Creative Generation Task Research Articles

Can Large Language Models Transform Computational Social Science?

Abstract Large language models (LLMs) are capable of successfully performing many language processing tasks zero-shot (without training data). If zero-shot LLMs can also reliably classify and explain social phenomena like persuasiveness and political ideology, then LLMs could augment the computational social science (CSS) pipeline in important ways. This work provides a road map for using LLMs as CSS tools. Towards this end, we contribute a set of prompting best practices and an extensive evaluation pipeline to measure the zero-shot performance of 13 language models on 25 representative English CSS benchmarks. On taxonomic labeling tasks (classification), LLMs fail to outperform the best fine-tuned models but still achieve fair levels of agreement with humans. On free-form coding tasks (generation), LLMs produce explanations that often exceed the quality of crowdworkers’ gold references. We conclude that the performance of today’s LLMs can augment the CSS research pipeline in two ways: (1) serving as zero-shot data annotators on human annotation teams, and (2) bootstrapping challenging creative generation tasks (e.g., explaining the underlying attributes of a text). In summary, LLMs are posed to meaningfully participate in social science analysis in partnership with humans.

Differences in brain activity patterns during creative idea generation between eminent and non-eminent thinkers

An influential model of the neural mechanisms of creative thought suggests that creativity is manifested in the joint contributions of the Default Mode Network (DMN; a set of regions in the medial PFC, lateral and medial parietal cortex, and the medial temporal lobes) and the executive networks within the dorsolateral PFC. Several empirical reports have offered support for this model by showing that complex interactions between these brain systems account for individual differences in creative performance. The present study examined whether the engagement of these regions in idea generation is modulated by one’s eminence in a creativity-related field. Twenty (n ​= ​20) healthy eminent creators from diverse fields of expertise and a ‘smart’ comparison group of sixteen (n ​= ​16) age- and education-matched non-eminent thinkers were administered a creative generation task (an adaptation of the Alternative Uses Task) and a control perceptual task, while undergoing functional magnetic resonance imaging (fMRI). The participants’ verbal responses were recorded through a noise-canceling microphone and were later coded for fluency and accuracy. Behavioral and fMRI analyses revealed commonalities between groups, but also distinct patterns of activation in default mode and executive brain regions between the eminent and the non-eminent participants during creative thinking. We interpret these findings in the context of the well-documented contributions of these regions in the generation of creative ideas as modulated, in this study, by participants’ creative eminence.

The relationship between constraining effects of examples in a creative generation task and mind wandering

The relationship between constraining effects of examples in a creative generation task and mind wandering

The relation between openness and creativity is moderated by attention to peers’ ideas in electronic brainstorming

ABSTRACT An emerging body of research has focused on students’ creativity in group contexts, with the assumption that students could be inspired by peers’ ideas. Although students’ openness and attention to peers’ ideas are claimed to play important roles in their creativity in group settings, there is little empirical research that tests this assumption. This study examined the moderating effect of attention to peers’ ideas in the relation between openness and creativity in electronic brainstorming. Participants were 91 undergraduate students who took about 10 min to complete a creative idea generation task during electronic brainstorming. Regression analyses found that students who were characterized by high openness were more creative, but only when they showed more attention to peers’ ideas. This suggests that electronic brainstorming can be useful for enhancing the creativity of some students.

Visiting the land of dream muses: The relationship between lucid dreaming and creativity

Dreaming is a very creative state of mind. Lucid dreamers – who can become aware in their dreams that they are dreaming – have an advantage of directly tapping into the creative potentials of the dreaming mind. This study explored the relationship between lucid dreaming and creativity. Twenty lucid dreamers were asked to accomplish a creative generation task once in a lucid dream and once while awake. Their performance was compared to a control group of 20 non-lucid dreamers who did the task once (i.e. awake). While according to the present findings, the lucid dream state did not appear to be more advantageous for creativity than waking imagination, lucid dreamers performed better on a creative task than non-lucid dreamers, which supports the notion of existing relationship between the ability for lucid dreaming and creativity. Future studies should explore the directionality of this relationship and make further comparisons of creative performance in wakefulness and the lucid dream state.

