Introduction The study of dispositions for effective teaching is a key topic in any discourse on educational reform. The purpose of identifying optimal dispositions are to a) screen potential teacher candidates during the admission process, and b) specifically develop and measure these dispositions and characteristics throughout the teacher education program. As we found in Phase One of this three-part study using the Teacher Behaviour Q-sort (TBQ), there is a consensus across teachers, administrators, and education faculty about how the diverse characteristics, skills and behaviours required by teachers should be ranked or ordered from most to least important. In this phase of the study we are examining the dispositions of 'Master' teachers at three distinct grade levels (primary, intermediate, and high-school) to assess whether there are distinct teaching styles, different ways to be excellent, within or across grade levels, and how well the consensus that we found earlier reflects these different styles. Theoretical Framework Assessing Dispositions Dispositions have been assessed by a variety of methods. For example, to assess the dispositions of applicants to teaching programs, B.Ed. programs may require a Letter of Intent in which the applicant outlines his or her reasons for pursuing a teaching career. Based on the Letter, the B.Ed. program can try to determine if the applicant has the necessary dispositions required to be a moral, ethical, and committed member of the teaching profession. Reference letters may also be required as a measure of an applicant’s aptitude and personal qualities. Other processes have focused on narratives, such writing a story (Kattner, 2010) or analyzing a case study (Mueller and Hindin, 2009), from which the admissions committee makes inferences about personal qualities and dispositions. Interviews are also utilized to assess a candidate’s dispositions and personal qualities. Byrnes, Kiger, and Schechtman (2003), Haberman (1987), and Calder Stegemann (2014) all reported a strong relationship between intake interviews and success on the practicum. All of these methods may provide some information about an applicant’s suitability vis-à-vis dispositions, ethical, and moral behaviour. However, assessment tool reliability, socially desirability, cost of training, and time constraints affect the choice and use of these methods. Some researchers have attempted to tap dispositions more directly and quickly using checklists or self-reports. For example, Anderson and Brydges (2010) developed a checklist and assessment rubric; however, their work found that TCs (particularly those who are the least successful on the practicum) tended to over-inflate their ratings. Similarly, Schulte, Edick, Edwards, and Mackiel’s Teacher Disposition Index (2004) lacked strong predictive power, but could be used as a self-screen for potential applicants. Rinaldo and Slepkov (2012) have also developed a checklist, with strong predictive power, to assess dispositional performance, however, the checklist is for use only during the teacher education program, and not as an admission assessment. The Q-sort Method The Q-sort method has been used extensively in the field of psychology to assess personality characteristics in children, youth and adults, and also used successfully in an educational context to measure teachers' beliefs about classroom discipline (Rimm-Kaufman, Storm, Sawyer, Pianto, & LaParo, 2006). Because the individual must place a specified number of cards into each category, he is ‘forced’ to choose between items – not all those that are desirable can be rated as 'most characteristic'. Instead, items must be ranked and priorities set, a process that has the further benefit of constraining social desirability. In the first phase of our study, we used the Q-sort method to rank the dispositions required for successful teaching. Practicing teachers, school administrators, and university education faculty described an ideal teacher by sorting 100 characteristics (each on separate cards) into nine groups ranging from most to least desirable or necessary for a particular grade level. As we reported in a poster session at CSSE in 2013 (Calder Stegemann & Roberts, 2013), (1) we found good agreement across our informants concerning the ranking or relative importance of the dispositions identified in our Teacher Behaviour Q-sort (TBQ). (2) When we examined individual items, we found several differences in those characteristics that were most highly valued at the primary versus high school levels. For example, being caring, warm, and responsive were more important for primary school teachers, whereas strong subject-area knowledge was more important for high school teaching. (3) Many qualities were shared across grade levels, for example valuing and respecting students and effectively instructing and understanding the instructional/assessment/reporting cycle. Having demonstrated that there is a reasonable consensus about which of the diverse behaviours of an ideal teacher are more important and which are less, we are using the Teacher Professional Q-sort to examine the issues of teaching styles, by having master teachers across three distinct grade levels describe their own teaching practices. If there are distinct teaching styles, different ways to be excellent, within or across grade levels, knowledge of these styles can aid us in screening and training teacher candidates. Research Objectives The immediate goal of this research is to assess whether there are different teaching styles among master teachers, different ways to be excellent in the classroom. The ultimate goals of this research are to a) improve the admission process to teacher education programs and b) measure growth in professional dispositions during teacher training. This paper reports the findings from the second year of our study, which focuses on the feedback of "Master" teachers. Method Participants and Procedures