The Long and Winding Road

1
The Long and Winding Road The Relationship
Between Leadership Practice and Student
Performance Jonathan Supovitz Philip
Sirinides University of Pennsylvania Consortium
for Policy Research in Education June, 2008 IES
Annual Conference
This research is supported by the Institute of
Education Sciences, grant R305C050041.  The
opinions expressed are those of the authors and
do not represent the views of the U.S. Department
of Education.
2
Study Overview
The literature on how school leadership improves
student outcomes stretches back at least 40
years. Initially, studies of school leadership
focused on the role and influence of school
principals. More recently, researchers have
started to explore the role and influence of
other school actors as well. In this study we
seek to examine the relative influence of
different school actors on instructional practice
and student outcomes.
3
Literature Review on School Leadership and
Student Outcomes
Several comprehensive literature reviews of the
effects of principal leadership on student
outcomes.

• Hallinger Heck (1997) examined 43 studies
  conducted between 1980 and 1995. Found little
  evidence of direct effects, with most evidence
  pointing to indirect effects. Concluded that
  principals have a measurable but indirect effect
  on school effectiveness and student achievement.
• Leithwood, Seashore Louis, Anderson and Wahlstrom
  (2004) conducted a wide-ranging analysis of
  educational leadership and concluded that
  leadership is second only to teaching among
  school-related factors in its impact on student
  learning.

4
Literature identifies several important principal
leadership characteristics
? Instructional mission for school (Goldring
Pasternak, 1994 Hallinger, Bickman Davis,
1996). ? Encouraging trust and collaboration
(Fink Resnick, 1991 Blase Blase, 1989) ?
Hands-on support of instruction (Hallinger
Murphy, 1987 Supovitz Poglinco, 2001)
5
Growing Literature on Range of Leadership
Activity in Schools

• Individuals at all levels of schools influence
  reform efforts (York-Barr Duke, 2004)
• Distributed leadership perspective (Gronn, 2000
  Spillane 2006)
• Interest in and attention to teacher leadership
  (Mangin Stoelinga, 2008)

6
Attributes of teacher influence on peers
identified in the literature
? Instructional conversation amongst peers
(Silva, Gimbert Nolan, 2000 LeBlanc Shelton,
1997) ? Active interaction amongst faculty
around teaching and learning (Wasley, 1991
Smylie Denny, 1990) ? Strong instructional
advice networks (Frank et al. 2004 Supovitz,
2008)
7
Mission Goals
Community Trust
Principal Leadership
Focus on Instruction
Student Learning
? Instruction
Instructional Conversation
Peer Influence
Interaction around TL
Advice Networks
Conceptual model
8
Research Questions

1. What is the relationship between teacher reported
   measures of leadership practice and peer
   influence with student achievement when including
   the indirect effects of teacher reported change
   in instructional practice?
2. What are the relative effects of teacher reports
   of principal leadership and teacher reports of
   peer influence on changes in instructional
   practice?

9
Data
? Survey data collected in 2007 via an IES funded
evaluation of a principal leadership
initiative. ? Survey of 1,839 school faculty
members in 38 elementary/middle schools (response
rate 81). ? Focused only on 1,079 responding
teachers of English and mathematics (and general
elementary). ? End of year state test results in
English and mathematics for 14,027 students in
2006 and 2007. ? Matching process resulted in
7,793 students and 519 teachers (English)/502
teachers (mathematics). ? 76 overlap in
teacher samples.
10
Measures Reliabilities

• Principal Leadership Scales
• Faculty Trust 6 items (ELA .93, Math .93)
• Instructional Goals 5 items (ELA.91 Math
  .90)
• Instructional Expertise 5 items (ELA .91
  Math .91)
• Peer Influence Scales
• Instructional Conversation 5 items (ELA .86
  Math .84)
• Interaction around Teaching and Learning 4
  items (ELA .78 Math .77)
• Advice Networks 2 items (ELA .82, Math .84)
• Change In Teacher Practice Scale 7 items (ELA
  .94 Math .95)

11
Two Phase Analysis
? Phase 1 Teachers linked to students Parallel
HLM analyses in English Math ? Phase 2
Structural equation models examining
relationships between indicators of principal
leadership, peer influence, and changes in
teacher instructional practice.
12
Measure of Student Learning
2006 and 2007 end-of-year state
assessment Different tests in different years ?
Transition to new state assessment by year,
grade ? Tests not vertically equated ?
Re-standardized within year and grade Change in
student rank ? Difference scores as the student
outcome ? Measure of students relative change in
ranking
13
Results Relationship between change in practice
and student learning (Random Effects Null Model)
ELA Math Student 91 88 Teacher 9 10 School
0 1
14
Results Relationship between change in practice
and student learning (Fixed Effects Model)
? Instruction
Student Learning
ELA Math
Change in Instruction .06 .01 (.02) (.02)
Note Model includes covariates for minority
status, gender, LEP, and Free-Reduced Lunch by
group centering at levels 1 and 2.
15
Results Model of Change in Instruction in ELA
Mission Goals
.87
Community Trust
Principal Leadership
.62
.98
.06
Focus on Instruction
? Instruction
Instructional Conversation
Peer Influence
Interaction around TL
Advice Networks
16
Results Model of Change in Instruction in ELA
Mission Goals
.87
Community Trust
Principal Leadership
.62
.98
.06
Focus on Instruction
? Instruction
.41
Instructional Conversation
.83
.28
Peer Influence
.61
Interaction around TL
Model Fit CFI .95 RMSEA .05
.27
Advice Networks
17
Results Model of Change in Instruction in Math
Mission Goals
.82
Community Trust
Principal Leadership
.61
.97
.07
Focus on Instruction
? Instruction
.39
Instructional Conversation
.79
.28
Peer Influence
.63
Interaction around TL
Model Fit CFI .94 RMSEA .05
.33
Advice Networks
18
Discussion

• Educational leadership influences practice which
  changes performance.
• Results are consistent in both English and math
  models.
• Principal instructional expertise is the most
  salient perceived principal leadership
  characteristic related to changing teacher
  instructional practice.
• Instructional collaboration amongst peers is the
  most influential teacher characteristic related
  to changing teacher instructional practice.
• Peer influence is more influential than principal
  leadership on changing teacher practice in both
  English and mathematics.

19
Next steps

• Combine both phases of the study into one
  multi-level model so we can examine both direct
  and indirect effects simultaneously.
• Look at effects of principal leadership on peer
  influence related to changing instruction.
• Add contextual effects of school characteristics
  into the models.
• Analysis of in-sample and out-of-sample teachers
  to assess attrition bias.

20
The Long and Winding Road The Relationship
Between Leadership Practice and Student
Performance Jonathan Supovitz Philip
Sirinides University of Pennsylvania Consortium
for Policy Research in Education June, 2008 IES
Annual Conference
This research is supported by the Institute of
Education Sciences, grant R305C050041.  The
opinions expressed are those of the authors and
do not represent the views of the U.S. Department
of Education.