Temporal and Spatial Patterns of Neural Activity Associated with Information Selection in Open-ended Creativity

Novel information selection is a crucial process in creativity and was found to be associated with frontal–temporal functional connectivity in the right brain in closed-ended creativity. Since it has distinct cognitive processing from closed-ended creativity, the information selection in open-ended creativity might be underlain by different neural activity. To address this issue, a creative generation task of Chinese two-part allegorical sayings was adopted, and the trials were classified into novel and normal solutions according to participants’ self-ratings. The results showed that (1) novel solutions induced a higher lower alpha power in the temporal area, which might be associated with the automatic, unconscious mental process of retrieving extensive semantic information, and (2) upper alpha power in both frontal and temporal areas and frontal–temporal alpha coherence were higher in novel solutions than in normal solutions, which might reflect the selective inhibition of semantic information. Furthermore, lower alpha power in the temporal area showed a reduction with time, while the frontal–temporal and temporal–temporal coherence in the upper alpha band appeared to increase from the early to the middle phase. These dynamic changes in neural activity might reflect the transformation from divergent thinking to convergent thinking in the creative progress. The advantage of the right brain in frontal–temporal connectivity was not found in the present work, which might result from the diversity of solutions in open-ended creativity.

Cultural Threats in Culturally Mixed Encounters Hamper Creative Performance for Individuals With Lower Openness to Experience

Past research has examined independently how openness to experience, as a personality trait, and the situational threat triggered by a foreign cultural encounter affect the emergence of creative benefits from a culture-mixing experience. The present research provides the first evidence for the interactive effect of openness to experience and cultural threat following culturally mixed encounters on creative performance. In Study 1, under heightened perceptions of cultural threat, exposing to the mixing of Chinese and American cultures (vs. a non-mixed situation) made close-minded Chinese participants to perform more poorly in a creative generation task. In Study 2, inducing cultural threat by having a foreign cultural icon spatially intrude a sacred space of the local culture caused Chinese participants with lower levels of openness to perform less creatively when the foreign icon was deemed highly symbolic of the foreign culture. These patterns of effects did not emerge among open-minded participants. These findings suggest that trait openness acts as a buffer against foreign cultural threat to sustain the creative benefits of culture mixing.

A (Creative) Portrait of the Uncertain Individual: Self-Uncertainty and Individualism Enhance Creative Generation.

Building on findings that self-uncertainty motivates attempts to restore certainty about the self, particularly in ways that highlight one's distinctiveness from others, we show that self-uncertainty, relative to uncertainty in general, increases creative generation among individualists. In Studies 1 to 3, high (but not low) individualists performed better on a creative generation task after being primed with self-uncertainty as opposed to general uncertainty. In Study 4, this effect emerged only among those who were told that the task measured creative as opposed to analytical thinking, suggesting that the positive effects of self-uncertainty on performance are specific to tasks that bolster perceptions of uniqueness. In Study 5, self-uncertain individualists experienced a restoration of self-clarity after being induced to think about themselves as more (vs. less) creative. Implications for compensatory responses to self-uncertainty and factors that influence creativity are discussed.

Workplace duties or opportunities? Challenge stressors, regulatory focus, and creativity

Previous research has produced contradictory findings about the impact of challenge stressors on individual and team creativity. Based on the challenge–hindrance stressors framework (LePine, Podsakoff, & LePine, 2005) and on regulatory focus theory (Higgins, 1997), we argue that the effect of challenge stressors on creativity is moderated by regulatory focus. We hypothesize that while promotion focus strengthens a positive relationship between challenge stressors and creativity, prevention focus reinforces a negative relationship. Experimental data showed that high demands led to better results in a creative insight task for individuals with a strong trait promotion focus, and that high demands combined with an induced promotion focus led to better results across both creative generation and insight tasks. These results were replicated in a field R&D sample. Furthermore, we found that team promotion focus moderated the effect of challenge stressors on team creativity. The results offer both theoretical insights and suggest practical implications.

