Teachers’ Social Networks and the Diffusion of Innovations

Download Report

Transcript Teachers’ Social Networks and the Diffusion of Innovations

Symposium Overview
Bill Penuel
SRI International
Complexity and Education
 August 2005: Gathering at SFI of scholars in education
inspired by how concepts of complex adaptive systems
might be applied to education
 Two areas of study had already begun to develop
– Complexity in education: Helping K-12 students
understand complexity
– Complexity of education: Studying and modeling
educational systems as complex systems
 Our symposium focuses on complexity in educational
systems
– Applications are nascent
– We are newcomers to the community
– We come bearing theory and preliminary models
Slide 2
Theoretical Frameworks
 Social capital theory
– Focus on resources and expertise that individuals access
through ties to others
– Resources and expertise are embedded in networks
 New institutionalism
– Isomorphism: Emergence of collaborative arrangements
as an adaptation to outside pressure
– Norms within institutional (professional) fields that are
constituted (and sometimes transformed) in local
interaction
 Complex systems theory
– Notion of making explicit assumptions so that systems
can be modeled
– Application to “systemic reform” efforts
Slide 3
Models and Methods
 Penuel and Riel: Integrating social network analysis into
multi-level models of school change processes
 Joshi: Using shadowing data to examine resources
exchanged in interaction and uncovering institutionalized
norms
 DiBello: Using simulations to help school districts
understand the costs of time away from educationally
relevant tasks
 Sabelli: Analyzing system levels toward developing
approaches to cross-case analysis
Slide 4
Agency and Timescales in Education
 Within education, there are agents who act at different
levels of the system
– Students: Classroom, Home, Out-of-School Time
– Teachers: Classroom and school
– Building leaders: School, district
– District administrators: School, district
– State and federal policy makers: State, federal agencies
– Reform intermediaries: All levels
 Institutional change happens on different timescales
– Age-graded classroom: 100 years
– Data-driven decision-making: 2 years
– Adopting a new curriculum: days, months
Slide 5
Applications
 Common framework for analyzing social capital in schools
– Gathering and interpreting network data
– Common data sets
 Knowledge building in educational research
– Unpacking “professional community”
– Cross-case analyses of schools (common language)
– Agent-based simulations for researchers
 Agent-based simulations for school leaders
– What-if scenarios for orchestrating collaboration
– Enabling leaders to understand schools and districts as
complex, multiscale systems
Slide 6
Teacher Networks and
the Diffusion of
Innovations
Bill Penuel and Margaret Riel
SRI International
Ken Frank and Ann Krause
Michigan State University
A Network Approach to Social Capital
Drawing on Portes (1998) and Lin (2001), we define social
capital in terms of:
 Ties: Interactions among faculty members in a school
 Resources and Expertise: The value of resources (e.g.,
curriculum) and expertise (e.g., wisdom of practice)
accessible through ties to others
This definition considers social capital as useful for
individual action, and secondarily as a social or collective
resource.
Diffusion of innovations is an emergent characteristic of the
school, which is facilitated by teacher talk and sharing of
resources about teaching.
Slide 8
A Network Approach to Social Capital
For us, analyzing networks is essential to measuring social
capital:
 Mapping the boundaries of networks and subgroups within
networks
 Including as part of our models the resources and expertise
one can access through those networks
 Considering the consequences in terms of changes in
teacher attitudes, teaching practice, and student
achievement
Several scholars have suggested that social capital has a
network structure (e.g., Burt, 2000; Lin, 2001), in that
valued resources are embedded within a network
Slide 9
A Network Approach to Social Capital
 Prior research
– Functions of network closure
• Resources and expertise flow freely within dense
networks
• Can protect a network from outside pressure
– Functions of bridging
• Critical source of new knowledge and skill
• People who play bridging functions can exert
considerable control over the flow of resources
 Implications
– Need to attend to network boundaries
– Boundaries exist within and across schools
Slide
A Crystallized Sociogram
Slide
Resources, Expertise, and Consequences
 Approaches to conceptualizing resources:
– Access to instructional materials
– Affordances of particular instructional materials (Stein & Kim, 2006)
– Resources-in-use (Cohen, Raudenbush, & Ball, 2003)
– Schoolwide norms (Bryk & Schneider, 2002)
 Approaches to conceptualizing expertise:
– Prior experience with a reform or activity
– Adaptive expertise framework (Bransford, Crawford)
– Pedagogical Content Knowledge (Ball, Hill)
– Formal Preparation (production-function literature)
– Teacher experience*
 Candidate consequences:
– Teaching quality
– Curriculum or reform implementation
– Student achievement
Slide
Our Research
Big Questions:
 How do interactions with colleagues affect teachers’ beliefs
and practices?
