Search results

The purpose of the study is to propose a normative approach for market segmentation, profile and monitoring using computing and information technology to analyze User-Generated Content (UGC).

The specific steps include performing a structural analysis of the UGC and extracting the base variables and values from it, generating a consumer characteristics matrix for segmenting process, and finally describing the segments' preferences, regional and dynamic characteristics. The authors verify the feasibility of the method with publicly available data. The external validity of the method is also tested through questionnaires and product regional sales data.

The authors apply the proposed methodology to analyze 53,526 UGCs in the New Energy Vehicle (NEV) market and classify consumers into four segments: Brand-Value Suitors (32%), Rational Consumers (21%), High-Quality Fanciers (26%) and Utility-driven Consumers (21%). The authors describe four segments' preferences, dynamic changes over the past six years and regional characteristics among China's top five sales cities. Then, the authors verify the external validity of the methodology through a questionnaire survey and actual NEV sales in China.

The proposed method enables companies to utilize computing and information technology to understand the market structure and grasp the dynamic trends of market segments, which assists them in developing R&D and marketing plans.

This study contributes to the research on UGC-based universal market segmentation methods. In addition, the proposed UGC structural analysis algorithm implements a more fine-grained data analysis.

The OSCE is regarded as the gold standard of competence assessment in many healthcare programs, however, there are numerous internal and external sources of variation contributing to checklist marks. There is concern amongst organisers that candidates may be unfairly disadvantaged if they follow an “excellent” preceding candidate. This study assessed if average checklist scores differed depending on who a candidate follows accounted for different sources of variation.

This study examined assessment data from final year MBChB OSCEs at the University of Aberdeen and categorised candidates into three levels dependent on examiner awarded global scores of preceding candidates for each station. Data were modelled using a linear mixed model incorporating fixed and random effects.

A total of 349 candidates sat the OSCEs. The predicted mean (95% CI) score for students following an “excellent” candidate was 21.6 (20.6, 22.6), followed “others” was 21.5 (20.5, 22.4) and followed an “unsatisfactory” student was 22.2 (21.1, 23.3). When accounted for individual, examiner and station levels variabilities, students following an “excellent” candidate did not have different mean scores compared to those who followed “other” (p = 0.829) or “unsatisfactory” candidates (p = 0.162), however, students who followed an “unsatisfactory” student scored slightly higher on average compared to those who followed “other” (p = 0.038).

There was weak evidence that candidate's checklist variations could be attributed to who they followed, particularly those following unsatisfactory students; the difference in predicted mean scores may be of little practical relevance. Further studies with multiple centres may be warranted assuring perceived fairness of the OSCE to candidates and educators.