17 April 2024
- Speaker: Professor Markus Bindemann, University of Kent
- Title: Identifying people we don’t know
- Time: 12.15pm - 1.05pm (GMT)
- Location: 10W 1.01
Some important tasks require the identification of unfamiliar people, who are not known to the observer. In Psychology, this is often studied with unfamiliar face-matching paradigms, in which observers must compare face images under highly controlled conditions. In this talk, Professor Bindermann will explain how we are applying virtual reality (VR) to study this problem in more complex contexts, focusing on the construction and validation of digital humans (avatars). I will also link this topic to automatic face recognition, using a case study from WW2.
24 April 2024
- Speaker: Adam Harris, University College London
- Title: The prevalence and consequences of the severity effect in probability communication
- Time: 12.15pm - 1.05pm (GMT)
- Location: 10W 1.01
Effective decision making relies on reliable assessment of probabilities. We receive clues to these probabilities from expert sources. Verbal communications are susceptible to a variety of pragmatic influences, and in this current talk we focus on the Severity Effect, whereby numeric interpretations of verbal probability expressions are higher for more severe events. The research has implications for likely amplification of perceived risks.
1 May 2024
- Speaker: ProfessorRichard Rowe, University of Sheffield
- Title: Modelling a general factor in developmental psychopathology: A multi-informant approach
- Time: 12.15pm - 1.05pm (GMT)
- Location: 10W 1.01
A growing body of research supports the utility of modelling a general psychopathology, or p-factor. The p-factor has been found to underlie multiple forms of mental ill health across development, including both internalising problems, such as depression, and externalising problems, such as hyperactivity. A number of theoretical accounts have been proposed to explain the p-factor, including that it may represent a dimension of cognitive or emotional dysfunction. Previous work has focussed on defining the p-factor on the basis of data from a single reporter, meaning that common method variance may contribute to modelled p-factors. This project characterises the p-factor based on combinations of self-, parent- and teacher- reported Strengths and Difficulties Questionnaire data. Models were constructed using the Mental Health of Children and Young People (MHCYP) Study, separately in children aged 5-10 and 11-15 and the Millennium Cohort Study (MCS) when members were aged 7. The results identified rater-specific and rater-general p factors. Longitudinal analyses identified p-factor outcomes including internalising and externalising disorders 3 years later in the MHCYP dataset and emotion problems, conduct problems and attempted suicide 10 years later in MCS. The MCS also identified age 5 predictors of the p-factor including bullying, family income and parental mental health problems. Overall, the findings confirmed the utility of the p-factor in multi-informant data.
2 May 2024
- Speaker: Laura Güdemann, University of Exeter
- Title: Causal treatment effect estimation using provider preference based Instrumental Variables - Applications in type 2 diabetes research
- Time: 12.30pm - 2pm (GMT)
- Location: 10W 1.01
Healthcare research using observational data has the potential to uncover valuable evidence for example for patient groups which are commonly excluded from clinical trials. But the quality of observational evidence depends on the application of suitable analysis methods which can adequately address potential sources of bias in data not necessarily collected for research purposes.
Instrumental Variable approaches provide a way of addressing bias due to unmeasured confounding when estimating treatment effects in observational data. One type of instrument is a proxy of healthcare provider prescription preference. In her talk, Laura Güdemann will describe a framework for the application of this type of instrument and discuss important methodological challenges for the application of preference-based instruments. Her talk will outline theoretical background in causal inference as well as application studies in Type 2 Diabetes research.
12 June 2024
- Speaker: Professor Dorothy Cowie, Durham University
- Title: TBC
- Time: 12.15pm - 1.05pm (GMT)
- Location: 10W 1.01