David Keatley

Studying behaviour outside of the laboratory is often difficult as information may be incomplete, idiosyncratic, fragmented, partly qualitative, and rapidly changing. Developing methods to detect and analyse the non-random, syntactic structure of behaviour allows researchers to understand and predict future behaviours. This symposium will provide a series of talks centred around Sequence Analysis methods, which allow the progression of real-world, complex behaviours (both animal and human) to be mapped and understood. The symposium will begin with a brief outline of the basic approach of Lag-Sequence Analysis (LSA), aimed at those with no prior experience or knowledge of Sequence Analysis research. This talk will also focus on Indicator Waves, a novel method of sequencing multiple, concurrent behaviours across time. The second talk will explain the use of proximity coefficients to analyse the interrelationships among sequences of behaviours within and between cases. Using examples from law enforcement interactions, the talk will show how this approach opens up conventional inference testing across variables that distinguish different sequences. The final talks will outline the T-System Approach in Sequence Analysis. The T-System is a formally defined set of probabilistic patterns of relations between behaviours or events on a discrete time scale. This may seem quite similar to LSA; however, the T-System approach allows for much more sophisticated, hierarchical analyses of sequential and concurrent behaviour over time. These complex T-System analyses will be clearly explained to show how hidden behavioural relationships can be detected, which would be impossible with traditional LSA methods. The T-Systems talks will include novel, extended analysis of T-Pattern sets for the new T-prediction and T-retrodiction. In addition, the newly updated THEME software, with increased speed of large dataset analysis, will be outlined, with applications to big data. Finally, drawing from the literature, examples will be provided on how the T-System approach can be applied to research in human and non-human samples. The symposium will end with a discussion and evaluation of which method is best for particular research questions or approaches, and the future of all these methods will be discussed.

Objectives: The main aim of this paper is to understand the pattern of actions involved in offending behaviours, in general. A particular case will be provided: A sequence analysis to describe the timeline of events between individuals deciding to consume alcohol, and then drive a vehicle. Design: The study used a sequence analysis design, in which a series of actions related to drinking alcohol and driving were presented and individuals’ patterns of behaviours investigated. Method: The method involves taking a series of recorded interviews into driving whilst under the influence of alcohol. The current sample of interviews (N=60) were transcribed and analysed until a mutually exclusive and exhaustive list of behaviours was agreed upon. Frequently occurring behaviours were then analyse for significant transitions. Results: Results showed clear patterns of actions in individuals’ choices to consume alcohol and drive a vehicle. A lag one sequential analysis approach was used (i.e., a-b, b-c, c-d). This analysis involves transitions between only the one preceding behaviour (antecedent), to give a chain of events. Transition maps showing the sequence of events were developed to clearly show the patterns of behaviours that lead to driving under the influence of alcohol. Conclusions: The present research highlights the sequences of actions and choices leading up to driving under the influence. The method has implications, which will be discussed, for policy and policing, as well as other related driving offences (i.e., drugs etc.). Limitations related to analysis are discussed and guidance provided. Relation to other areas of Forensic Psychology (i.e., rape cases, violence) will also be outlined.