Duncan Guest is a Senior Lecturer (Associate Professor) in Psychology at Nottingham Trent University (NTU) and Programme Leader of the BSc Psychology course there. Prior to arriving at NTU he worked in the Institute of Marketing and Communications at the University of Lugano, the Marketing Department of Bocconi University and Psychology Departments at Oxford Brookes University and Warwick University.
His research interests span cognitive psychology and consumer psychology. Within cognitive psychology he specialises in visual perception, exploring the decisional and perceptual processes (often with mathematical models) underlying tasks including perceptual categorization, object identification and visual search. In consumer behaviour he explores how current ideas in cognitive psychology (and psychology more broadly) can help us understand consumer behaviour. This work comprises a broad range of topics including; how people think about relations between products and brands; how properties of linguistic information (e.g., valence, arousal, spatial association) contained in branding (e.g., brand names or slogans) influences attention and perception; how co-created value in services can be influenced by psychological factors (e.g., psychological distance).
The overall aim of this workshop is to help participants enhance their understanding of statistical approaches used to analyse data from experiments both at a theoretical level and a practical level by giving the participants the skills required to carry out these analyses. The focus will be on a robust and flexible statistical technique that is used in a great variety of research called Analysis of Variance (ANOVA).
What is ANOVA? In experimental settings, we typically have an outcome measure (e.g., how well people can remember a brand name) and we manipulate another variable in order to measure its effect on this outcome variable (e.g., manipulate whether people saw this brand name in an advertisement in the morning, afternoon or evening). Broadly speaking, ANOVA looks at the data for these different groups (e.g., morning group, afternoon group, evening group) and tries to establish the extent to which any variance in the data was caused by the experimental manipulation. In other words, the main aim of ANOVA is to establish whether experimental manipulations have any observable effects.
ANOVA is a powerful tool and is widely used across a number of fields. It is therefore essential for anyone who will be designing experiments in their research to understand the analysis and be skilled in using it. This course will explore the broad theoretical underpinnings of the method and a major part will involve putting that theory into practice by learning how to compute ANOVA using IBM SPSS (a statistics package).
Throughout the course the format of the sessions will be to first go over the theoretical basis for a particular analysis and then spend time learning how to do that analysis in IBM SPSS using pre-existing data sets. Finally, we will discuss how to write up each analyses for publication.
The course will start by considering what circumstances we use ANOVA in (what kind of designs with what kind of data). It will then introduce the t-test, the understanding of which is a pre-requisite to understanding ANOVA. It will then consider how ANOVA differs depending on the design; whether the design is between subjects or within subjects and whether the design has multiple variables which are manipulated (factorial ANOVA). Many experiments studies are designed to see how multiple variables interact (influence) with each other. As such we will spend time considering interactions in factorial ANOVA, what they are, and how they can be broken down and understood (e.g., through simple effects analysis).
Toward the end of the course we will start examining more complex forms of ANOVA, including; Analysis of Co-Variance (ANCOVA) which can account for other sources of variance between participants that are not due to the experimental manipulation (e.g., participant age); and Multivariate Analysis of Variance (MANOVA) whereby the design has multiple outcome variables, Depending on time we may also look at more complex issues such as comparing the use of ANOVA to alternative data analysis techniques and exploring the debate around null hypothesis significant testing.
The following are books that I like. However, there are a large number of books out there and understanding statistics is partly based upon you finding a book that speaks to you about statistics at a level that you understand. As my background is in Psychology, these books also reflect that. A good book in terms of understanding the theory behind statistical tests is
The most enjoyable book on Statistics and SPSS (in my opinion) is
Given the above, I suggest reading a couple of chapters at the beginning of any statistics book to refresh you and to attend the Preliminary workshop (SPSS/Statistics Refresher)