A Structural Model of Prospective Science Teachers' Nature of Science Views

Abstract

This study aims to establish a viable structural model of prospective science teachers' nature of science (NOS) views, which could be used as an analytical tool for understanding the complex relationships between prospective teachers' conceptions of NOS and factors possibly affecting their conceptions. In order to construct such a model, likely factors that might influence prospective teachers' NOS views were hypothesized. These included science process skills; attitudes toward science teaching; academic achievement in pedagogical and science courses; and social, religious, economic, political, aesthetic, and theoretical values. The hypothetical model was then developed and modified using structural equation modeling methodology. The final viable model indicates that attitudes toward science teaching, science process skills, academic achievement in pedagogical courses, religious values, and economic values explain NOS views with low predictive power.

Keywords:

• nature of science
• prospective teachers
• model

Notes

¹There are criticisms about these kinds of assessment tools (Lederman, Wade, & Bell, 1998; Lederman, Abd El Khalick, Bell, & Schwartz, 2002). These tests do not give the opportunity to participants to express themselves freely, and therefore the researcher may not get complete information.

²Grade Point Average of Physics I, Physics II, Chemistry I, Chemistry II, Biology I, Biology II, Mathematics I, and Mathematics II courses, which are required courses in both universities.

³Grade Point Average of Introduction to Education, Development and Learning, Planning and Assessment, Educational Technology and Material Development, Teaching Science, and Classroom Management courses, which are required courses in both universities.

⁴For details about chi‐square analysis and fit indices please see Klein, 1998; Byrne, 2001.

⁵In the practice of structural equation modelling, it is often found that the researcher's model does not coincide with the data, as shown by statistical tests. The researcher may then modify the model by freeing substantively interesting parameters by looking at the modification indices (MI) and the expected parameter change statistics (EPC). This approach to model modification was developed by Saris, Satorra, and Sörbom (1987) and has been advocated by Kaplan (1990) as a way of modifying models taking into account power considerations (George & Kaplan, 1998, p.100).