writing model was found to be narrower than the same interval for integrated writing. The results show that the sentiment and sentence similarity analysis algorithms can be used to generate complementary data to improve more complex multivariate L2 writing performance prediction models.

A shared learning outcome for baccalaureate programs is for students to think critically and utilize critical thinking skills. However, education in Saudi Arabia seems to has focused more on rote memorization. Critical Thinking (CT), one of several essential skills regarded in college outcomes, has become a vital citizenry skill in Saudi. In this study, the researcher examined whether or not instructors apply CT or use its components in their teaching. This study offers professional development specialists insight on what to focus on when planning and developing critical thinking workshops. Participants in this study included instructors from a small-sized university in Saudi Arabia. A descriptive analysis, multiple regression, and MANOVA tests were implemented. The results of this study indicated that the majority of instructors use CT within their lectures. Furthermore, the results showed that younger instructors seem more likely to implement CT activities. There was no significant difference between science disciplines and humanities disciplines when implementing CT in their classes.

This study aims to discover groups of students enrolled in the emergency remote teaching online course based on the various course-related data collected throughout the first year of COVID-19 pandemic. Research was conducted among 222 students enrolled in the course "Business Informatics" at the Faculty of Organization and Informatics of the University of Zagreb in the academic year 2020/2021. Overlays were used to model students' success on the various quizzes and exams within the course. The k-means clustering was employed to classify students into groups, based on combination of students' overlay values, frequency of accessing course lessons and the final grades. Three distinct clusters (i.e., students' groups) were discovered and explained in the given context. The identified groups of students can be used for future adaptations of the online course design in order to improve the retention and their final grades.

Bibliometric mapping serves as a method to systematically evaluate and visually demonstrate the development of a research field. CiteSpace and VOSviewer, two research tools of bibliometric mapping, were used in the present study to analyze, synthesize, and visualize the hot topics as well as frontier evolution of science education. Co-authorship analysis, co-citation analysis, co-occurrence analysis, cluster analysis, and content analysis were conducted based on 6278 articles selected from seven SSCI journals. Researchers from countries/territories in North America, Europe, Oceania, and West and East Asia had maintained relatively tighter cooperation with each other. Highly influential literature mainly focused on the standards, methods, practice, and reflection of science education. In the past two decades, the literature on science education covered seven hot topics: conceptual issues in science education, gender, scientific argumentation, professional development, science learning, evolution, and peer review. The research on science education in the past 20 years can be divided into three phases: the first stage focused on knowledge learning, identity, and informal education; the second stage emphasized formal education, scientific literacy, and social-science issues; and the third stage highlighted scientific argumentation and STEM education.

The authoritative school climate construct, or the degree to which schools demonstrate student support and disciplinary structure, predicts several important student outcomes (e.g., racial suspension gap, student disengagement). To better understand this construct, we used multilevel latent class analysis to identify latent classes of student perceptions of school climate, model school-level variation in student experiences, and examine the relation between school climate and student and school demographics. Using a nationally representative sample of public high students in the USA, results show that student perceptions of school climate fall into four classes, "Authoritative," "Permissive," "Authoritarian," and "Uninvolved," which are consistent with authoritative school climate literature. Student gender, ethnicity/race, and SES all impact a student's likelihood of membership in these classes. Schools tend to have a predominant school climate experience, although considerable variability in individual student experiences within a school exists. Findings address gaps in the authoritative school climate literature and have important implications for future research and policy.

Development of graphical methods for representing data has not kept up with progress in statistical techniques. This article presents a brief history of graphical representations of research findings and makes the case for a revival of methods developed in the early and mid-twentieth century, notably ISOTYPE and Chernoff's faces. It resurrects and improves a procedure, clustered iconography, which enables the presentation of multidimensional data through which readers engage more effectively with the presentation's central message by way of an easier understanding of relationships between variables. The proposed technique is especially well adapted to the needs and protocols of open-source research.

