Development and validation of the Visualize image-based assessment of the Five-Factor Model of Personality
This project aims to develop and validate the Visualize behavioural assessment, the only instrument that integrates text and images to measure the Big Five model of personality. Using computer vision techniques to improve the construct validity and with the support from a team of EDI specialists, the images used in Visualize were selected to promote diversity, multiculturalism and inclusion. The Visualize items complement the mainstream perspectives on assessing personality traits as they not only measure affects, behaviours, cognitions and desires using sentences, but create visual scenarios that consider contemporary topics in society, including the use of social media, technology adoption, social, cultural and environmental awareness, and ethical and social citizenship.
As a result of the inclusion criteria, a combination of intra-individual response variability, identical consecutive responses, response time and social desirability bias scale, we collected data on 1,055 workers, who responded to eight blocks of items at different points in time, totalling 1,200 items measuring 108 facets. The items, dataset and R scripts will be made publicly available to researchers in three languages (English, Spanish and Portuguese) through the Open Science Framework to facilitate reproducible research. The construct validity studies are being carried out using machine learning algorithms, item response theory and Thurstonian item response theory.
The ultimate goal of this project is to deliver the first image-based adaptive behavioural assessment measuring the Big Five model of personality. Users will respond to fewer items with an adaptive platform, significantly improving test efficacy and measurement precision. By engaging with a shorter version, they are expected to complete it up to three times quicker than they would in a traditional assessment, improving the candidate experience and helping to increase the efficiency of recruitment processes.
This project is in collaboration with the School of Computing, National University of Singapore (University of Singapore), the Intelligent Vision Research Laboratory (Federal University of Bahia, Brazil) and Visiu Analytics Ltd.
Prospect theory and employee turnover propensity
Employee turnover has become one of the most important topics in organisational behaviour given its costs and consequent impacts on the organisation strategic planning. Costs of turnover vary, and depending on the employment sector, they can range from 20% to an enormous 213% of an employee’s annual salary among positions requiring a significant level of higher education and training, such as senior executives (Boushey & Glynn, 2012).
Despite the growing number of studies involving antecedents and measures of employee turnover intentions, to our knowledge, there are no behavioural models that use the prospect theory to investigate the employees’ intentions to quit considering scenarios that involve uncertainty and risk (e.g., 10% raise to stay in a job that I dislike vs New job I like for the same as my current pay).
As the decision to quit entails a trade-off between the perceived gains and losses of leaving the current job and moving to a new one, in this study decision weights are combined with antecedents of turnover to provide additional insights into the employee intentions to leave or stay. Based on a three-part longitudinal study, we will survey the same participants over the course of a year to explore what patterns emerge from the responses of those who decided to leave their job and those who decided to stay. Finally, a predictive model using the antecedent variables will be fit to estimate the employee turnover propensity, an index able to flag workers with prior-to-leaving mindsets that can help organisations develop appropriate retention policies.
This project is in collaboration with the Future Of Global Work And Leadership Institute (University of Hull), the People Analytics and Psychometrics Laboratory (University of Hull) and Department of Health Policy & Management (Columbia University).
Development and validation of the Agile Team Dynamics instrument
Agile Team Dynamics (ATD) is a network approach to team dynamics that measures the psychological forces that come into play while individuals interact and work as a team. It captures the individuals’ perceptions of their team, focusing on collective aspects beyond individual aspects such as personality and individual skills. The ATD was developed to provide an in-depth analysis of the team dynamics, including three dimensions and nine sub-dimensions:
- Psychological safety: safe to learn, safe to be different, safe to challenge
- Growth mindset: collective resilience, collective outsight and collective learning
- Connectedness: identification, cohesion and helping behaviours
By proposing an integrative approach to team dynamics, the ATD considers teams as complex systems that emerge organically due to the individuals’ interactions with their different components. Psychometric network models are used to explore the network structure and lay the groundwork for building a network theory of team dynamics.
This project is in collaboration with Ollintel Ltd.