460 likes | 606 Views
Effects of the My Teaching Partner Intervention in Secondary School Classrooms. Joseph P. Allen Robert C. Pianta University of Virginia. Co-Collaborators: Amori Mikami Anne Gregory. Project Team: Chris Hafen Sharon Deal Judith Wasserman Rachel Boren Janetta Lun. Context.
E N D
Effects of the My Teaching Partner Intervention in Secondary School Classrooms Joseph P. Allen Robert C. Pianta University of Virginia Co-Collaborators: Amori Mikami Anne Gregory Project Team: Chris Hafen Sharon Deal Judith Wasserman Rachel Boren Janetta Lun
Context Number of Secondary School Students in U.S.: 24 million Number of Secondary School Classes being taught each week 6 million % of 9th graders who won’t finish High school by the end of 12th grade 25% Number of programs in ‘What Works’ Clearinghouse with demonstrated efficacy improving teaching quality enough to improve student achievement in these classrooms 0
Key Questions • Can we identify teacher-student interaction qualities that predict student engagement and achievement? • Can we change these qualities? • Will changes lead to sustainable student achievement gains? • What are the mechanisms of change?
Key Questions • Can we identify teacher-student interaction qualities that predict student engagement and achievement? • Can we change these qualities? • Will changes lead to sustainable student achievement gains? • What are the mechanisms of change?
Classroom Learning Assessment & Scoring System- Secondary (CLASS-S) Instructional Support Instructional Learning Format Content Understanding Analysis & Problem Solving Quality of Feedback Classroom Organization Behavior Management Productivity Emotional Support Positive Climate Teacher Sensitivity Regard for Adolescent Perspectives Negative Climate Student Outcomes Student Engagement
Observational Assessment of Classroom Environment • Videotaped observations of a classroom • spread throughout course of year • Two 20-minute segments per class session/tape • Each tape rated by 2 raters • Coded Using CLASS-S System • High inter-rater reliabilities; ICC’s range from • .73 - .82 for overarching domains • .50 -.78 for specific dimensions (all but one dimension > .64)
Student Academic Success • Score on State “Standards of Learning” End of Year Subject Test • The measure by which schools/students are judged for accreditation/graduation. • Extensive seven-year validation/standardization process.
Evaluation Design 88 classrooms 45 Treatment 43 Control (Classrooms Randomized within school) (640 Control Students in Year 1) 2237 Students across two years
Classroom Characteristics • School type:39% High school; 61% Middle School • Subject:52% Language/Social Studies; 48% Math/Science • Average class size:23 students • Gender: 47% girls 53% boys • Ethnicity:23% African American 2% Asian 4% Hispanic 70% European-American
Teachers • 88Teachers (45 treatment; 43 control) within 8 schools • 1 focal classroom selected per teacher • Teacher Demographics: • 64% female • 83% White, 8% African-American; 6% Mixed-Ethnicity; 3% Other • 54 middle school, 34 high school • 35% BA degree; 65% at least a year of course work beyond BA • Average 8 years of teaching experience
Analytic Approach • Initial analyses only with control group. • All models covary: • Student factors: • Grade level • Gender • Family poverty status • Classroom factors: • Classroom size • Teacher Factors • Teacher experience • Teacher education • Teacher gender and race • Moderating effects of covariates are also examined.
Preliminary Findings • School-level effects • Appear minimal (classes, not schools matter most) • Teacher/Classroom-level effects • Larger and more consistent • Baseline Ach. scores account for 44% of variance in End of Year Ach. scores (i.e., Achievement is relatively stable) • After accounting for baseline scores, 27% of remaining variance is explained at the classroom/teacher level (i.e., classroom effects do matter).
Predicting Student Achievement(Control Group Only) Covariates were all non-significant. No effect of subject matter/content area.
Key Questions ✔Can we identify teacher-student interaction qualities that predict student engagement and achievement? • Can we change these qualities? • Will changes lead to sustainable student achievement gains? • Why?