Clinical prediction models are widely used in modern clinical practice. Such models are often developed using individual patient data (IPD) from a single study, but often there are IPD available from multiple studies. This allows using meta-analytical methods for developing prediction models, increasing power and precision. Different studies, however, often measure different sets of predictors, which may result to systematically missing predictors, that is, when not all studies collect all predictors of interest. This situation poses challenges in model development. We hereby describe various approaches that can be used to develop prediction models for continuous outcomes in such situations. We compare four approaches: a "restrict predictors" approach, where the model is developed using only predictors measured in all studies; a multiple imputation approach that ignores study-level clustering; a multiple imputation approach that accounts for study-level clustering; and a new approach that develops a prediction model in each study separately using all predictors reported, and then synthesizes all predictions in a multi-study ensemble. We explore in simulations the performance of all approaches under various scenarios. We find that imputation methods and our new method outperform the restrict predictors approach. In several scenarios, our method outperformed imputation methods, especially for few studies, when predictor effects were small, and in case of large heterogeneity. We use a real dataset of 12 trials in psychotherapies for depression to illustrate all methods in practice, and we provide code in R.

A sample suffers range restriction (RR) when its variance is reduced comparing with its population variance and, in turn, it fails representing such population. If the RR occurs over the latent factor, not directly over the observed variable, the researcher deals with an indirect RR, common when using convenience samples. This work explores how this problem affects different outputs of the factor analysis: multivariate normality (MVN), estimation process, goodness-of-fit, recovery of factor loadings, and reliability. In doing so, a Monte Carlo study was conducted. Data were generated following the linear selective sampling model, simulating tests varying their sample size (N = 200 and 500 cases), test size (J = 6, 12, 18, and 24 items), loading size (L = 0.50, 0.70, and 0.90), and restriction size (from R = 1, 0.90, 0.80, and so on till 0.10 selection ratio). Our results systematically suggest that an interaction between decreasing the loading size and increasing the restriction size affects the MVN assessment, obstructs the estimation process, and leads to an underestimation of the factor loadings and reliability. However, most of the MVN tests and most of the fit indices employed were nonsensitive to the RR problem. We provide some recommendations to applied researchers.

Although many textbooks on multivariate statistics discuss the common factor analysis model, few of these books mention the problem of factor score indeterminacy (FSI). Thus, many students and contemporary researchers are unaware of an important fact. Namely, for any common factor model with known (or estimated) model parameters, infinite sets of factor scores can be constructed to fit the model. Because all sets are mathematically exchangeable, factor scores are indeterminate. Our professional silence on this topic is difficult to explain given that FSI was first noted almost 100 years ago by E. B. Wilson, the 24th president (1929) of the American Statistical Association. To help disseminate Wilson's insights, we demonstrate the underlying mathematics of FSI using the language of finite-dimensional vector spaces and well-known ideas of regression theory. We then illustrate the numerical implications of FSI by describing new and easily implemented methods for transforming factor scores into alternative sets of factor scores. An online supplement (and the fungible R library) includes R functions for illustrating FSI.

Growth mixture models (GMMs) are applied to intervention studies with repeated measures to explore heterogeneity in the intervention effect. However, traditional GMMs are known to be difficult to estimate, especially at sample sizes common in single-center interventions. Common strategies to coerce GMMs to converge involve post hoc adjustments to the model, particularly constraining covariance parameters to equality across classes. Methodological studies have shown that although convergence is improved with post hoc adjustments, they embed additional tenuous assumptions into the model that can adversely impact key aspects of the model such as number of classes extracted and the estimated growth trajectories in each class. To facilitate convergence without post hoc adjustments, this paper reviews the recent literature on covariance pattern mixture models, which approach GMMs from a marginal modeling tradition rather than the random effect modeling tradition used by traditional GMMs. We discuss how the marginal modeling tradition can avoid complexities in estimation encountered by GMMs that feature random effects, and we use data from a lifestyle intervention for increasing insulin sensitivity (a risk factor for type 2 diabetes) among 90 Latino adolescents with obesity to demonstrate our point. Specifically, GMMs featuring random effects--even with post hoc adjustments--fail to converge due to estimation errors, whereas covariance pattern mixture models following the marginal model tradition encounter no issues with estimation while maintaining the ability to answer all the research questions. [For the corresponding grantee submission, see ED618691.]

