Co-Principal Investigators: Brian Junker, Carnegie Mellon University; Richard Correnti, University of Pittsburgh

Purpose: The purpose of this project is to assess the efficacy of The Learning Walk routine as a strategy for developing school leaders. The Learning Walk routine, developed by the Institute for Learning at the University of Pittsburgh, is a form of "walkthrough" or "learning walk" practice in which school leaders conduct brief classroom visits on a regular basis for the purpose of observing classroom instruction and providing feedback to teachers.

Project Activities: The research team will conduct an experimental study involving approximately 80 elementary schools in a large urban school district. Half of the schools will be randomly assigned to using The Learning Walk routine and the other half will be randomly assigned to conduct "business-as-usual" according to school and district practice. At the treatment schools, the leadership teams (the principal and three instructional leaders) will be trained in The Learning Walk routine. The project will compare the outcomes between treatment and control schools on both proximal outcomes, such as school practice including collaborative routines, structure of practice, norms of practice, and classroom instruction and on the more distal outcomes of student performance on state standardized tests in reading and mathematics.

Products: Products from this project include articles published in peer reviewed journals on the efficacy of The Learning Walk routine on school practice and student achievement.

Setting: The study will take place within a large urban school district.

Population: A sample of approximately 80 elementary schools will be drawn from one of the United States' largest urban school districts serving a predominantly high-poverty, high-minority, and low-achieving student population.

Intervention: The Learning Walk routine, and its associated training and observational tools, were developed by the Institute for Learning at the University of Pittsburgh as the core of its leadership development program. It is comprised of a highly structured set of activities for the observation and interpretation of teaching and learning by a team composed of the principal and three additional leaders per school. The routine is based on repeated observation and interpretation of teaching and student learning anchored both in the official curriculum of the school and the enacted curriculum of the classroom. Its purpose is to build trust and mutual access among staff and provide routes by which new knowledge can enter a school's community of practice through training, observations, and discussions. In the first year of the project, all teachers are to rotate onto a Learning Walk team so that they will understand the routine and all teachers will be visited in their classroom by a team. At the end of the day, the team is to meet with the teachers whose classrooms were visited for a joint discussion of what was observed.

Research Design and Methods: The project will use a group-randomized control trial in which a sample of 80 schools willing to take part in the study will be randomly assigned to the treatment or control groups. Baseline data collection on school practice will be carried out at all schools and then the treatment schools will receive training in The Learning Walk routine. At the treatment schools, the school leadership teams (the principal and three instructional leaders at each school) are to receive 5 full-day training sessions and 5 half-day training sessions and the principals are to receive an additional 3 full-day and 5 half-day sessions. The comparisons between treatment and comparison school practice, classroom instruction, and student achievement will be based on data obtained through a school staff survey, student achievement data, and samples of teacher and student work. At 8 schools (4 treatment and 4 control), case studies will be done to examine whether and how school-level conditions mediate any observed effects. Specifically 8–10 staff members (all trained leaders, 2 informal leaders, and 4 non-leaders) will be interviewed each year and 3 days will be spent observing the use of The Learning Walk routine or other collaborative routines at each school.

Control Condition: Schools randomly assigned to the control group will conduct "business as usual" using existing school and district practices.

Key Measures: A web-based School Staff Questionnaire (SSQ) will be administered in both the treatment and control schools twice per year over 3 years to collect data on school practice–collaborative routines, structure of practice, norms of practice, classroom instruction, and also leadership team members' expertise in judging and critiquing the quality of classroom instruction. Data on classroom instruction will come from two sources: (1) items from the SSQ concerning instructional practices in both reading and writing and (2) collecting and reviewing teacher assignments and student work in order to assess the academic rigor of student-teacher interactions over content. Data on treatment dosage will be collected from three sources: (1) district and project staff logs of participation in training sessions, (2) artifacts from learning walks conducted in the treatment schools, and (3) items from the SSQ regarding school staff participation in The Learning Walk activities. Student scores on the state's standardized reading and math tests will be obtained from administrative data.

Data Analytic Strategy: Several levels of analyses will be done depending on the level of outcome. Impacts on school level practices will be estimated using a single level model, impacts on teacher practices will be estimated using two-level models (teacher in school), and impacts on student outcomes will be estimated using three-level models (student with teacher within a school). Additional analyses will be done to examine whether impacts change or accumulate over time.

Products and Publications

Journal article, monograph, or newsletter

Spillane, J.P. (2012). Data in Practice: Conceptualizing the Data-Based Decision-Making Phenomena. American Journal of Education, 118(2): 113–141.