Module 06-23856 (2018)

The aim of the module is to provide an introduction to the use of experimental design and statistics for the purpose of investigating human behaviour. The module is targeted at computer scientists with an interest in (i) understanding empirical studies concerning human behaviour, including studies of cognitive, social, and economic behaviour, and (ii) designing and conducting empirical research into the interaction between people and computers. The module may be of interest to computer scientists who look to an understanding of human behaviour to help constrain the development of computational systems, including novel forms of social media, information visualisation, information retrieval system, decision support system, robotics, and dynamic control systems. The module focuses on the implications of methodology and statistics (through lectures) and the practical implementation of research methodologies on real world datasets. Students will learn about how to design experiments, how to analyse data (using a statistical programming language or package), and how to write evidence-based reports.

• Provide the student with knowledge and skills necessary to assess and conduct empirical research for HCI and computational science
• Give students practical and mathematical knowledge in basic statistical techniques

1. Identify and discuss research methodologies for investigating human behaviour
2. Recognise the appropriateness of statistical techniques in data analysis
3. Conduct and report a variety of statistical tests using appropriate software
4. Interpret research findings from a variety of statistical techniques to a high level
5. Discuss issues related to conducting research on human participants (sampling, recruitment, ethics etc)

None

1 hr lecture, 2hr tutorial/practical a week

Contact Hours: 33

Sessional: 1.5hr examination (80%), continuous assessment (20%). The continuous assessment is based on two class tests (10% each).

Supplementary (where allowed): examination only (100%).

1. Writing about scientific research; evidence based argumentation; traditional versus rational authority; claims, data, warrants, and qualifiers; good science / bad science; what counts as a good theory.
2. The use of the R programming language for inferential statistics.
3. Hypotheses. The null hypotheses. Type I and type II errors; data types: nominal, ordinal, interval, ratio.
4. Plotting data, boxplots, graphical representations of dispersion and extreme scores; frequencies, distributions, measures of central tendency; measures of variability, standard deviation, variance, confidence; Central Limit Theorem; the importance of data screening.
5. Correlation; what is a correlation? What does it mean? The value of correlation; methods for computing a correlation statistic.
6. Inferential statistics: independent and dependent variables, comparing means, power analysis, p-value, significance, confidence intervals, effect sizes; the limitations of inferential statistics.
7. Regression. What is a regression analysis? Calculation of model fit. Residuals.
8. t-tests; What is a t-test? Why use t-tests? Between and within subjects design.
9. ANOVA; Why use ANOVA? How to calculate an analysis of variance; Multiple Independent variables (2 way ANOVA); interaction effects.
10. Alternatives to inferential statistics will also be discussed.