Over the past few years, academics have undertaken initiatives to bridge the gap between theory and practice in the ever-growing field of business analytics, including implementing real-life student projects in all shapes and forms. Every year since 2015, Manhattan College has invited student teams from across North America and elsewhere in the world to its campus in order to participate in an intercollegiate business analytics competition (BAC@MC). This well-received event and the objectives behind it are described in this article. The program is shown to serve as an effective experiential learning adventure for the undergraduate students as it hones their data analytic skills in the context of an engaging real-world business problem. The roles various stakeholders play in this high-impact practice are highlighted. Furthermore, an example of a recent competition question is presented (along with a summary of the analytical approaches attempted) by the student teams. Descriptive visualizations, regression, and cluster algorithms implemented using python, R, Excel, or Tableau are among the typical analyses utilized by participating students. As witnessed by the students, faculty advisors, and the industry practitioners who attended the event, competitions such as BAC@MC can be rewarding, community-building, and transformative experiences for undergraduate students who will soon become tomorrow's business analysts.

Previous research and experiences have indicated the potential that games have in educational settings. One of the possible uses of games in education is as game-based assessments (GBA), using game tasks to generate evidence about skills and content knowledge that can be valuable. There are different approaches in the literature to implement the assessment machinery of these GBA, all of them having strengths and drawbacks. In this article, we propose using multivariate Elo-based learner modeling, as we believe it has a strong potential in the context of GBA for three aims: first, to simultaneously measure students competence across several knowledge components in a game; second, to predict task performance; and finally, to estimate task difficulty within the game. To do so, we present our GBA Shadowspect, which is focused on solving geometry puzzles, and we depict our implementation using data collected from several high schools across the USA. We obtain high-performing results (AUC of 0.87) and demonstrate that the model enables analysis of how each student's competency evolves after each puzzle attempt. Moreover, the model provides accurate estimations of each task's difficulty, enabling iterative improvement of the game design. This study highlights the potential that multivariate Elo-based learner modeling has within the context of GBA, sharing lessons learned, and encouraging future researchers in the field to consider this algorithm to build their assessment machinery.

The Major Field Test in Business (MFTB) is a nationally administered student evaluation that measures the accumulated knowledge of undergraduates enrolled in a four-year degree program. Research to date has focused primarily on understanding how different variables correlate with performance on this standardized test -- such as student grades, gender and ethnicity. The research objectives of this essay are to analyse the ambiguous results found in previous studies and to highlight how interaction effects among variables can be used to better explain test success. Using data on student performance collected over 11 semesters at a Canadian university, this essay uses a multi-variable regression model to understand the factors affecting scores on the MFTB. The model results suggest that examining the interaction between variables provides important insights and can help to better explain the ambiguity in prior studies. This study is unique in that it uses a statistical measure known as the extra sum of squares F-test to demonstrate the significance of interaction variables.

The goal of this study was to evaluate the measurement invariance of an adapted assessment of motivations for social withdrawal ("Social Preference Scale--Revised"; SPS-R) across cultural contexts and explore associations with loneliness. Participants were a large sample of university students (N = 4,397; M[subscript age] = 20.08 years, SD = 2.96; 66% females) from 10 countries (Argentina, Australia, Canada, China, India, Italy, South Korea, Norway, Turkey, and the United States). With this cross-cultural focus, we illustrate the "multiple-group factor analysis alignment method," an approach developed to assess measurement invariance when there are several groups. Results indicated approximate measurement invariance across the 10 country groups. Additional analyses indicated that overall, shyness, avoidance, and unsociability are three related, but distinct factors, with some notable country differences evident (e.g., in China, India, and Turkey). Shyness and avoidance were related positively to loneliness in all countries, but the strength of the association between shyness and loneliness differed in Italy and India relative to the other countries. Results also indicated that unsociability was related positively to loneliness in the United States only. Theoretical and assessment implications are discussed.

