ľ¹Ï¸£ÀûÓ°ÊÓ

Technological developments such as smartphones have enabled data on emotions, thoughts and behaviour to be collected in real time. Although these data provide a wealth of information about the dynamics of time in psychopathology, one of the greatest hopes in digital mental health care has yet to be realised: a smart real-time personalised alert system for signalling clinically relevant changes in symptoms. In this pilot study, Emmeke Aarts is developing and validating a new dynamic statistical algorithm that should enable a smart personalised digital predictive tool for mental health.

Researcher:  Emmeke Aarts 
Project: Towards a real-time personalised alert system in mental health care

Low literacy is a major problem, but it’s difficult to identify people who have low-literacy abilities. The current testing methods are very intense and intimidating for the target group, because they require reading – something very uncomfortable for them. However, most people feel more at ease with spoken language. The central question in the proposal by Hans Marien is whether it’s possible to detect low literacy through speech. This would avoid uncomfortable situations and improve the detection of low literacy. Marien is using speech-to-text technology to research the extent to which people are able to recognise their own spoken words in text form and whether recognition scores predict their literacy level.

Researcher: Hans Marien
Project: Detection of low-literacy: Assessing visual word recognition with Speech-to-Text technology

Further research into the attitude of various populations to homosexuality has produced a puzzling finding: in countries where homosexuals are frequently discriminated against, the respondents are not aware of this discrimination. In contrast, respondents in more tolerant countries report discrimination more often. The way in which respondents interpret the questions may possibly differ between and within countries, making surveys potentially misleading.
In her project, Katharina Meitinger combines web probing with small area estimation to explain regional differences in attitude and interpretation. The results will help policymakers to interpret survey results and will provide essential input for improving anti-discrimination measures.

Researcher: Katharina Meitinger
Project: Regional validity of survey questions on attitudes towards homosexuality in Europe: combining web probing with small area estimation

Teachers are the first authority figures whom young people encounter outside their home environment. The way that teachers and young people interact helps to determine their attitude to authority within and outside school. Through these interactions, young people develop their attitude towards rules and legal authorities such as the police. The details of how this learning takes place are still not well understood. In her research project, Amy Nivette observes young people and their teachers as they discuss a rule-breaking situation and she develops a coding system to identify frequently applied strategies that may lead to young people accepting and following rules. 

 Researcher: Amy Nivette 
Project: Measuring the micro-processes of legal socialisation during interactions between teachers and students

Forming part of the group and acquiring social status is very important to adolescents. Risk behaviour such as smoking and use of alcohol and cannabis is often influenced by the degree to which this behaviour increases popularity status. The hierarchy in the class (popularity, acceptance, rejection) can result in stressful situations in which adolescents feel the need to increase or defend their status. These stressful situations can influence and perpetuate risk behaviour. In her project, Margot Peeters aims to understand problematic risk behaviour as a consequence of popularity status and associated stress responses.

Researcher: Margot Peeters
Project: Stress and popularity status as important pathways in understanding adolescents’ problematic risk behavior

Data-driven analysis techniques are being used increasingly frequently in the field of data science and AI. The application of these techniques requires a high level of specific knowledge. However, if the data are incomplete then more statistical expertise is required and correctly specified imputation models are needed in order to draw valid conclusions. In his project, Gerko Vink presents ‘mindless imputation’: a data-driven automatic imputation technique with minimal user intervention. Applied researchers can use mindless imputation to focus their attention on the intended analysis. Mindless imputation enables the use of modern analysis methods for obtaining valid inferences and reliable predictions on the basis of incomplete datasets.

Onderzoeker: Gerko Vink
Project: Mindless Mice: Multiple imputation without explicit model specification