MyTeachingPartner Overview • Consultant and teacher work together using the CLASS-S in cultivating: • Observation • Reflection • Development of knowledge and expertise Reflection Knowledge Expertise Support Teaching Practice Classroom Observation
MTPS Website www.mtpsecondary.net
Video Library www.mtpsecondary.net
Detailed Video Examples www.mtpsecondary.net
Key Questions ✔Can we identify teacher-student interaction qualities that predict student engagement and achievement? • Can we change these qualities? • Will changes lead to sustainable student achievement gains? • What are the mechanisms of change?
Evaluation Design Treatment group: Year 1: • Introductory Workshop (late summer) • Ongoing consultancy • ~ 2 days total in-service time Year 2: • Booster Workshop (late summer) only + Web site access • Control group: Usual in-service practice.
Evaluation Design 88 classrooms 45 Tx. 43 Control (Classrooms Randomized within school) 2237 Students
Year 1 Change in OverallTeacher-Student Interactions Standardized Effects: Baseline = .45*** Intervention = .19* MTPS participation predicts higher quality teacher-student interactions
Intervention Effect on Change in Indirectly Targeted Classroom Qualities
Key Questions ✔Can we identify teacher-student interaction qualities that predict student engagement and achievement? ✔Can we change these qualities? • Will changes lead to sustainable student achievement gains? • What are the mechanisms of change?
Year 1Intervention Effects on Achievement • No relation of intervention to either baseline or exit achievement test scores in Year 1 (all p’s > .35). • Why? • No evidence we changed the classroom until the very end of the year when most teaching was past.
Year 2Change in Achievement Standardized Effects: Pre-test = .54*** Intervention = .22* MTPS is predicting increases in End of Course Achievement Tests
Year 2Intervention Effects on Achievement • Real-world effect size = .22 SD increment in Achievement Test scores • Average ‘Bump’ of students in MTP from 50th to 59th percentile in achievement • If effect applies equally at all parts of achievement spectrum (as appears to be the case): a 16 point boost would reduce failure rates from: 14% without the intervention to 10% with it Reducing the number of failing students each year by 29% *** This occurs in the year AFTER the intervention year (i.e., sustainability), across diverse subject matter/content areas.
Key Questions ✔Can we identify teacher-student interaction qualities that predict student engagement and achievement? ✔Can we change these qualities? ✔Will changes lead to sustainable student achievement gains? • What are the mechanisms of change?
A Preliminary Mediational Analysis Intervention Environmental Outcome Observed Change in Student Achievement “My Teaching Partner” Intervention *
A Preliminary Mediational Analysis Intervention Intervention Target Environmental Outcome ?? Observed Change in Student Achievement “My Teaching Partner” Intervention
Mediational Analyses • Assessed via Multi-level Structural Equation Modelling, followed up via parametric bootstrapping analysis (Preacher et al., 2010) • Focus on target of intervention (Teacher-student interactions assessed via CLASS-S) • Using Centered/Standardized data for ease of interpretation.
MTP-S Effect as Mediated via Observed Interactions Intervention Intervention Target Student Outcome Observed Teacher-Student Interactions .16** .37** .12* Change in Student Achievement “My Teaching Partner” Intervention Initial Model (Simple Direct Effects)
MTP-S Effect as Mediated via Observed Interactions Intervention Intervention Target Student Outcome .06* * Indirect effect Observed Teacher-Student Interactions .16** .37** .12* Change in Student Achievement “My Teaching Partner” Intervention .06 ns Initial Model (Simple Direct Effects) Final Model (Including Mediated Effect)
Limitations • Design only supports causal interpretations for outcomes, not for mediating processes with analyses thus far. • Attrition Concerns (though positing attrition ‘sleeper’ effects in year 2 achievement data seems like a stretch). • Modest statistical significance
Conclusions • We CAN identify elements of the classroom environment that predict student achievement. • We CAN change these environmental factors. • If we do, student achievement will change as well, eventually. • Changes can be sustained over time and in new classrooms, post-intervention. • We can identify potential mechanisms of change linked to the intervention.
Potential Significance – Costs vs. Benefits (BOE* Calculation) *BOE = Back of Envelope