Search

A study demonstrated utility of cluster analysis to classify a racially diverse group of children from kindergarten to Grade 3. Four victim subtypes were identified: nonaggressive nonasocial; aggressive; asocial; and both aggressive and asocial. Early aggression levels predicted increases in victimization and chronicity. The role of asocial behavior was less clear. Victims who were rejected were more likely to remain victimized than their better accepted counterparts. (Author/KB)

Standardized sensory, perceptual, linguistic, intellectual, and cognitive tests were administered to 470 entering first graders. Factor analysis found four factors, which were reading related skills, visual cognition, verbal cognition, and speech processing. Cluster analysis identified nine profiles. The strongest predictors of later reading and other academic achievement were first, the reading-related skills factor and second, the visual cognition factor. (Contains references.) (Author/DB)

Six core profiles were identified on four fluid reasoning to comprehension knowledge constructs previously found to be both theoretically and empirically linked to children's reading acquisition: auditory processing, crystallized ability, processing speed, and short-term memory. Cluster analysis was used to develop this taxonomy and the six-cluster solution was supported through cross-validation. (Author/VWL)

Introduces a novel approach to unsupervised text summarization. Proposes an "information-centric" approach to evaluation, where the quality of summaries is judged not in terms of how well they match human-created summaries but in terms of how well they represent their source documents in information retrieval tasks such as document retrieval and text categorization. (Author/AEF)

Students completed the Approaches and Study Skills Inventory. Results suggest that small numbers of poorly attaining learners using dissonant study methods occur in a variety of educational settings. These learners are distinct from those lowly attaining ones who unreservedly use surface methods of studying and have low academic aspirations. (EV)

Examines the phenomenon of dissonant forms of memorizing and repetition from both location and structural perspectives. Attention is drawn to clear evidence that supports, in particular, an argument in favor of gender-response differences in both location and structural dissonance. Also illustrates how the isolation of dissonant response patterns may be approached analytically. (Author/EV)

Explored whether children's experiences in child care programs related to individual (ethnicity, age, gender) and classroom (environmental quality) characteristics. Identified six patterns in how children spent their time: creative, language arts, didactic, gross motor, high-level adult involvement, and individualized adult interactions. Found that ethnicity, gender, and environmental quality related to children's activity patterns, but associations for children 10- to 36- months of age differed from those for older children. (Author/KB)

Suggests that groups, or clusters, of industries form partnerships with community colleges in order to positively impact economic development. Asserts that a cluster-oriented community college system requires innovation, specialized resources and expertise, knowledge of trends, and links to industry. Offers suggestions for developing such a system. (NB)

Research fronts, defined as clusters of documents that tend to cite a fixed, time invariant set of base documents, are plotted as time lines for visualization and exploration. Illustrates the construction, exploration, and interpretation of time lines for identifying and visualizing temporal changes in research activity through journal articles. (Author/LRW)

Explores the development of a classification system for two-year institutions that can provide a framework for analysis and contribute to the discourse in public policy. The proposed classification system is based on cluster analyses using data from the National Center for Education Statistics' Integrated Postsecondary Education Data System (IPEDS). (Contains four figures and four references.) (NB)

Studied aspects of listening comprehension skills in 107 second, fourth, and sixth graders in Finland using the Sentence Verification technique. Findings, comparable with previous research, show age-related improvement in listening comprehension for narrative, but not expository, passages. Cluster analysis produced four comprehension strategies. (SLD)

Employed college students (n=401) completed the Self-Directed Search and NEO Personality Inventory-Revised. Person-environment fit across Holland's six personality types predicted job satisfaction and skill development. Five-Factor Model traits significantly predicted intrinsic career outcomes. Use of the five-factor, person-centered approach to complement congruence studies could increase predictive validity. (Contains 41 references.)

