The Exploration-Exploitation Trade-Off during Divergent Thinking on an Alternative Use Task
This study explores internal divergent thinking with the Alternative Use Task. The main aim is to create a scoring system for exploration-exploitation trade-off within participants. Internal divergent thinking can
be described with the marginal value theory, where one is expected to exploit a category of related terms until an optimal moment and then start exploring alternatives for a new category. An alternative theory involves semantic similarities between terms. The reaction sequence is expected to be created by pursuing small steps in semantic distance. The overlying component between these theories are the reaction times. In both theories the reaction times are used as confirmatory analysis, either in relation to exploration-exploitation jumps, or related to semantic distances. In this study we combine these theories to investigate strategies and their trade-offs with the data of over 2000 students on divergent thinking tasks.
Fleur Doorman & Claire Stevenson
Amsterdam AUT (A-AUT) – Dutch AUT database and automated scoring algorithm
In this project, we created the Amsterdam Alternative Uses Task (A-AUT). This task is based on the original, and widely popular, Alternative Uses Task from Guilford (1950). Some of the common issues of the AUT are the lack of consistency between scorings, and the heavy time investment of manual scoring. In our current project, we extend the existing task by bringing together big data and psychometrics, with the aim to create an automated scoring algorithm for the AUT. To achieve this, we are combining AUT data from various Dutch researchers into one large database and create an algorithm that can identify semantically similar responses. Ultimately, based on the existing dataset, the algorithm will be able to automatically identify if a new response is similar to an existing one in the database and assign the appropriate score to the new response. This way, we can increase the reliability of its creativity scores as the same responses will receive the same score.
For a first pilot of this algorithm, please see the previous project “scoring algorithm for creativity”.
Dr. Claire Stevenson, dr. Matthijs Baas & Han van der Maas
A-AUT Validation Study
This project is part of the A-AUT (Amsterdam Alternative Uses Task)-project. The generated algorithm automatically scores responses to the A-AUT to reduce the usual time-intensive (and often variable) manual scoring by experts. This validation study aims to collect a new set of data on the A-AUT and similar tasks (i.e., verbal fluency task, K-DOCS, creative achievement questionnaire, Remote associate task) to, first, validate the accuracy of the algorithm by comparing the algorithm scores to those of expert raters and people’s ratings of their own creativity. Second, to test the reliability of the algorithm by comparing the test-retest and alternate form scores of the algorithm for different AUT-items.
Dr. Claire Stevenson, dr. Matthijs Baas, Iris Smal & Maike Dahrendorf
Automated categorization of AUT responses
Animals forage for food by moving through different patches of resources, being, for example, a patch of ants or a patch of fruits in a tree. Humans’ are thought to forage through their cognitive memory in a similar way, moving from one patch of information to another. Prior research suggests that in humans’ these patch switches are made when one has exploited a patch of semantically similar information (Hills et al., 2012; Hills et al., 2015). One of the abilities that depend on memory retrieval is called ‘creative ideation’, which is the ability to create novel useful ideas based on prior stored information within the cognitive memory. Based on this knowledge it is probable that creative ideation holds similar principals as animal foraging theories. Creative ideation is often tested by divergent thinking tasks, such as the Alternative Uses Task (AUT). To test animal foraging theories on AUT responses, patch switches need to be made based on semantic similarity. Therefore, AUT data needs to be grouped and classified based on semantic similarity. In previous research, this has been done either manually or by the use of an outdated model. In the current study, we aim to classify AUT responses automatically and compare machine versus human coders. You can download Emma’s thesis here.
Dr. Claire Stevenson & Emma Schreurs
Bayesian cognitive model of decision making in creativity
The first perspective considers the production of creative ideas as a decision process, where numerous ideas come to mind, but some are discarded along the way as they are not ‘useful’ enough, whereas other ideas – reaching some internal threshold – are expressed. The diffusion model, often applied to simple speeded decision tasks, can be adapted to examine this process. In this project, an adapted form of the diffusion model will be used to examine the trade-off that takes place when moving to and between the idea generation and evaluation phases of the creative process. Click here to see a video presentation of the results.
Michelle Donzallaz, Julia Haaf & Claire Stevenson
Is creativity domain-general or domain-specific? A systematic review of neuroimaging studies of creativity
Many neuroscientific reviews of creativity have tried in vain to answer the question of whether creativity is domain-general or domain-specific. Very little evidence of overlapping brain regions across creativity domains was observed by them, causing the neural mechanisms underlying creativity to remain poorly understood. Almost all these reviews base their results on studies that used divergent thinking or insight as measures of creativity. However, both divergent thinking and insight tests have been considered domain-general tests, which makes them a biased source when studying the generality-specificity dilemma. Therefore, the current review aimed to provide a systematic review of neuroimaging and neuroelectric studies assessing creativity from multiple domains, such as creative writing, visual arts and musical creativity, whilst deliberately excluding studies on divergent thinking and insight. A total of 41 studies were included in this review belonging to a variety of creativity domains, as derived from the creative achievement questionnaire (CAQ). The results of this review portrayed a highly variegated landscape of reported brain regions, both within and across creativity domains. Among the most consistent findings for neuroimaging studies was the involvement of brain regions belonging to the executive control network and the default-mode network for creative compared to control tasks, though lateralization differed between studies. Neuroelectric studies most consistently reported an association between frontal alpha-band activity in creative compared to control tasks, though the direction of activity (increased or decreased alpha activity) varied between studies. These findings are in line with results from studies assessing divergent thinking, insight problem solving and intelligence. Although methodological artifacts within and across creativity domains compromise drawing reliable conclusions, the outcome of this review suggests that a specific creativity network does not exist, but rather that creativity is dependent on general cognitive processes.
