This longitudinal study explored change in teachers' knowledge of subjects they teach from preservice training through 17 years of professional experience. It followed secondary school science teachers in Australia, through sequences of individual interviews in which change in content knowledge (mainly energy-related) was probed primarily via concept profiles-a word-association method. Change was found to be multifaceted, with details of unused content fading from memory, alongside growth that results from improved understanding and reorganization of structure more than from accretion of new material. Across personal and professional life tracks that produce variation between individuals, development is facilitated by "critical mass" of teachers' knowledge and interest in their chosen disciplines of study and certification, whereas deficiencies tend to persist in the other subjects they are asked to teach. The required curriculum is the single most powerful determinant of teacher knowledge, serving as both its organizer and source. Based on the study, a three-phase model of teacher content-knowledge development is proposed, and the discussion highlights the need for career-long support for growth, even in teachers' major subjects where expertise is taken for granted.C 2007 Wiley Periodicals, Inc. Sci Ed 92: 221 -251, 2008 How does teachers' knowledge of subject matter change over their professional careers? Research into this question is not abundant. Most studies on science teachers' content knowledge are limited to the preservice period, more often done with elementary rather than secondary student-teachers, and few continue through the early career years (Cochran & Jones, 1998;Davis, Petish, & Smithey, 2006;Gess-Newsome & Lederman, 1999).An earlier version of this paper was presented at the Annual Meeting of the American Educational Research Association, San Diego, CA, April 2004.Hanna J. Arzi was at the Faculty of Education, Monash University, during the first stage of the reported longitudinal study, 1985 -1987. Correspondence to: Hanna J. Arzi; e-mail: arzi hj@netvision.net.il C 2007 Wiley Periodicals, Inc.
ARZI AND WHITEIn the absence of longitudinal data, inferences on later years rely on questionable extrapolation of early data and comparisons between preservice students and experienced teachers.There are reasons for the paucity of research on teacher development over long periods, starting with difficulties involved in any long-term longitudinal study due to research maintenance issues of sample attrition or continued funding (challenges of longitudinal studies are discussed in Arzi, 1988aArzi, , 2004White & Arzi, 2005). In the case of teacher content knowledge, a further concern is the sensitivity of teachers who can find repetitive probing threatening, reflecting disrespect for their professional credentials, particularly so for experienced secondary teachers who are expected to be subject experts. Consequently, teacher cooperation beyond their preservice courses is not easy to obtain a...