 What patterns of expertise flow in a school promote the
diffusion of innovations across a school?
 How do informal interactions combine with professional
development and intentional efforts to promote teacher
collaboration?
 How do teachers choose with whom to interact around
practice?
Slide
Study of Schoolwide Reforms
 Looking at different “home-grown” school-wide reform
initiatives
– Technology integration
– Literacy
– Data-driven decision making
– Creating standards-aligned assessments
 Measured social capital and self-reported influence on
practice at two points in time (Penuel, Frank, & Krause,
2006)
– We used “implementation levels” at Time 1 as an
indicator of expertise
– We looked at how getting help from an “expert”
Slide
influenced teachers reports that their school’s initiative
Sample Characteristics
 Schools
– 23 schools selected for commitment to collaboration and
engaged in whole school reform
– 13 elementary; 3 K-8; 5 middle; 2 high schools
– 8 high SES <10% free/reduced price lunch
– 3 low SES >80% free/reduced price lunch
 Teachers
– 499 teachers (with matched data for both surveys)
– teaching experience: 125 <5 years; 236 6-15 years; 190 teachers
16+ years
 School Leaders
– 29 informal leaders from 14 of the schools: designated coaches
for schoolwide initiatives; generally team or grade-level leaders
Slide
Two Analyses for Today
 Case studies of two schools’ reforms
– A look at the role of critical role of between-subgroup
dynamics in a school with respect to flows of resources
and expertise
 Hierarchical linear model looking at what predicts change in
instructional practice
– A look at between and within-subgroup dynamics and
their influence on teachers’ attitudes
Slide
Glade and Crosswinds
 Two schools in California’s II/USP Program
 Similar demographics and challenges in improving literacy
outcomes for English Language Learners
 Adopted similar approaches to reform: Promoting teacher
community facilitated by instructional coaches
 Coaches were expected to play slightly different roles
 Two schools met with dramatically different results:
– Glade: Still struggling for reform to take hold and gain
legitimacy
– Crosswinds: Steady gains in achievement, strong shared
commitment to reform goals
 Comparative case study analysis set out to test rival
explanations (Yin, 2003) for why
Slide
Formal Meetings at Glade and Crosswinds
Slide
Density of Interaction
Slide
Glade’s Subgroups
Veteran
Group
New Immigrant
Teacher Group
Slide
Crosswinds’ Subgroups
Early
Elementary
Group
Slide
Access through Ties to Expertise
Size and color
indicate extent of
use at time 1
A
B
provides help
to
C
D
Slide
HLM Analyses of Innovation Diffusion
 Models take into account expertise of colleagues with whom
teachers interact
 Three levels to model
– School
– Subgroup
– Individual teacher
Slide
Access through Ties to Expertise
Ripple around A indicates increase
in use between time 1 and time 2
A
provides help
to
B
C
D
Change in A is a function of
interaction with people with expertise;
the greater the mean expertise of
Helpers B, C, and D, the greater the
change in A
Slide
3-Level Model of Change in Practice
Slide
Discussion
 The case study results point to the importance of
the informal network in supporting or inhibiting the
flow of resources and expertise across subgroups
 Relationship to school (i.e., perceived collective
responsibility) is filtered by experience within
subgroup (i.e., fit with subgroup).
 The HLM analyses are suggestive of a way that
between and within-subgroup dynamics can
induce changes in teachers’ practices.
