Teacher self-efficacy, an adapted version of self-efficacy in teaching context (Fives & Alexander, 2004) is related to teachers’ belief in their ability to teach mathematics (personal efficacy) and their belief that students’ learning can be influenced by effective teaching. Tschannen-Moran and Woolfolk Hoy (2001) stated that teacher efficacy is an important determinant for teachers’ persistence, enthusiasm, commitment, and instructional decisions.

Development of efficacy-belief is important to investigate since it is hard to change well-established belief systems (Woolfolk Hoy, 2001). Thus, conducting longitudinal studies is important to see the changes in self-efficacy across time (Woodcock, 2011). As a continuation of Işıksal-Bostan’s work (2016), the aim of this longitudinal study was to investigate prospective middle school mathematics teachers’ (PMMT) efficacy belief.

Thus, following research question guide the present study:
1. How does PMMT mathematics teaching efficacy belief change during their enrollment in teacher education program and at the end of their first year and second year employment as a teacher?.

A longitudinal survey design was used. Data were collected from 30 prospective mathematics teachers enrolled in large public university in Ankara, Turkey. The first data were collected from PMMT before their graduation (Time 1). The second data set were collected at the end of their first year employment (Time 2). Last data set were collected at the end of their second year of employment as a teacher (Time 3). Mathematics Teaching Efficacy Belief Instrument (MTEBI, Enochs et al., 2000) was used to collect data.

Findings showed that the total mathematics teaching efficacy scores differed significantly between time periods (p=.01). More specifically, total efficacy score decreased significantly (p=.01) from Time 1( M= 85.8, SD= 7.4) to Time 2 (M= 81.2, SD= 7.4). Participants’ efficacy scores started to increase after Time 2 and created significant difference (p=.01) between Time 2 and Time 3 (M= 84.7, SD= 5.9). Findings revealed that there was no significant difference in mean scores between Time 1 and Time 3 (p=.5). Parallel to the overall efficacy scores, findings revealed that participants self-efficacy and outcome scores decreased significantly from Time 1 to Time 2 increased significantly in the Time period 3.

Findings revealed that teachers’ efficacy belief decreased significantly after their graduation. When the first data set were they had completed their course in which they had many opportunities to prepare activities, observe students and their mentor teachers. Thus, those engagements could have positively influenced their efficacy beliefs (Yürekli, 2015). However, findings revealed that there was a sharp decrease in teachers’ efficacy score after their recruitment as a teacher. However when novice teachers confronted with complexities of actual teaching, they may lose their confidence in their ability to teach mathematics effectively (Li & Zhang, 2000). Teachers’ efficacy belief increased at the end of their second year of employment. Better classroom experiences (Holzberger et al., 2013) and being more familiar with the content may be the factors that support that increase. Further qualitative research studies could be conducted to investigate the factors that support or inhibit teaching efficacy belief.

Keywords: Teacher efficacy, Longitudinal study.