The Research Digest blog was five years old in February. As part of an ongoing celebratory series, I’ve asked Dr Gavin Nobes of the University of East Anglia to look back on his research on children’s naive models of the earth that I covered in March 2005, to reflect on that study and the field more generally. Here’s what he had to say:
“Almost 15 years ago the late George Butterworth visited UEL and inspired a group of us to follow up some work he and Michael Siegal had started in Australia. Using a novel, forced-choice question task, they were testing the claim, based on children’s drawings, that children have theory-like ‘naive mental models’ of the Earth; that is, children believe it to be (for example) flat, or a hollow sphere in which we live. This area of research has important implications for our understanding of the acquisition of knowledge, and for science education. For example, if children are influenced primarily by their intuitions and observations (as proponents of the naive mental model approach claim), they would be expected to think the Earth is flat; but if cultural communication is the principal source of information, children’s first concept of the Earth should be a rudimentary version of the scientific, spherical model.
In the study featured in the Digest five years ago, Georgia Panagiotaki, Alan Martin and I asked children not to draw but to choose, from a set of pictures, those that they thought best represented the Earth. As in the Australian study, we found that children knew much more about the Earth than previous researchers had claimed, and found no evidence of naive mental models.
Despite this apparently strong evidence from two different methods, the debate continued. Our recognition (forced-choice questions and picture selection) methods were criticised on the grounds that, unlike the earlier studies based on children’s drawings, they failed properly to elicit children’s understanding. We responded to these criticisms by giving university students the same drawing task and open-ended questions that had been given to children in the earlier studies. We were amazed to find that many of these adults drew exactly the same pictures, and gave identical non-scientific answers, as had children who were supposed to have naive mental models. Subsequent interviews revealed that the students had drawn and answered in these ways because they didn’t understand the questions – despite them being designed for 5-year-olds! Further experiments with a new version of the task, in which we rephrased the original open questions to reduce their ambiguity, led both adults and children to give substantially fewer non-scientific answers. We concluded that naive mental models are methodological artifacts: children and adults give these responses to the original instrument because the questions are poorly worded.
One recommendation that arises from this work is that, wherever possible, different methods should be used to test the same hypotheses. Another is that, however simple your children’s task might be, try it out first on adults: this is quick, easy, and can be remarkably revealing. And third, don’t be too dispirited by negative reviews: especially early on, editors sent our submissions to proponents of the naive mental model view, whose disparaging reviews resulted in rejections. Had it not been for Michael’s and George’s generous support and encouragement, we would probably have given up and turned to less controversial areas of research.”
Nobes, G., Martin, A., & Panagiotaki, G. (2005). The development of scientific knowledge of the Earth. British Journal of Developmental Psychology, 23 (1), 47-64 DOI: 10.1348/026151004×20649
Look out for more of these ‘looking back‘ guest posts in the coming months.