Investigating early years teachers' understanding and response to children's preconceptions

Abstract

This paper focuses on young children’s scientific preconceptions and discusses teachers’ identification of these preconceptions when teaching science in the early years, on which research is still limited. This paper is based on the theoretical framework of constructivism and it defines preconceptions as children’s erroneous concepts prior to formal education. A two phase case study approach was employed, facilitating in-depth investigation though the use of questionnaires, interviews and observations. The results indicate that the teachers did not dedicate time to identify children’s preconceptions when planning and teaching science, even when acknowledging preconceptions’ possible existence. This indicates a lack of appreciation of the importance of children’s preconceptions of the consequences when ignoring them. The results also indicate the need for further training and professional development in relation to the teaching of early-years science, especially since only a very small percentage of early years teachers tend to study science during their years compulsory education. A number of suggestions are also provided that can be useful for practice and policy as well.

Keywords:

• early years education
• science education
• preconceptions/misconceptions/alternative ideas
• conceptual development
• research and practice

Notes

1. All questionnaires were completed by female early years teachers which indicates that the profession is female dominated. This observation is also confirmed by the Ministry of Education and Culture based on a list of all in-service early years teachers, 99.9% of whom are women (www.eey.gov.cy).

2. Cronbach's Alpha can lie between zero (0) and +1 - any value over +0.7 is acceptable and indicates that the scale is reliable.

3. The two samples, private and public schools, are independent, the appropriate test to use was the independent samples t-test’ (when p<0.05 then fewer than five out of a 100 cases are spurious).

4. The assumptions necessary to use this test were: (1) the sample data had to be normally distributed; (2) the sampling distribution had to be normally distributed; and (3) data had to be interval or ratio (Field, 2009). For a positive correlation Pearson's r lies between 0 < r ≥ 1 whereas for a negative correlation Pearson's r lies between −1 and 0.00.