These two studies extended research on the recently developed Almost Perfect Scale-Revised (APS-R; R.B. Slaney, M. Mobley, J. Trippi, J.S. Ashby, & D. Johnson, 1996). Cluster analyses were used to identify adaptive and maladaptive perfectionists and nonperfectionists. Differences between the clusters emerged on measures of self-esteem, depression, anxiety, and positive and negative affects. (Author)

Discusses information visualization and cocitation analysis and demonstrates the use of an integrative approach to visualizing and tracking the development of competing scientific paradigms. Assumes that a cluster of highly cited and cocited scientific publications in a cocitation network represents the core of a predominant scientific paradigm. (Author/LRW)

Discusses Web site hyperlink-affiliation network structures based on a study of Korea's top 152 Web sites. Provides a theoretical framework and literature review; shows through cluster analysis that financial Web sites occupy the most central position in the network; and discusses these findings from the perspective of Web site credibility. (Contains 60 references.) (Author/LRW)

In a quantitative content analysis, 128 repots written by German special education teachers about children (ages 7-11) with learning disabilities were analyzed. A cluster analysis of the 34 most frequent statements led to five clusters that differed clearly. No significant correlation between the children's characteristics and membership of a cluster could be found. (Contains references.) (Author/CR)

Explored the problem of choosing the correct number of clusters in cluster analysis of high dimensional empirical binary data. Findings from a simulation that included 162 binary data sets resulted in recommendations about the number of clusters for each index under consideration. Compared and analyzed the performance of index results. (SLD)

Reports on an exploratory study that investigated the relationship between participants' learning style preferences and their perceptions of a professional workshop on collaboration and technology to support collaboration. Discusses educational effectiveness measures and results of cluster analysis that showed learning styles should be considered as a variable that influences learning outcomes and preferences. (Author/LRW)

Uses co-occurrence analysis of INSPEC classification codes and thesaurus terms assigned to medical informatics (biomedical information) journal articles and proceedings papers to reveal a more complete perspective of how information science and information technology (IS/IT) authors view medical informatics. Discusses results of cluster analysis nd multidimensional scaling analysis. (Author/LRW)

Presents the initial phase of a comparison of cocitation and social network analysis methods for the study of scholarly communication, based on a case study of the intellectual contributions of Kurt Lewin, a social psychologist. Highlights include an author tri-citation analysis; cluster analysis; multidimensional scaling; and diffusion of ideas. (Author/LRW)

Factor analysis of data from the Career Decision Scale, Career Factors Inventory, and Career Decision Difficulties Questionnaire for 434 undecided college students resulted in six factors: lack of information, need for information, trait indecision, disagreement, identity diffusion, and choice anxiety. Cluster analysis yielded three clusters: information deficit/identity diffusion, decision process inhibitors, and choice inhibitors. (Contains 46 references.) (SK)

Analyzed annual electronic journal usage data for the NorthEast Research Library (NERL) consortium. Discusses the high degree of skew in journal use; describes results of a cluster analysis that revealed three distinct groups of institutions based on journal use; and recommends that institutions consider their consortial membership and organize themselves into groups of homogeneous institutions. (Author/LRW)

This study developed and examined a multidimensional, interdisciplinary model of social status among 487 fourth- through sixth-grade students. Seven subtypes were identified based on dimensions of likability, perceived popularity, and social dominance (high status, perceived popular/dominant, well-liked/dominant, average, low dominant/unpopular, disliked, and low status). The cluster solution was supported by cluster differences in peer-reported prerogatives of status and self- reported social characteristics. Each cluster had a distinct behavioral profile. (Author/KB)

Evaluated the performance of several alternative cluster analytic approaches to initial model specification using population parameter analyses and a Monte Carlo simulation. Of the six cluster approaches evaluated, the one using the correlations of item correlations as a proximity metric and average linking as a clustering algorithm performed the best. (Author/SLD)

Combines procedures for single-level mediational analysis with multilevel modeling techniques to test mediational effects in clustered data appropriately. Compared, through simulation, the performance of these multilevel mediational models with that of single-level models in clustered data with various real-world characteristics. (SLD)

Presents a variable selection heuristic for nonhierarchical (K-means) cluster analysis based on the adjusted Rand index for measuring cluster recovery. Subjected the heuristic to Monte Carlo testing across more than 2,200 datasets. Results indicate that the heuristic is extremely effective at eliminating masking variables. (SLD)

