Educating and Measuring Choice: A Test of the Transfer of Design Thinking in Problem Solving and Learning

Abstract

Educators aim to equip students with learning strategies they can apply when approaching new problems on their own. Teaching design-thinking strategies may support this goal. A first test would show that the strategies are good for learning and that students spontaneously transfer them beyond classroom instruction. To examine this, we introduce choice-based assessments (CBAs). CBAs measure how people learn when there is minimal guidance and they must make decisions as independent learners. Here, sixth-grade students completed multiple design activities that emphasized either seeking constructive criticism or exploring a space of alternatives. Afterward, they completed the CBAs, which measured strategy transfer. Results showed that lower-achieving students benefitted most from instruction, exhibiting a relative increase in their use of design-thinking strategies. In addition, strategy choices correlated with prior achievement measures and appeared to mediate performance in and learning from the CBAs. The choices to use the two strategies themselves were not correlated, which indicates that they are not subsets of a larger construct, such as growth mindset. In sum, CBAs enabled a double demonstration: design-thinking strategies may improve learning and problem solving, and design-thinking instruction may improve the likelihood of lower-achieving students choosing to use effective strategies in novel settings that require new learning.

Acknowledgments

The authors would like to thank the students, teachers, and administrators, particularly Erik Burmeister of Hillview Middle School, Menlo Park, California. Without their enthusiasm and support, this study would not have happened.

Supplementary Material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/10508406.2019.1570933

Notes

¹ Multiple imputation of a given variable uses other available data to predict the value of a missing data point (e.g., using multiple regression). Given the estimate of the value, the imputation uses the variance in the existing data to construct a distribution around the value and then randomly samples from that distribution. Each imputation gives a slightly different value for each missing data point because of the random sampling. Further details about the imputation model used for this analysis are available in the Appendix.

² Note that the bin width values were determined by the distribution of SAM from the imputations, not the original data set.

³ We did not develop performance or learning measures for Apex-Explore.

Additional information

Funding

This material is based on work supported by the National Science Foundation (Grant Nos. 1228831 and 1337414) and the Gordon and Betty Moore Foundation. Any opinions, findings, and conclusions or recommendations expressed in this material are the authors’ alone and do not necessarily reflect the views of the granting agencies.