XS grants from SSH Open Competition for six FSBS researchers

Low literacy is a serious problem, yet it is difficult to find people who are functionally illiterate. The target group finds the current test methods to be quite drastic and intimidating because these methods require reading, which is extremely uncomfortable for them. By contrast, most people are proficient in spoken language. This proposal centres on the question of whether it is possible to detect low literacy through a person’s speech. This would make the process less embarrassing and increase the low literacy detection rate. Using Speech-to-Text technology, this project explores the extent to which people are able to recognise their own spoken words in text form and whether recognition scores are an accurate predictor of their literacy level.

Earlier research into the attitudes of various population groups towards homosexuality yielded an enigmatic finding: in countries where anti-homosexual discrimination is prevalent, the respondents are unaware of that discrimination. Respondents in more tolerant countries, on the other hand, tend to report discrimination more often.

Because the way in which respondents interpret the questions may vary from one country to another, questionnaires can potentially be misleading. This project combines web probing with small area estimation in order to explain regional variations in attitude and interpretation. The results will provide policymakers with a starting point for interpreting the survey results and supply essential input for the improvement of measures to combat discrimination.

Teachers are the first authority figures young people encounter outside their respective home environments. How teachers and young people treat one another helps to shape the attitude of those young people towards authority, both in and outside school. Through these interactions, pupils develop their individual attitudes toward rules and legal authorities such as the police. The details of how that learning takes place are not yet clearly understood. This project observes young people and their teachers as they discuss a situation involving rule-breaking. It also aims to develop a coding system for identifying common strategies that may encourage young people to accept and comply with the rules.

Belonging to a group and acquiring social status are extremely important to adolescents. Risk behaviour such as drinking alcohol, smoking and using cannabis is often influenced by the extent to which these behaviours increase one’s popularity status. The social hierarchy at school (popularity, acceptance, rejection) can result in stressful situations in which young people feel a need to increase or defend their status. These stressful situations may influence and promote ongoing risk behaviour. This project strives to understand problematic risk behaviour as a consequence of popularity status and the associated stress responses.

Data-driven analysis techniques are increasingly being used in the field of data science and AI. The application of these techniques requires a great deal of specialised knowledge. And when the data is incomplete, the amount of statistic expertise required increases and correctly specified imputation models are needed in order to draw valid conclusions. This project suggests mindless imputation: a data-driven automatic imputation technique with minimal user intervention. With mindless imputation, applied researchers will be able to focus their attention on the intended analysis. Mindless imputation will make it possible to use modern methods of data analysis to obtain valid inferences and reliable predictions based on incomplete data sets.