Naming unrelated words reliably predicts creativity

Abstract

Several theories posit that creative people are able to generate more divergent ideas. If this is correct, the simple act of naming unrelated words and then measuring the semantic distance between them could serve as an objective measure of creativity. To test this hypothesis, we asked 8,892 participants to name 10 words that are as different from each other as possible. A computational algorithm then estimated the average semantic distance between the words; related words (e.g., “cat” and “dog”) have shorter distances than unrelated ones (e.g., “cat” and “thimble”). We predicted that people producing greater semantic distances would also score higher on traditional creativity measures. In Study 1, there were moderate to strong correlations between semantic distance and two other creativity measures (the Alternative Uses Task and the Bridge-the-Associative-Gap Task). In Study 2, with participants from 98 countries, semantic distances varied only slightly by demographic variables which suggests that the measure can be used without modification across diverse populations. There was also a positive correlation between semantic distance and performance on problem solving tasks known to predict creativity. Overall, semantic distance correlated at least as strongly with established creativity measures as those measures did with each other. Naming unrelated words in what we call the Divergent Association Task can thus serve as a brief, reliable, and objective measure of creativity.

Forward flow and creative thought: Assessing associative cognition and its role in divergent thinking

“…to type proper nouns (such as names, brands, etc. ), which are not commonly included in text corpora used to compute semantic distance (cf., Olson et al, 2020). Each trial presented a starting word followed by 19 text boxes, where participants typed their chained free association responses.…”

A nudge-based intervention to reduce problematic smartphone use: Randomised controlled trial

“…We are not aware of any other empirical studies that have tested this link 7 although it has frequently been hypothesised 72 . However, in a follow-up pre-registered study (https://osf.io/ztdbk) with a more diverse sample (N = 300), we were unable to replicate this relationship across two measures of smartphone use (screen time and the Smartphone Addiction Scale) 70 and two measures of creativity (the Divergent Association Task and the Creative Achievement Questionnaire) 73,74 . The link between creativity and smartphone use thus remains unclear.…”

“…An algorithm then computed the semantic distance between each of the words; higher average distances indicate greater creativity. The task correlates well with traditional creativity measures 73 .…”

“…Divergent Association Task (DAT) As an exploratory predictor, a subset of the participants (N = 62) additionally completed the DAT, a measure of creativity 73 . The task involves writing 10 words that are as different from each other as possible in all meanings and uses of the words.…”

Construction and validation of a computerized creativity assessment tool with automated scoring based on deep-learning techniques.

“…As LSA has been increasingly employed for automated DT test scoring, researchers have also begun to identify its limitations (Forthmann et al, 2019; Heinen & Johnson, 2018; Kenett, 2019; Olson et al, 2020). In our view, the limitation that most severely restricts its application is the conflict between the open-ended nature of DT and the computational expense of LSA: The enormous challenge of providing a set of all possible responses to DT tests means that a comprehensive semantic model is needed to form the groundwork of any automated scoring system.…”