Doing data analysis: Some tools and techniques

Paul Collett, Trevor Holster
Saturday, 14 September 2013 - 6:30pm to 8:00pm

Session 1:

Introduction to the R statistical package
Paul Collett

In this workshop, we will introduce the R statistical package, which is a free software environment for statistical computing and graphics (http://www.r-project.org). It is a powerful, popular alternative to such commercial statistical packages as SPSS, with a large user base and extensive documentation and support.
After outlining the basis of obtaining and installing the package we will demonstrate some applications using real data sets to show how it can be used for different types of data analysis, such as generating basic descriptive statistics, correlations, scale validation, and multivariate analysis like cluster analysis, principal components analysis or factor analysis.
We will finish up with a brief look at some of the substantive issues and critiques related to quantitative statistical analysis.

Biodata: The presenter teaches and is involved in ongoing research at Shimonoseki City University.

Session 2:

Objective measurement: An introduction to Rasch analysis
Trevor Holster
Fukuoka Women's University

Although raw percentage scores are commonly used for classroom assessment, research instruments and standardized tests require more sophisticated analysis. Georg Rasch developed a model of "objective measurement" based on the simple conceptual insight that responses to test questions follow probabilistic patterns rather than machine-like deterministic patterns of success and failure. Higher ability persons have a higher expectation of success on tasks, but the assumption of Rasch models is that sometimes low ability persons succeed and high ability persons fail. Linacre extended Rasch's original model to allow analysis of other facets of performance, such as rater severity in judged performance tests, and this is now one of the most widely used techniques in analyzing language performance tests. This presentation will briefly review Rasch's simple concept of "specific objectivity" and then demonstrate Rasch analysis using data from a multiple-choice vocabulary test and from judged ratings of classroom presentations. Audience members who wish to try the analysis themselves will be provided with the datasets and the free student versions of the Ministeps and Minifacs software packages (for Microsoft Windows operating systems only).

Biodata: Trevor Holster has a Master’s degree in TESOL, specializing in curriculum design and materials writing. His research interests include the integration of classroom assessment with instruction, performance based testing, peer-assessment, and placement testing. He is the Membership Chair of the JALT Testing and Evaluation SIG.