Ensuring comparability of Likert-style items across different countries is a widespread challenge for authors of large-scale international surveys. Using data from the EUROGRADUATE Pilot Survey, this study employs a series of latent class analyses to explore which response patterns emerge from self-assessment of acquired and required skills of higher education graduates and how these patterns vary between eight participating countries. Results show that countries differ in the number of classes which most accurately fit the data structure. The effort to overcome national specifics by combining the levels of acquired and required skills into a single measurement of skill surplus/deficit reduces heterogeneity of patterns across countries and slightly increases comparability, yet the notion that respondents understand the scale in the same way across countries (measurement invariance) is not supported.

In cluster randomized evaluations, a treatment or intervention is randomly assigned to a set of clusters each with constituent individual units of observations (e.g., student units that attend schools, which are assigned to treatment). One consideration of these designs is how many units are needed per cluster to achieve adequate statistical power. Typically, researchers state that "about 30 units per cluster" is the most that will yield benefit towards statistical precision. To avoid rules of thumb not grounded in statistical theory and practical considerations, and instead provide guidance for this question, the ratio of the minimum detectable effect size (MDES) to the larger MDES with one less unit per cluster is related to the key parameters of the cluster randomized design. Formulas for this subsequent difference effect size ratio (SDESR) at a given number of units are provided, as are formulas for finding the number of units for an assumed SDESR. In general, the point of diminishing returns occurs with smaller numbers of units for larger values of the intraclass correlation. [For the corresponding grantee submission, see ED625939.]

Purpose: Identifying appropriate targets for vocabulary instruction and determining the optimal sequence for instruction continue to be a challenge. The purpose of this study is to investigate how previously studied lexical characteristics collectively influence children's word learning. Method: A secondary data analysis was conducted using the word learning results of 350 first-, second-, and third-grade students who participated in an investigation examining the effects of a supplemental vocabulary intervention. We investigated the influence of the following lexical characteristics on the learning of 377 words: word frequency, level of concreteness, phonotactic probabilities, neighborhood density, and age of acquisition using multivariate adaptive regression splines (MARS). Results: MARS modeled the influence lexical characteristics had on word learning and determined the relative importance of each variable for each grade-level model. Results revealed age of acquisition was the most important factor related to word learning in all grades, but contributions of other lexical characteristics and their level of importance differed across models. All respective models fit well, with root-mean-square error values ranging from 0.11 to 0.15 and generalized cross validation scores of 0.01 and 0.03. Conclusions: Nuanced information from the MARS analysis provides insights into how lexical characteristics affect word learning differently for children in different grade levels. This information is key to understanding the vocabulary acquisition of school-aged children. The findings from this research have the potential to inform the development of a word selection framework that will organize vocabulary targets into an appropriate sequence based on relevant predictors. [For the correspoinding grantee submission, see ED628272.]

Latent class analysis (LCA) is a useful statistical approach for understanding heterogeneity in a population. This article provides a pedagogical introduction to LCA modeling and provides an example of its use to understand youths' daily coping strategies. The analytic procedures are outlined for choosing the number of classes and integration of the LCA variable within a structural equation model framework, specifically a latent class moderation model, and a detailed table provides a summary of relevant modeling steps. This applied example demonstrates the modeling context when the LCA variable is moderating the association between a covariate and two outcome variables. Results indicate that students' coping strategies moderate the association between social stress and negative mood; however, they do not moderate the social stress-positive mood association. Online supplemental materials include R (MplusAutomation) code to automate the enumeration procedure, ML three-step auxiliary variable integration, and the generation of figures for visually depicting LCA results.