Does design thinking training increase creativity? Results from a field experiment with middle-school students

ABSTRACT

Design thinking remains mired in controversy. Its proponents claim that it enhances not just confidence but also creativity, while the opponents question whether it does anything beyond building unfounded confidence. To bring rigorous evidence to this debate, we designed a randomised field experiment amongst school children served by a major non-governmental organisation in rural India. The findings reveal that the design thinking training did not just increase confidence: it also significantly increased ideational fluency and elaboration in a divergent thinking task, although the originality and flexibility of the generated ideas was on average lower in the treatment group than the control group. We also find that the increase in confidence occurred primarily among female students, whereas the increase in ideational fluency and elaboration occurred for both genders in our study.

KEYWORDS:

• Design thinking
• innovation
• creativity
• experiment

Acknowledgments

We thank Agastya International Foundation for allowing us to conduct this study in their campus in Kuppam, India. We also thank Stanford University and INSEAD for research funding. Abhishek Naresh provided research assistance, and Panchali Guha gave us useful feedback.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. https://en.wikipedia.org/wiki/Agastya_International_Foundation.

2. https://www.dfcworld.com/SITE/dfcstory.

3. We should note that perspective taking differs from empathy: the former relates to ‘Can I see the world from your point of view?’ and the latter pertains to ‘Can I feel what you feel?’.

4. More details related to the pilot study are available upon request from the authors.

5. Naturally, we cannot be sure that the impact of covering the entire workbook (whose details are available upon request) would necessarily have the same impact as going through just these specific (even if core) activities as interventions in the actual experiment. However, it seems reasonable to expect that going through the entire workbook would only have strengthened the design thinking training treatment. Therefore, to the extent that even this subset of the activities produces an effect on creativity and other outcomes, the findings can be considered a lower bound.

6. As full disclosure, we note that we also attempted to conduct an additional (and shorter) single-day experiment based on a smaller subset of DFC workbook activities and a new sample of students. However, the randomisation in that case was unsuccessful (covariate balance was not achieved across groups), making that data unusable for further study.

7. This measure is based on the argument by Torrance (1981) that ‘adding details that help to bring the idea to life and create a better understanding and appreciation of it. These additional details can often transform an old idea into a new one with greater potential’.

8. While Agastya leaders informed us that the sample of students coming to Agastya is representative of the student population in the region in general, we do not have precise data to back this up. Therefore, we cannot conclusively claim generalisability of our results to students beyond those with pre-exposure to Agastya activities.

9. There was no significant difference between participants in the treatment groups (pooled) and the control group in the absolute number of suggestions to improve their village, but that is not relevant for measuring confidence.

10. Recall that ‘valid’ here means that a drawn object adhered to the specific instructions on how to use the circles for drawing images. But the same result holds even if we measure fluency using just the raw count for the number of ideas generated (more circles filled in: 23.6 vs. 19.9, p=0.02 in a t-test).

11. In randomisation inference tests, it is possible to account for the possibility that the difference between treatment and control is simply based on which units are assigned to the treatment group-even if the treatment itself has no effect. The procedure involves reassigning ‘treatment’ at random hundreds or thousands of times, to compute the probability of differences under the null hypothesis that the treatment has no effect. Athey and Imbens (2016) urge those who conduct field experiments to use randomisation inference because uncertainty in estimates arises naturally from the random assignment of the treatments, rather than only from hypothesised sampling from a large population.

12. We also developed interim measures for Days 1 and 2 that were not part of the analysis for this project, which only relies on final (Day 3 measures). We had piloted those to help develop additional measures of confidence and perspective taking for possible use by Agastya for future purposes. For perspective taking, for instance, we asked ‘One of your classmates often does not do his homework. Can you list reasons why he may not have done his homework?’ For confidence, we asked ”A new project has begun in your school to help keep the playground clean; would you join such a project”? and for creativity, ‘ Make a list of all the things you can do to stop littering in your school’. These interim measures were administered uniformly to all treatment and control groups, and while these still need refinement and validation, the pattern of results for these Day 1 and Day 2 measures seem broadly similar to what we obtained with the final day 3 measures after the experiment was competed.

13. Yet another possibility arises from the fact that, although the design thinking workbook was provided in the local language (Telugu), the perspective taking exercise did rely on English to the extent that the participants had to draw the letter ‘E’. As these were middle school students who had all taken English for several years, we felt reasonably confident that they had the requisite competence at least to be able to draw the ‘E’ correctly. Nonetheless, we cannot fully rule out that the non-finding for this exercise arose in part from their limited familiarity with English.