Stable and Dynamic Properties of Category Structure Guide Imaginative Thought

Previous research has shown that category exemplars vary in how accessible they are within their categories, and that more accessible exemplars are more likely than less accessible ones to be used as starting points in creative idea generation. In the present study, specific exemplars of fruit and tools that varied in their baseline levels of unprimed accessibility were made more accessible by way of an initial rating task, which led to an increased likelihood of those primed exemplars being used in a subsequent creative generation task. At the same time, items with higher baseline levels of unprimed accessibility continued to be used more often than items with lower unprimed accessibility, providing evidence that long-term associative strengths and recent experiences with exemplars both play a role in creative generation. Explicit recall of primed items did not correlate with their use in imagination, consistent with the idea that reliance on recent experiences in generating novel ideas may be inadvertent rather than deliberate.

The role of domain knowledge in creative generation

Previous studies have shown that a predominant tendency in creative generation tasks is to base new ideas on well-known, specific instances of previous ideas (e.g., basing ideas for imaginary aliens on dogs, cats or bears). However, a substantial minority of individuals has been shown to adopt more abstract approaches to the task and to develop more original products as a result. Because domain knowledge is sometimes thought to be associated with more abstract representations, it is possible that these individual differences are linked to how much people have learned about the domain in question. The present study explored the relationship between domain knowledge, the tendency to rely on specific domain instances and both the originality and practicality of outcomes in a task in which participants designed novel sports. In addition to designing novel sports, participants rated their own knowledge about and participation in sports, and they took a short test of sport knowledge. As in previous studies, there were individual differences in approaches to the task, and participants who relied on specific known sports (e.g., basketball) produced less original designs. Neither the measures of sport knowledge nor the measure of sport participation related to the tendency to rely on specific instances in developing the novel sports or to the originality of those creations. However, there were links between tested knowledge and the rated practicality of the designs.

Interactive Effects of Mood and Task Framing on Creative Generation

ABSTRACT This study introduced a motivational compatibility account for the influence of mood on creative generation. Building upon the feelings-as-information framework, it was proposed that positive moods signal to individuals that they are safe, motivating them to take advantage of this presumed safety by seeking stimulation and incentives (i.e., having fun), whereas negative moods signal to individuals that there are problems at hand, motivating them to solve these problems. Based on these assumptions, it was predicted that positive and negative moods should enhance effort on creative generation tasks construed as compatible with the motivational orientations they respectively elicit. Specifically, positive, relative to negative, moods were predicted to enhance effort on tasks construed as fun and silly, whereas negative, relative to positive, moods were predicted to bolster effort on tasks construed as serious and important. Evidence for this model, and several of its underlying assumptions, was adduced in 3 experiments in which mood was manipulated and participants completed creative generation tasks that were framed as either fun or serious. Results are discussed with an eye toward addressing alternative theoretical explanations.

Implications of Counterfactual Structure for Creative Generation and Analytical Problem Solving

In the present research, the authors hypothesized that additive counterfactual thinking mind-sets, activated by adding new antecedent elements to reconstruct reality, promote an expansive processing style that broadens conceptual attention and facilitates performance on creative generation tasks, whereas subtractive counter-factual thinking mind-sets, activated by removing antecedent elements to reconstruct reality, promote a relational processing style that enhances tendencies to consider relationships and associations and facilitates performance on analytical problem-solving tasks. A reanalysis of a published data set suggested that the counterfactual mind-set primes previously used in the literature tend to evoke subtractive counterfactuals. Studies 1 and 2 then demonstrated that subtractive counterfactual mind-sets enhanced performance on analytical problem-solving tasks relative to additive counterfactual mind-sets, whereas Studies 3 and 4 found that additive counterfactual mind-sets enhanced performance on creative generation tasks relative to subtractive counterfactual mind-sets.

Thinking within the box: The relational processing style elicited by counterfactual mind-sets.

By comparing reality to what might have been, counterfactuals promote a relational processing style characterized by a tendency to consider relationships and associations among a set of stimuli. As such, counterfactual mind-sets were expected to improve performance on tasks involving the consideration of relationships and associations but to impair performance on tasks requiring novel ideas that are uninfluenced by salient associations. The authors conducted several experiments to test this hypothesis. In Experiments 1a and 1b, the authors determined that counterfactual mind-sets increase mental states and preferences for thinking styles consistent with relational thought. Experiment 2 demonstrated a facilitative effect of counterfactual mind-sets on an analytic task involving logical relationships; Experiments 3 and 4 demonstrated that counterfactual mind-sets structure thought and imagination around salient associations and therefore impaired performance on creative generation tasks. In Experiment 5, the authors demonstrated that the detrimental effect of counterfactual mind-sets is limited to creative tasks involving novel idea generation; in a creative association task involving the consideration of relationships between task stimuli, counterfactual mind-sets improved performance.

