Learning Leadership Matters: Teachers’ experiences of innovatively and conventionally prepared principals

1. Learning Leadership Matters:Teachers’ experiences of innovatively and conventionally prepared principals Paper presented by: Terry Orr, Bank Street College Stelios Orphanos, Stanford University AERA, Chicago, April 2007

2. Purpose To investigate the relationship between innovative leadership preparation and perceptions of graduates’ leadership practices and impact on selected school improvement and teacher satisfaction, collaboration and engagement

3. Background • Centrality of leadership to school improvement and performance • Emerging policy emphasis on highly effective principals, through recent NCLB Commission recommendations • State policy emphasis on leadership preparation through licensure and program accreditation • Program reforms to meet national standards and local district conditions

4. Theoretical background • Transformational/instructional leadership as a mediating influence on school improvement (Leithwood & Jantzi, 1999) • Leadership as a set of skills and qualities that can be developed through strategically designed content and active adult theory-based strategies (Dvir, et al, 2004; Jackson & Kelly, 2002; Orr, 2006)

5. Research background • Teacher engagement, collaboration and job satisfaction as significant mediating factors in improved student outcomes (Geijsel, et al, 2001; Pearson & Moomaw, 2005) • Leadership-teacher outcome relationship (Leithwood, et al, 2004; Waters, et al 2005) • Mediating effects of school improvement strategies (professional development, distributed leadership, parental involvement) (Geijsel, et al, 2001; Sweetland & Hoy, 2000; Sebring, et al, 2006) • Leadership preparation-leadership practices relationship (Leithwood, et al 1996; Orr & Barber, 2006)

6. Study support • Study was made possible by a grant from the Wallace Foundation for the Stanford University School Leader Study • Study was directed by Linda Darling-Hammond and Debra Meyerson, Stanford University; Michelle LaPointe is the study’s research director. • The results are based on the combined work of the research team to document eight innovative programs and construct case studies of sampled schools in each site.

7. dislead teacher satisfaction profdev Preparation leadership parental support teacher engagement Figure 1 Revised conceptual model

8. Methodology • nonexperimental research design • survey research methods • two teacher samples • confirmatory factor analysis and structural equation modeling

9. Hypotheses • Hypothesis 1:The more positive teachers’ rating of principal leadership, the greater their perceived job satisfaction and teacher engagement and collaboration. • Hypothesis 2: Whether principals had innovative preparation experiences will have a moderating influence on the principal leadership-teacher outcome relationship. • Hypothesis 3a. The extent of professional development participation and of distributed leadership in the school will mediate the leadership-teacher outcome relationship • Hypothesis 3b: The extent of parental support will mediate the leadership-teacher outcome relationship

10. Samples • Stanford University study sample (2005): • Five states (CA, CT, KY, MS, NY) • Principals in one of four innovative leadership preparation or one of four innovative in-service programs • 16 elementary and middle schools • 389 teachers (214 with exemplary prepared prinpls.) • SASS study sample (1999-2000): • Restricted to urban schools and principals with five years or less experience • 16 elementary and middle schools • 855 teachers

11. Survey instrument • Stanford University teacher survey, with SASS items • SASS teacher survey

12. Sample characteristics

13. Innovative Preparation Programs

14. Teacher perceptions of teacher cooperation and engagement

15. Principal leadership measure

16. *

17. Evaluation of the Dataset • The scale items and other measures were evaluated for their measurement error: • Most of the data are ordered categorical data, so polychoric (PC) correlations are more appropriate. • The assumption of Univariate normality was rejected for all variables, and the assumption of multivariate normality is assumed to not hold either. • Our estimation method was Weighted Least Squares (WLS), to take the non-normality of the variables in consideration. • Our sample (n=1253) satisfies WLS requirement of a large sample size.

18. Confirmatory Factor Analysis • Tested the hypothesized measurement (factorial) model using confirmatory factor analysis (CFA) with LISREL 8.8. • Thirteen indicator variables were specified that loaded on four different latent figures that were left to be freely correlated • The data fitted the model well. • The data did not fit the model well when conventional leadership preparation was used instead of innovative leadership preparation.

19. CFA fit indices • The chi-square statistic (as an index of absolute fit) was significant enough to reject the model, but the large sample size and use of WLS inflates this statistic. • The goodness-of-fit index (GFI)-- analogous to the familiar R2 statistic -- was 0.993. • Using a cut-off rule of 0.05 for all other indexes (Hu & Bentler, 1999) shows a good fit: • The RMSEA is 0.0317 with a 90% confidence interval of (0.0242, 0.0392), • the Normed Fit Index (NFI) is 0.969, • the Non-Normed Fit Index (NNFI) is 0.977 and the Comparative Fit Index (CFI) = 0.983 • the Root Mean Square Residual (RMR) -- the average residual value derived from fitting the variance-covariance matrix of the hypothesized model to the variance – covariance matrix of the sample data is 0.040 –also less than 0.05 (Bollen, 1989). • Of the 91 standardized residuals 16% are greater than 2.58 (15 out of 91) a satisfactory fit, although not perfect

20. Assessing the structural equation model (using WLS) • All reported effects are statistically significant at the 0.05 level of significance. • Innovative preparation predicts leadership practices. • Leadership practices predict teacher professional development, distributed leadership and teacher job satisfaction and engagement, both directly and indirectly. • Teacher job satisfaction predicts engagement and collaboration.

21. Figure 2: revised model

22. Other findings • Perceived parental support has a moderating influence on leadership and teacher job satisfaction, suggesting that it serves as a context measure. • The moderating influence of leadership preparation on teacher professional development participation may reflect the unaccounted for influence of strategic district emphasis on both leader preparation and teacher professional development. • Distributed leadership has an indirect influence on teacher outcomes, through professional development participation.

23. Conclusions • Conceptual validation of the influence of leadership preparation on leadership and school outcomes • Findings consistent with foundational research • Methodological direction: • Refined measurement, with greater differentiation on leadership preparation • Accounting for district support in the model • Other school improvement related outcomes

24. Contact information Margaret Terry Orr Bank Street College morr@bnkst.edu; 212-875-4546 Stelios Orphanos Stanford University orphanos@stanford.edu