Gwybodaeth Modiwlau

This course includes an introductory lecture in each of at least six advanced statistical topics. From this set of topics the students choose two to study in more detail. The detailed study involves completing a workbook for each topic, for which the statistical programming language R is required.

The second part of the course is on statistical consultancy. Students will be exposed to a range of different data sets. This section of the course is taught in a series of seminars, where group discussion and critical reflection are encouraged. Students are expected to assess the validity of the data collection methods, discuss possible approaches to analysis and think critically about what inferences may meaningfully be drawn from the available data. This includes discussions of published papers from various statistical fields.

The topics available (introductory lecture followed by computer-based workbook) include: principal component analysis, statistical epidemiology, multivariate ANOVA (MANOVA), time series, multifactorial ANOVA and generalized linear models (GLMs). The two-hour introductory lectures take place in the first three weeks of the semester. Each workbook is associated with three two-hour computer practical sessions spread out over three weeks. The statistical programming language R is used for these workbooks. Each student will need to choose two of the workbooks to complete, and the marks for the workbooks will constitute the major part of the assessment for the module.

The statistical consultancy is presented as a set of weekly, one-hour seminars. Data sets are presented by postgraduate students and/or staff members, and through group discussions and individual reflection, students gain an understanding of how to approach new data sets and how to critically appraise different approaches to data analysis. Some of the seminars consist of discussions of peer-reviewed statistical articles, chosen to expose the group to contemporary issues in statistics and encourage critical thinking about statistical methodology.