There are three phases to this study, extending over a 30-month period. During the first phase we tested the Q-sort method and refined character descriptors. This entailed having practicing teachers, administrators, and B.Ed. faculty members complete the Q-sort to establish the teacher dispositions deemed most necessary for successful teaching. The second phase of the study involves "Master" teachers at the primary, elementary, and high school levels. Phase three will involve TCs from their First and Second Years of our program doing the same Q-sort as that used in Phase Two. Phase two - Field testing the Teacher Behaviour Q-sort. School principals in five surrounding school districts were approached via school district email and asked to nominate one or two individuals from their staff who they considered to be “Master” teachers. The school principals then forwarded our letter of invitation to these nominees. Participants were sent the Q-sort cards and a return envelope. 54 Master teachers (MTs) from three different school districts in our region agreed to participate – 19 primary (grades K-3), 14 intermediate (grades 4-8), and 21 high school (grades 9-12). Analysis Following Block (1971), we assessed possible heterogeneity in our master teachers by performing a maximum likelihood factor analysis (MLFA) of their Q-sorts, which is recommended by many authorities for exploratory purposes (e.g., Costello & Osborne, 2005). In Block’s approach, the data matrix is transposed so that teachers form the columns and Q-Sort items the rows. The factor analysis then clusters teachers who are similar to one another instead of variables that are similar to one another. The transposed data set can also be used to calculate correlations between criterion Q-sorts (mentioned above) and individual master teachers. Differences between primary, intermediate, and high school master teachers were initially assessed by analyses of variance, with significant differences being further evaluated by discriminant analyses, as reported below. Results and Interpretations Are there different types of master teachers? Within each level (primary, intermediate, and high school) only one factor emerged for each group with an eigenvalue greater than one. For the 19 primary master teachers, this factor accounted for 70% of the variance, and loadings for the individual teachers ranged from .70 to .94. For the 14 intermediate teachers, the first factor accounted for 71% of the variance and loadings for the individual teachers ranged from .74 to .91. For the 21 high school teachers, the first factor accounted for 72% of the variance, and loadings for individual teachers ranged from .80 to .89. Thus within each level, teachers described themselves in very similar ways. This homogeneity was also reflected in consistently high correlations between the individual teachers and the ideal descriptions that we developed earlier (Calder Stegemann, & Roberts, 2013). Correlations for primary teachers ranged from .69 to .92 (mean = .81), while those for intermediate teachers ranged from .64 to .87 (mean = .78). Correlations for high school teachers ranged from .51 to .87 (mean = .78). How are master teachers at different levels different? Although teaching requires a core set of skills across elementary, intermediate, and high school, we did expect some differences to emerge, especially between primary and high school teachers (with intermediate teachers showing fewer differences because of the intermediate status of their students). We explored this possibility by conducting a series of discriminant analyses. Primary teachers versus high school teachers. As shown in Table 1, significant differences (p < .05) between primary and high school teachers emerged for 14 of the 100 items in the TBQ, and binomial tests rejected the omnibus null hypothesis that all differences were zero, p < .001. Many of these differences reflected the different demands faced by primary and high school teachers. For example, being knowledgeable about children’s social and cognitive development, their individual differences and special needs, and considering the advice of parents about their children, were all more salient characteristics for primary teachers than for high school teachers. In addition, primary teachers described themselves as relatively more physically active. In contrast, contributing their expertise to professional activities, being responsible, and behaving professionally were more salient characteristics for high school teachers. In a discriminate analysis, knowledge of children’s social development correctly classified 89% of primary teachers and 71% of high school teachers, for an overall success rate of 80%. Intermediate teachers versus primary teachers; intermediate teachers versus high school teachers. As anticipated, fewer significant comparisons were found between intermediate teachers and the other two groups. Only eight of 100 comparisons with primary teachers were significant at .05, and binomial tests retained the omnibus null hypothesis that all differences were zero, p = .23. Similarly, only seven of 100 comparisons between intermediate and high school teachers reached statistical significance, and binomial tests again retained the omnibus null hypothesis, p = .13. Although, overall, differences occurred only at chance levels, some made theoretical sense. We will describe these briefly, because it is appropriate to be concerned about type II error in an exploratory study. For example, warmth and empathy were more salient for primary teachers than for intermediate teachers, as were knowledge of children’s social and cognitive development (as they had been when high school teachers were the comparison group). In a supplementary discriminant analysis, knowledge of children’s social development correctly classified 82% of primary teachers and 64% of immediate teachers, for an overall success rate of 74%, Wilks’ Lambda (1,1,29) = 0.82, p = .018. Similarly, compared to high school teachers, intermediate teachers described themselves as relatively more physically active (just as primary teachers had done). Intermediate teachers were also more likely to consult with parents about their children. In contrast, being responsible was a more salient characteristic for high school teachers (as it was when primary teachers were the contrast group). In a supplementary discriminant analysis, physical activity and consulting with parents correctly classified 79% of intermediate teachers and 75% of high school teachers, for an overall success rate of 76%, Wilks’ Lambda (2,1,32) = 0.68, p = .003. Table 1. Mean differences in the self-descriptions of primary school (N = 17) and high school (N = 20) master teachers. Item primary mean (SD) high school mean (SD) eta-squared 7. Knowledgeable about children’s social development 7.8 (1.2) 6.0 (1.3) .37 12. Considers the advice of parents about their children 6.8 (1.1) 5.7 (1.2) .23 71. A genuinely dependable and responsible person. 6.0 (1.2) 7.3 (1.2) .22 8. Knowledgeable about children’s cognitive development and learning abilities 8.0 (1.1) 6.8 (1.3) .21 55. Shy and reserved, makes social contacts slowly. 3.7 (0.9) 3.0 (0.6) .20 20. Contributes own expertise to activities offered by schools, districts, professional organizations 4.7 (0.8) 5.8 (1.3) .19 10. Knowledgeable about children with special needs 6.8 (1.2) 5.9 (1.0) .16 73. Open and straightforward. 5.3 (1.2) 6.2 (1.2) .14 75. Tends to undermine, obstruct or sabotage others. 1.3 (0.6) 1.0 (0.0) .14 68. Responds to humour. 5.7 (1.5) 6.7 (1.3) .12 57. Physically active. 5.0 (1.3) 4.3 (0.6) .12 79. Tends to keep thoughts, feelings, or products to self. 3.9 (0.9) 3.4 (0.7) .12 91. Restless and fidgety. 3.7 (0.9) 3.2 (0.4) .12 9. Knowledgeable about individual differences in how children learn 7.9 (1.0) 7.1 (1.3) .11 Note. All eta-squared values significant at p < .05. Items in bold entered the discriminant analysis. Conclusions We are in the second of three phases of a study to develop and refine an innovative method of assessing teacher dispositions - the Teacher Professional Q-Sort. In this phase, we found that master teachers within each of the three levels that we examined (primary, intermediate, and high school) were reasonably consistent with one another in their self-descriptions, and consistently showed strong correlations with descriptions of an ideal teacher. This suggests that there is a common core of dispositions and values that inform good teaching. Our results also suggest that teaching practices differ in sensible ways across grade levels. For example, knowledge of children’s social development and cognitive development are more salient for primary teachers than for either intermediate or high school teachers. High school teachers are less likely to consult with parents than are either primary or intermediate teachers, consistent with the greater maturity and independence of their students; and they are more likely to emphasize professional service and responsibility in their self-descriptions. We expect similar analyses and findings to guide our exploration of teacher candidates, as we assess admission characteristics and progress in a teacher training program. References Anderson, D., & Brydges, B.C. (2010, November). Dispositional field of dreams: We built it and they came. Now what? An analysis of four semesters of assessment data identifying and improving the performance of struggling and high risk teacher candidates through the use of a dispositional assessment rubric. Paper presented at the Seventh Annual Symposium of Educator Dispositions. Kentucky. Byrnes, D.A., Kiger, G., & Schechtman, Z. (2003). Evaluating the use of group interviews to select students into teacher – education programs. Journal of Teacher Education, 54(2), 163 - 172. Calder Stegemann, K.J. (2014). Predicting teacher candidates’ success in their final school practicum. US-China Education Review A, 4(2), 77-86. Calder Stegemann, K.J., & Roberts, W.L. (2013). Assessing teacher dispositions using the Q-sort method. Poster presented at the annual meeting of the Canadian Society for Study of Education, Victoria, BC. Costello, A. & Osborne, J. (2005). Best practices in the exploratory factor analysis: four recommendations for getting the most from your analysis. Practical Assessment, Research and Evaluation, 10, 1-9. Retrieved from http://pareonline.net/getvn.asp?v=10&n=7 Haberman, M. (1987). Recruiting and selecting teachers for urban schools. ERIC Clearinghouse on Urban Education, New York, NY. ED 292942 Kattner, T. (2011). Creative assessments can uncover promising applicants. Recruitment and Retention In Higher Education, 25(3), 2-4. Kosnik, C., Brown, R., & Beck, C. (2005). The pre-service admissions process: what qualities do future teachers need and how can they be identified in applicants? New Educator, 1(2), 101-123. Mueller, M. & Hindin, A. (2009). An examination of preservice teachers’ development of dispositions and the factors that influence them. Teacher Education and Practice, 22(2), 143 -164. Rimm-Kaufmn, S.E., Storm, M.D., Sawyer, B.E., Pianta, R.C., & LaParo, K.M. (2006). The Teacher Belief Q-sort: A measure of teachers’ priorities in relation to disciplinary practices, teaching practices, and beliefs about children. Journal of School Psychology, 44, 141-165. Rinaldo, V., & Slepkov, H. (2012, May). Measuring teacher dispositional performance in a three-semester program. Paper presented as the annual meeting of the Society for the Study of Education, Waterloo, Ont. Schulte, L, Edick, N., Edwards, S., & Mackiel, D. (2004). The development and validation of the Teacher Dispositions Index. Essays in Education, 12. Retrieved from http://www.usca.edu/essays/vol12winter2004.html SSESSING TEACHER DISPOSITIONS USING THE Q-SORT METHOD - Phase 2 - MASTER TEACHERS CATE – CSSE 2014 MASTER TEACHERS: CONSISTENCY AND DIVERSITY AMONG DISPOSITIONS CATE – CSSE 2014