Describes methods for fitting a fuzzy consensus partition to a set of partitions of the same set of objects. Describes and illustrates three models defining median partitions and compares these methods to an alternative approach to obtaining a consensus fuzzy partition. Discusses interesting differences in the results. (SLD)

Used Monte Carlo methods to evaluate the adequacy of cluster analysis to recover group membership based on simulated latent growth curve (LCG) models. Cluster analysis failed to recover growth subtypes adequately when the difference between growth curves was shape only. Discusses circumstances under which it was more successful. (SLD)

Explores the intellectual subject structure and research themes in software engineering through the identification and analysis of a core journal literature via two expert perspectives: the author, through cocitation analysis; and the indexer, through subject terms selected to facilitate retrieval. Topics include cluster analysis, multidimensional scaling, PFNets, object orientation, and further research. (Author/LRW)

Investigates ways in which people group or categorize documents associated with a voluminous work to guide the construction of organized displays for information retrieval systems. Discusses results of cluster analyses based on "A Christmas Carol" (Charles Dickens) and suggests implications for metadata standards and digital libraries as well as future research. (Author/LRW)

Used cluster analysis of survey data for 506 respondents to create a taxonomy of welfare leavers in Illinois based on their self-reported well-being after leaving welfare. Used classification tree analysis to identify factors associated with different types of leavers. Findings highlight the existence of many marginally successful leavers who could be served with a few additional supports. (SLD)

Study examined data from students reporting driving within one hour of drinking two or more drinks. Three-cluster typology included: Marginals engage in least amount of delinquent behaviors, alcohol use, and drinking-driving; Heavy Drinkers are heaviest drinkers, most frequent drinking-drivers, and have greatest driving exposure; Delinquents exhibit most delinquency, have least driving exposure, and greatest experience of accidents and drinking-driving consequences. (Contains 33 references, 2 tables, and 1 figure.) (ADT)

Evaluated a cross-validation approach for cluster analyses through two simulation studies. Although on the average replication indices peaked in the region of the true number of clusters, choosing the number of clusters by maximum replication results in a negatively biased estimate. Using a "scree test" to attenuate this bias is discussed. (SLD)

Classifies a spectrum of user behavior with electronic journals based on transaction log analysis of SuperJournal and K-Means cluster analysis. Describes results of user interviews and questionnaires from 13 United Kingdom higher education institutions that showed content and ease of use were the most significant factors affecting patterns of use. (Author/LRW)

Provides a general method that adjusts for the inflation of information associated with a test containing item clusters and a computational scheme for the evaluation of the factors of adjustment for clusters in the restrictive and general cases. Illustrates the approach with an analysis of National Assessment of Educational Progress data. (SLD)

Examined antecedents and moderators of antisocial behavior among children interviewed annually through high school and in early adulthood. Analyses identified four configurations differing in early school dropout and criminal arrests. School extracurricular activity participation related to reduced rates of early dropout and criminal arrest among high-risk children. Decline in antisocial patterns depended on social network participation in extracurricular activities. (Author/KB)

Examines various aspects of the motives, perceptions, and expectations connected with the introduction of Internet banking in Danish retail banking. Responses from questionnaires and results from a factor analysis and a hierarchical cluster analysis indicate a belief that Internet banking will become more important in the future. (Author/LRW)

Discussion of models for information retrieval focuses on an application of linear algebra to text clustering, namely, a metric for measuring cluster quality based on the theory that cluster quality is proportional to the number of terms that are disjoint across the clusters. Explains term-document matrices and clustering algorithms. (Author/LRW)

Using cluster analysis, a study empirically divided 54 children with attention-deficit/hyperactivity disorder (ADHD) based on their Full Scale IQ and reading ability. Clusters had different patterns of cognitive, behavioral, and neurochemical functions, as determined by discrepancies in Verbal-Performance IQ, academic achievement, parent aggression ratings, and a measure of noradrenergic function. (Contains references.) (Author/CR)