Automatic effects of deviancy cues on creative cognition

Three experiments test the existence of an automatic deviancy-creativity link. Using a lexical decision task, in Experiment 1 we found a semantic link between deviancy and creativity words in that decision times for creativity-related words were enhanced after subliminal deviancy priming. In Experiment 2, participants were led to think about either a punk or an engineer and afterwards were administered creative insight and analytical reasoning problems. According to a pretest, punks and engineers were judged as differing in uniqueness but not in creativity. Participants given ‘punk’ priming solved more creative insight problems and fewer analytical reasoning problems than those given ‘engineer’ priming. In Experiment 3, participants were incidentally exposed to abstract artworks symbolically expressing either the concept of conformity or deviancy and were subsequently asked to solve a creative generation task. Exposure to the artwork representing deviancy led to generation of more creative solutions than exposure to that representing conformity. Copyright © 2005 John Wiley & Sons, Ltd.

The Role of Specificity and Abstraction in Creative Idea Generation

Three experiments examined the relationship between approaches to a creative generation task and the novelty of the resulting products. Participants were given the task of imagining life on other planets and they received instructions intended to encourage them to formulate the task in either very specific ways (e.g., thinking of specific Earth animals) or more abstract ways (e.g., thinking of environmental conditions and general survival needs). The latter instructions led to more novel creations. The results are discussed in terms of the malleability of people's approaches to creative generation, the role of problem formulation and the link between abstraction and novelty.

Creative cognition in gifted adolescents

Gifted adolescents performed a creative generation task in which they imagined and drew fruit that might exist on another planet. They produced creations that tended to include central properties of Earth fruit, such as seeds and stems, but also more unusual properties, such as poisonousness. They also developed imaginary fruit that were rated as more original than those developed by a comparison sample of college students, and they did so regardless of whether they were explicitly instructed to be creative. They were also less likely than college students to base their creations on specific exemplars of Earth fruit, but among those who did so, they tended to rely on highly representative types of fruit (e.g., apples) as starting points for their novel ideas. The results are discussed in terms of how existing knowledge directs imaginative activities in individuals that exhibit different degrees of generative or creative performance.

How examples may (and may not) constrain creativity.

Three experiments were performed to test Smith, Ward, and Schumacher's (1993) conformity hypothesis-that people's ideas will conform to examples they are shown in a creative generation task. Conformity was observed in all three experiments; participants tended to incorporate critical features of experimenter-provided examples. However, examination of total output, elaborateness of design, and the noncritical features did not confirm that the conformity effect constrained creative output in any of the three experiments. Increasing the number of examples increased the conformity effect (Experiment 1). Examples that covaried features that are naturally uncorrelated in the real world led to a greater subjective rating of creativity (Experiment 2). A delay between presentation and test increased conformity (Experiment 3), just as models of inadvertent plagiarism would predict. The explanatory power of theoretical accounts such as activation, retrieval blocking, structured imagination, and category abstraction are evaluated.

Constraining effects of examples in a creative generation task.

In three experiments we tested the conformity hypothesis--that subjects' ideas would conform to examples they had been shown--by using a creative generation paradigm in which subjects imagined and sketched new exemplars of experimenter-defined categories. Designs made by subjects who had first seen three examples of ideas were compared with those of control subjects, who received no examples. In all three experiments, the designs of subjects who had seen the examples were more likely to contain features of the examples. This conformity effect did not significantly decrease in Experiment 2, when a 23-min task was interpolated between viewing the examples and generating related ideas. The hypothesis that the observed conformity effects may have been caused by subjects' assumptions that they should try to generate ideas similar to the examples was refuted in Experiment 3; explicitly instructing subjects to create ideas that were very different from the examples did not decrease conformity to the examples, and instructing them to conform to the examples significantly increased conformity. The results show that recent experience can lead to unintentional conformity, constraining the generation of creative ideas.