Analysing large-scale assessment data скачать в хорошем качестве

Analysing large-scale assessment data

Session chair: Prof. George Leckie, Professor of Social Statistics; Co-Director, Centre for Multilevel Modelling, School of Education, University of Bristol Issues with using police data to investigate offending: A research perspective Dr Ioana Crivatu, Research Fellow at the University of Birmingham Dr Ruth Spence, Senior Research Fellow at Middlesex University Police data is an important source of information for researchers about investigations, suspects, and victims. However, crime records can be problematic to work with. Here we outline three key issues along with our approach in combining and quantitatively analysing police data from several police forces in England and Wales which used different crime recording systems. We discuss data quality, which reflects missing and misclassified values; inconsistency, which refers to the vague and at times different definitions provided; and granularity, which reflects the lack of detailed information included in the datasets. We recommend developing a robust strategy for working with missing data, triangulating across different sources, creating higher-order categories where necessary, and creating a detailed data governance plan before analysis begins. Link to published paper: https://journals.sagepub.com/doi/full... Preparation of a Large-scale Assessment in Education and its use in a Quantitative Intersectional analysis in R Dr Natalia López-Hornickel, Postdoctoral Research Associate at Roehampton University; SWDTP alumni In this presentation, first, I aim to show the considerations and challenges of preparing large-scale assessment data, using the International Civic and Citizenship Education Study (ICCS) from 2016. This includes the sources of the data and the merging process, which is usually an overlooked but crucial step before proceeding with the analysis. Second, I will refer to the analysis steps to obtain descriptives and models. Particularly, I will use the case of the Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) to develop an intersectional analysis of students’ endorsement of the gender equality scale (Fifth paper of my thesis). This technique is a parsimonious alternative to multiplicative terms in regressions. All the explanations will be conceptual and also accompanied by a description of some R syntax.