Fabienne van Rossenberg, Claire Stevenson & Matthijs Baas
Training the Brain for Creative Thinking: a Meta-Analytic Comparison of Neurostimulation and Cognitive Training Studies
Creativity research, whether fundamental or applied, often ultimately aims to “improve creativity”. However, we currently lack insight into the most efficient way to do so. Two lines of interventions emerge from the literature as especially promising: cognitive training and neurostimulation. The primary objective of this study is therefore to compare the effectiveness of cognitive training and neurostimulation programs in improving divergent thinking in a meta-analysis. For more details please see the OSF project page.
Hanna Haas, Claire Stevenson & Baptiste Barbot
Creativity Games for Children
We are working on an exciting new game for the Math Garden and Language Sea online learning platforms that aim to measure creativity. One of the games is inspired by verbal analogies (e.g., sheep:lamb, cow: ? ). The other game is inspired by Bongard problems. Bongard problems are puzzles where you are asked to find a pattern or a rule that holds for six geometric figures but does not hold for the other six figures. We are developing a child-friendly version where the children will be asked to point to pictures that ‘go together’ on a tablet or smartphone. This project is funded by the Jacobs Foundation.
Claire Stevenson & Naomi van Bergen
Reliability of divergent thinking tasks
We perform a meta-analysis regarding the reliability of divergent thinking tasks. Divergent thinking tasks are often used to measure creative potential and make claims about how creative abilities change over time or after an intervention. However, when you measure a person’s creative potential one day and then repeat this a few days later, the performance can be quite different. This implies that divergent thinking tasks are unreliable for longitudinal or intervention studies. Repeated assessment of creativity is of great interest for educators, developmental psychologists and those interested in the effects of interventions on creativity. For more details about the project, have a look at the Open Science Framework website.
Claire Stevenson, Hanna Haas, Lea Naczenski & Baptiste Barbot
Towards a minimal theory of creativity: Theoretical predictions tested with data from a large-scale adaptive learning platform
The minimal theory of creativity (MTC) is a theory that tries to explain creativity in the most logical and parsimonious way possible. It states that only intelligence and domain expertise are substantially related to creativity. In my research project, I seek to assess the plausibility of the MTC. For that purpose, specific predictions have been derived from the MTC and will be tested empirically. The two predictions are a) that intelligence and domain expertise are able to adequately predict creativity levels and b) that a nested model of intelligence + domain expertise model which also includes motivation will not perform any better than the basic intelligence + domain expertise model. To test these hypotheses, structural equation models will be computed and assessed regarding their goodness of fit, as well as compared to each other. The models will be fitted to an extract of a large database that stems from an online adaptive learning system called “Oefenweb”. Testing whether the predictions hold will allow for conclusions about the MTC and hopefully give basic grounds for further endeavors related to the investigation of the MTC.
Niklas Frerichs & Claire Stevenson
Optimal foraging strategies in semantic search
The second perspective views the creative ideation process as a search process for possible solutions through long term memory – comparable to how people solve verbal fluency tasks (e.g. “List as many countries as possible within one minute.”). In recent papers, researchers found that the search through semantic memory appears to be similar to search in physical space, which involves a dynamic process of mediating between local exploitation and global exploration of clusters of information in much the same way that animals forage among patches of food in their environment. In this research project, this optimal foraging strategy will be applied to data of a divergent thinking task (Alternative Uses Task data) to see whether the search process in memory resembles the search process for animals for food.
Leonie Poelstra & Claire Stevenson
Scoring algorithms of creativity
These research projects focused on the development of automated scoring algorithms to assign ratings of creativity to answers on the Alternative Uses Task. The performance of these algorithms are compared to traditional scoring methods in terms of reliability and validity. The aim of this project is to develop a time-efficient and low-cost scoring method that can compete with existing methods. See Charlotte’s thesis for the results of the first version of such a scoring algorithm. And see Yina’s thesis for a newer version.
Claire Stevenson, Raoul Grasman, Charlotte Tanis & Yina Tsai
The fourth research project focuses on the structure of the semantic memory as an explanation of individual differences in creative potential. We use network models to mimic these semantic memory structures, making it easier to relate the results to existing theories about creativity and memory. In this project responses to the Alternative Uses Task will be used to create a novel task that will be used to construct individual networks. See this thesis for a more thorough theoretic background and the first results.
Iris Smal & Claire Stevenson