Intended to serve as a `bridging' course between the basic data collection and analysis modules and a wider range of short courses dealing with particular data analysis and statistical approaches. It would be taken between such courses, and would need to be scheduled accordingly.
This is envisaged as a key course to discuss the key assumptions of the multiple linear regression model and the kinds of diagnostic tools available. It will provide a grounding in the statistical approach to analysing social science data.
A key objective is to provide a brief survey of the range of other statistical methods available, to enable informed choices about other courses.
A variety of software packages would suffice, but (given current site licences) it is envisaged that SPSS would be the main package used.
Learning Outcomes
Students will develop skills in:
learning & understanding the assumptions required for the linear regression model to have the best linear unbiased estimators
the importance of looking at a range of diagnostic information (particularly relating to examining the residuals) and the dangers of over-reliance on some popular summary statistics
be able to critique existing research, and produce regression results of their own
how this model may be extended to logistic regression where the dependent variable is categorical, how the approach differs, and alternative model specifications (`probit') and extensions to multi-category dependent variables with and without an `order'
show awareness of other statistical approaches that draw on the techniques
Assessment
21873-01 : 2,000 word data analysis report : Coursework (100%)