Several studies have confirmed that the teacher’s knowledge is essentially practical and personal, and that it is closely related to the appreciation of his own experiences (Clandinin, 1986; Clandinin & Connely, 1987; Elbaz, 1983; Onofre, Carreiro da Costa & Marcelo, 2002). Studies concerning occupational socialization (Lacey, 1977; Lortie, 1975; Templin & Schempp, 1989) clearly identify the phases of the teacher’s learning trajectory: anticipatory socialization, preservice training, and inservice training. Teacher Personal Efficacy (TPE) which is representative of the belief of teacher capacity has been noticed as a basilar variable to the quality of teaching (Woolfolk Hoy & Spero, 2005; Bandura, 1997; Henson, 2002). Several studies have demonstrated the way how this belief is related to the teacher's commitment, effort and motivation (Bandura 1997), his(er) practical knowledge and teaching behaviour (Onofre, Carreiro da Costa & Marcelo, 2002). According to the integral model of teacher efficacy (Tschannen-Morgan, Woolfolk Hoy, & Hoy, 1998), the major influences on efficacy beliefs are the attributional analysis and interpretation of the four sources of teacher personal efficacy: enactive mastery experience, vicarious experience, verbal persuasion and physical and affective states. Recent research concerning teaching in general, tries to understand better how is the teacher’s education reflected on the alteration of the TPE. The results are not totally clear about the relative importance of the several sources of the TPE, especially when considering some professional specificity of teachers' activity like their own experience, the gender and the disciplinary area. In effect, Tschannen-Moran & Woolfolk Hoy (2007) have shown that at the beginning of the teachers’ career this belief is particularly sensitive to verbal persuasion, despite Tschannen-Morgan, Woolfolk Hoy, & Hoy (1998) pointed out vicarious experience as a privileged source for those teachers, and concluded simultaneously that the mastery experiences were the ones that influenced the most the teachers’ TPE in general. Zeldin e Pajares (2000) had already verified that vicarious experience and verbal persuasion were the most influent on women mathematics teachers’ TPE. Pereira (2002) verified that, in Portugal, the support teachers’ TPE was globally determined by mastery experiences, vicarious learning, verbal persuasion and satisfaction or anxiety. Jardim (2007) verified that the most preponderant sources of preservice teachers’ TPE were mastery experience and verbal persuasion. The clarification of the relative importance of these sources and its expression during teachers' professional socialization requires more studies about sources of teacher personal efficacy (Tschannen-Moran & Woolfolk Hoy, 2007; Labone 2004). In Portugal, as far as this ambit is concerned and relatively to Physical Education subject, studies are almost non existent. Therefore, this study aimed to identify: - the main sources of physical education teachers’ TPE; - in what extent do these sources differentiate higher and lower teachers’ TPE; - how are the teaching experience and gender factors differentiating these teachers.

This study was developed in two phases: an extensive and an intensive one. During the extensive phase, to the characterization of the TPE of physical education teachers of Portuguese elementary and high schools, a Portuguese version of the Scale of the Teacher’s Efficacy (Gibson e Dembo, 1984, apud Onofre, 2000) was used. This phase was developed with the purpose of selecting a number of representative cases of two groups of teachers with high and low TPE, to be studied later in the intensive phase. In this second phase, the cases were selected according to three criteria. Considering the level of teachers’ TPE two groups were formed. Four teachers with a lower TPE (less than or equal to percentile 25: 3,8) and four teachers with a higher TPE (greater than percentile 75: 4,6) were selected. The second and the third criteria, respectively the teaching experience and the gender, were used in order to control their probable influence in the main studied concept– TPE. Each group was composed by two experienced teachers and two novice ones, two men and two women. This phase was developed according to an inductive approach as a multicase study (Yin, 1994). Teachers were inquired in order to identify their TPE sources. A semi-structured interview was used. The teachers answers were transcribed in written protocol, and then submitted to a thematic content analysis using Aquad software, version 6.0 (Huber, 2005). This strategy allowed the safeguard of the conditions for an analytic generalization of the evidences (Yin, 1994), enabling the exercise of literal and theoretical replication of the data obtained in the different cases, and so, pointing out the similarities and differences in the experiences that led to the variation of TPE.

Results have shown that the experiences indicated by teachers as being the most influent on their TPE are connected to the mastery experiences either in the preservice and inservice training. As far as vicarious experience is concerned, it was mentioned only by two teachers, and verbal persuasion indicated by a single one. In what concerns TPE sources, no differences were found between groups. We believe that the non existence of those differences may be associated to the “power” of other variables. Analysing that hypothesis and focusing on the variable gender, we could verify that only the female teachers mentioned the social persuasion experiences as TPE sources , while masculine gender ones referred almost exclusively mastery experiences. These results seem to meet the conclusions of Zeldin and Pajares (2000) and indicate a gender differentiation regarding the mastery experiences.