Doctoral course - Graduate Schools MADIS / ENGSYS (Fall 2022)

Experimental research and statistical methods
for Human-Computer Interaction

Description

This special topics graduate course is an introduction to experimental research that enables the identification of causal relationships through the design of controlled experiments with participants. The course covers fundamental concepts to conduct successful experimental research through appropriate experimental designs to answer clearly defined research hypothesis. The course will cover major types of experimental designs (within-group, between-group and split-plot designs) with their pro and cons. The course will also adress the statistical analysis of the data collected during such experiments. Main parametric (e.g. ANOVA) and non-parametric (e.g. Wilcoxon) tests will be covered and illustrated using R. The concepts will be tackled through lectures, and tutorials in R and Rmd will help mastering them with examples. The course will be illustrated with numerous examples from the domain of Human-Computer Interaction (HCI). It is fundamental for any student working in HCI and any student willing to design and analyze experiments with human participants.

Keywords

Experimental research, experimental design, statistical analysis, R.

Prior knowledge

No specific knowledge in statistics ; knowledge in programming to start more easily with R.

Expected outcomes

• Concepts of experimental research,
• Understanding of experimental designs and ability to choose the right one to answer a research hypothesis,
• Understanding of existing statistical tests to analyze experimental data collected during such experiments.

Courses

• Experimental research
• Experimental design
• Stats [Examples from the course]

Tutorials

• T1: Hello R and RStudio [solution]
• T2: Data visualization [solution][access to paper][Use python and R in Rmd]
• T3: Parametric tests [solution]
• T4: Non-parametric tests [solution]
• T5: Conduct your own analysis [solution]