Slide
The Role of
Institutionalized
Norms of Autonomy
and Equality in
Shaping Interactions
of Teachers
Aasha Joshi
William R. Penuel
SRI International
Institutionalized Norms
 Norms emerge out of interactions
 “Specify how things should be done…designate appropriate
ways to pursue them [goals or objectives]” (Scott, 2001,
p55)
Slide
Equality as a Norm in Schools
 Network configuration
– Many ties within and across subgroups
– No single leader
 Direction of help
– Bias against seeking and giving help
 Interaction structure
– Collaborative “war” stories
– Experience-swapping
– Simply ignore reform (don’t believe there is anything to learn)
 Focus of talk
– Common challenges and struggles (e.g., students)
– Shared tasks
Slide
Autonomy as a Norm in Schools
 Network configuration
– Limited interaction among colleagues (Lortie, 1977)
 Direction of help
– Limited evidence of any kind of helping interaction
 Interaction structure
– Experience-swapping
– Collaboration as distribution of responsibility for tasks
 Focus of talk
– Talk about broad principles (e.g., standards) but not
much about teaching
Slide
Creating Disequilibrium: Efforts to Create
Teacher Community
 Network configuration
– Informal and formal leaders emerge in bridging roles
– Teachers also perform bridging roles to colleagues
 Direction of help
– Creates press to give and receive help
 Interaction structure
– Apprenticeship learning (e.g., model teaching)
 Focus of talk
– On teaching and on challenging and critiquing peers’
practice
Slide
Shadowing as a Way to Study
Transformation of Norms in Interaction
 The shadowing task:
– Single researcher followed shadowees from arrival to the school until
the end of their work day; debrief interview following the shadowing
day
 Documentation:
– Duration
– physical location (e.g., staff room, hallway, or telephone)
– intentionality (e.g., scheduled or impromptu)
– Topics
– Participants
– Comments by the participants during the interactions
 Data set:
– 6 schools
– 14 teachers
– 6 school leaders
– Coded: 798 interactions
Slide
Case Study Schools
Goals and Key Strategies of Reforms in Sample Schools
School
School 1
Goa ls
Improving reading achieve ment,
especially for English Language
Learners
Key Strategies
Use of common instructional planning te mplate
Mentoring and coaching of teachers
School 2
[Charter]
Improving instructional decisionma king through better use of data
Analysis of align ment among national standards,
state standards, and assessments (especially in
English/Language Arts)
Structured process for reviewing curricula r
resources
Frequent monitoring and assessment of students
School 3
Shared leadership with respect to
instructional decision ma king
Consensus process for deciding on instructional
strategies based on data (prima rily
English/Language Arts and mathemat ics)
Frequent monitoring and assessment of students
School 4
[Charter]
Analyzing and coordinating
effective instruction
“Cycle of Inquiry” in which teachers test and
collect data on efficacy of particu lar
instructional strategies
School 5
Improving reading achieve ment,
especially for English Language
Learners
Frequent monitoring and assessment of students
Mentoring and coaching of teachers
School 6
Improving reading and language
development
Two-way immersion curriculu m
Train ing and assistance from resource teachers
in school
Slide
Case Study Schools
Demographic Characteristics of the Schools
School
School 1
School 2
School 3
School 4
School 5
School 6
Student
Enrollment
Faculty Size
(FTEs)
Percent
Minority
817
521
297
274
724
520
39.0
27.5
16.4
15.3
45.0
26.0
88.7
55.3
50.2
16.4
73.2
71.2
Percent
English
Language
Learners
70.4
27.3
11.1
5.8
43.6
27.3
Percent Free
and Reduced
Price Lunch
100
18.6
47.5
5.5
66.4
49.8
Slide
Network Configuration
Slide
Direction of Help
 Bi-directional help is common among both school leaders and teachers
 School leaders give help more than they receive help
 Teachers receive help more than they give help
Aspect of Interaction
Direction of Help
Shadowee is giving help
Shadowee is receiving help
Bi-directional help
School Leaders
(n = 6)
M
SD
14.17
6.00
14.17
6.43
2.68
4.54
Teachers
(n = 14)
M
SD
6.43
9.43
20.29
4.62
4.16
9.45
t
3.06**
-1.85†
-1.50
Slide
Interaction Structure
 Modeling as help is rare; talk as help is most common
Aspect of Interaction
Nature of Interaction
Exchange of resources
Modeling instruction
Talk
School Leaders
(n = 6)
M
SD
11.67
2.50
27.00
6.68
2.59
5.76
Teachers
(n = 14)
M
SD
6.78
3.14
29.14
4.49
2.03
10.43
t
1.93†
-0.60
0.47
Slide
Focus of Interaction
 Instruction, coordination of activities, and school-level problems are
common topics of talk as help
Aspect of Interaction
Focus of Talk
Instruction
Assessment
Students
Other colleagues
School-level problems
Coordination of interactions,
resources, and activities
Uncertain/Shadowee is witness
School Leaders
(n = 6)
M
SD
Teachers
(n = 14)
M
SD
t
4.83
4.50
2.67
0.83
5.33
5.50
4.26
4.04
2.07
0.75
2.88
2.59
6.64
4.21
4.57
0.86
5.43
6.29
3.93
3.19
3.46
0.95
4.40
3.56
-0.92
0.17
-1.25
-0.05
-0.62
-0.49
3.17
2.32
0.79
1.12
-3.15**
Slide
Discussion
 Network configuration, helping patterns show some
evidence of perturbations with respect to norms of
autonomy and equality
– Networks are dense in schools, and interaction among
teachers is frequent, both in planned and impromptu
encounters
– There is some asymmetrical help, with designated
leaders providing it
 In these schools at least, the nature of interaction and focus
of talk there is evidence of equality and autonomy norms
– Lots of “experience swapping” remains
– Talk about instruction happened, but was not the main
focus of talk
Slide
Complex Systems and
Educational Change
Nora Sabelli, SRI International
Jay L. Lemke, University of Michigan
Balcones Conference
 2001: Meeting of 4 major NSF-funded projects in which
researcher-educator partnerships had sustained successful
collaborations for over 10 years
 Guiding Questions:
– What is the relationship between educational system,
research on the system, and the models of change used
in designing the work?
