Toolkit for Modern Psychometrics and Statistics

Facilitator
Helen Baron

Who should attend
This workshop is designed to provide an introduction to some modern analytic techniques. It would be particularly useful for two groups of people. Firstly, for those wishing to bettercritically understand the current psychological literature or evaluate new psychometric instruments developed using these techniques. Secondly, for those who are considering using any of the tools discussed for developing measurement scales or for more general research purposes. It assumes only a basic level of familiarity with simple statistical techniques (e.g. correlation) and does not require any knowledge of statistical equations or algebra.

Workshop overview
Statistical techniques are continually developing. This workshop will provide an introduction to some of the techniques that are becoming necessary to understanding the psychological research literature, but are typically beyond the scope of many statistical courses for psychologists. They are powerful additions to the analyst’s toolbox and enable a better understanding of trends in data.

The workshop covers two techniques:
1. Structural Equation Modelling (SEM), which is a synthesis of path and factor analysis and has both theoretical and practical applications. For example, it can be used to explore exercise and competency effects in an assessment centre (confirmatory factor analysis) or to look for causal relations between theoretical constructs.
2. Item Response Theory, which is revolutionising test development procedures. It enables a better understanding of how tests and questionnaires are working and enables applications such as item banking and adaptive testing.There are now easily available commercial tests, which use Item Response Theory in development and/or scoring and norming. Researchers may also find the techniques useful in developing and calibrating research instruments.

This is a non-technical presentation focusing on understanding what the techniques can add to our understanding of psychological data and variables in a range of applications from test development to developing psychological theory. It will enable participants to apply the technique in the contexts of their own research interests. The approach is one of developing understanding and will not go into the detail of running computer programs for these techniques. There will be a mix of presentation and small group exercises.Participants will be encouraged to work on their own variable sets paradigms in the exercises wherever possible, but content will be provided for those that prefer this.

Aims of the workshop

• To develop knowledge of relevant statistical techniques which are new and/or rarely
• covered during training.
• To enable participants to understand the growing part of the research literature which
• uses these techniques.
• To enable participants to know when these techniques might be usefully applied.

Projected outcomes and benefits of attending
By developing an understanding of these important techniques participants will be able to critically evaluate research and development based on these techniques and understand when it is appropriate to apply them to their own or other’s research. They will also be ready to start learning how to actively use the techniques for analysis if this is relevant for them.

Psychological theory underpinning the workshop
Structural Equation Modelling is a powerful statistical technique that allows a confirmatory approach to be taken to data analysis in many situations. It combines elements of regression, measurement theory and factor analysis. It is now used frequently in all areas of psychological research, such that it can be difficult to read the research literature without an understanding of SEM. Many researchers have helped develop the theory and technique, but the most important contributions are from Joreskog (1970) and Bentler (1980). Item Response Theory provides a compelling set of tools for the test developer, which
allows greater understanding of measurement tools and enables applications such as tailored testing and item banking. Based on the work of Birnbaum (Lord & Novick, 1968) and Rasch (1960) the theory has more recently begun to be used extensively in applied psychometrics. Hambleton (van der Linden & Hambleton, 1996) has been highly instrumental in popularising its use along with the easy availability of computing power.

Bentler, P.M. (1980). Multivariate analysis with latent variables: causal modeling. Annual Review of Psychology, 31, 419-456.
Joreskog, K.G. (1970). A general method for analysis of covariance structures. Biometrika, 57, 239-251.
Lord, F.N. & Novick, M.R. (1968). Statistical theories of mental test scores. Reading: MA, Addison-Wesley.
Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests. Chicago: University of Chicago Press. van der Linden, W.J. & Hambleton, R.K. (1996). Handbook of modern item response theory. New York: Springer.

Pre/post work required
N/A

Date and venue
19 June 2008, 09.30 - 17.00.
The British Psychological Society, 30 Tabernacle Street, London, EC2A 4UE.

Facilitator details
Helen Baron is a Chartered Psychologist and an Associate Fellow of the BPS. She has over 20 years experience in research,development and application of psychometric tests,including both cognitive and personality related measures. Her research interests include equal opportunities and cross cultural work. She has international experience in adapting instruments appropriately for use in other cultures. She now continues with this work as an independent consultant. She is an experienced trainer in statistical and test development methods as well as using these techniques in her own work.She is a member of the BPS Standing Committee on Test Standards. She has published widely in the area of psychometric test development and is a regular speaker at conferences.

Registration is now open

Reg Form: Baron 19.06.08

or book on-line

The Division of Occupational
Psychology is part of:
The British Psychological Society