– What are the open problems with the existing models
that could shape the future of this type of work?
– Does the existence of models of change facilitate
scaling and adaptation of reform efforts?
– Is there a taxonomy of such models where these
projects and others have generated knowledge?
Slide
Internal and External Complexity of
Complex Adaptive Systems
 A complex adaptive system is situated in an environment:
– That environment is always more complex than the
system itself, and therefore, it can never be completely
predictable for the system, but the system depends on
regularities of the environment for maintaining [the]
energy supply needed to support its internal structures
and processes .
SFI Working Paper Abstract; 2003
Author: Juergen Jost
Paper # 03-12-070
Slide
Changes in Paradigms Based on Complex
Systems Theory
 Causal loops vs. causal chains (nonlinear networks)
 Integrated systems vs. isolable units of analysis
 Dynamical models and simulations vs. input-output
modeling
 Unique systems vs. generic systems
 Emergence vs. determinacy (surprise)
Slide
Complex systems theory provides a
thinking tool for:
• Qualitative reasoning about complex socio-natural
systems
• Making the infrastructure (human and technical)
assumptions, needs and opportunities more explicit.
• Quantitative modeling and simulation
Slide
Complex Systems Theory Provides a
Thinking Tool For
 Qualitative reasoning about complex socio-natural systems
 Making the infrastructure (human and technical)
assumptions, needs and opportunities more explicit.
 Quantitative modeling and simulation
Slide
Creating a Possible Framework to Make
the Infrastructure Visible
 Capacity to order and simplify
 Identification of significant features
 Congruence with reality
 Communicative power
 Explanation of a total process
 A basis for inquiry and research:
– How to build a model that specifies the relationships
between concepts
Slide
Goal: Cross-case Research
• It takes a village to study a village: who’s on the
team?
– The education system is a system
• Cross-project cumulativity of cases: meta-models
for research
• Education is local, research is general
• One example of a meta-model
• To annotate local case studies
Slide
Goal: Cross-Case Research
 It takes a village to study a village: who’s on the team?
– The education system is a system
 Cross-project cumulativity of cases: meta-models for
research
– Education is local, research is general
 One example of a meta-model
– To annotate local case studies
Slide
How is Learning Organized?
Transition across Levels
Standards
Demographic Trends
Standards
Instructional
workforce
Coherence &
Accountability
Teacher Recruitment and
policies
Teacher Certification
Standardized
StandardsTesting
System options
and constraints
Incentives
Local education needs
Instructional Leadership
and coherence
Distribution of internal
and external resources
Incentives
Alignment
Teacher Professional
Development
Available resources
Teacher Expectations
Assessment data available
Pedagogical Content
Knowledge
Which people learn (equity)
Why people learn (context)
How people learn
(cognition)
What people learn (content)
Slide
How is Learning Organized?
NSF Systemic Change
Drivers
Cohen et al.
Confrey et al.
Incentives and
Accountability
What is Known about Learning
Standards-based Curriculum
Evidence
Incentives
Resources
Coordination
Accountability
Student Outcomes
Instruction
Standards-based
curriculum
Professional Development
Why people learn (context)
Which people learn (equity)
How people learn (cognition)
What people learn (content)
Slide
Ups and Downs of Multiscale Selforganization
 Lower levels define the range of possibilities
 Higher levels constrain the emergence of consistent
patterns
 Intermediate levels buffer higher levels against fluctuations
from below
Slide
Some Key Questions for Adequate Models
 What are the range of timescales of the critical processes
that enable the system to maintain itself?
 What are its significant levels of organization (control
hierarchies) by characteristic structures and emergent
processes and patterns at each level?
 What kinds of material resource and information flows
connect adjacent and non-adjacent levels?
 How is information transformed, filtered, re-organized,
and added to from level to level?
Slide