190 likes | 314 Views
Using a Dose-Response Analysis Strategy to Measure Outcomes of Place-based Education. Prepared by: Michael Duffin, PEER Associates, Inc. With support from: the Place-based Education Evaluation Collaborative (PEEC) For presentation at: the NAAEE Pre-conference Research Symposium
E N D
Using a Dose-Response Analysis Strategy to Measure Outcomes of Place-based Education Prepared by: Michael Duffin, PEER Associates, Inc. With support from: the Place-based Education Evaluation Collaborative (PEEC) For presentation at: the NAAEE Pre-conference Research Symposium October 10, 2006
For more details and related information, seewww.PEECworks.orgorwww.PromiseOfPlace.org(as of Dec. 2006)
The Place-based EducationEvaluation Collaborative (PEEC) • CO-SEED • Community Mapping Program • Litzsinger Road Ecology Center • Wellborn Ecology Fund • Trail To Every Classroom • Sustainable Schools Project • Forest For Every Classroom
Program “Response” (Measures of Intended Outcome) Lower Higher Less More Program “Dose” (Exposure + Implementation) PEEC’s Dose-Response Measurement Strategy • If participants with less dose report lower outcomes, and those with more dose report higher outcomes, then the program is likely to be an active ingredient • Coleman (1966) claimed that schooling accounted for only 10% of the variance in student achievement (or rR2 = .10) • Marzano (2003) claims that that number is actually closer to 20%, with 13% deriving from teacher-level factors, and 7% attributable to school-level factors • Wang (2002) found that weight status predicts 17-19% of cost for treating cardiovascular disease
Constructing Program “Dose” • Survey items estimating # of times participated in each part of the program delivery • Each item multiplied by staff estimate for average duration of typical event resulted in rough estimate of total hours • Converted program exposure to 1 to 4 scale and averaged with responses to items for implementation level to get dose composite • Very blunt instrument, but does seem to capture variability in dose (for the regression calculation, ordinal sequence was more important than gross accuracy of construct)
Measuring Program “Response” or Outcomes • Survey items reflecting behavior changes represented in program logic • Modules reflect major ideas • Indices represent a range of sub-dimensions of major ideas • Two to six items for each index • Used Bonferoni correction to be conservative about statistical significance • Indices, modules, and overall modules rationally constructed (factor analysis and systematic internal reliability tests still to be done)
PEEC Cross-Program Survey Results 2003-2004Changes in Educator Practice • 342 educator surveys • Very diverse sample (4 programs in 55 schools; Whole school change & Prof. development models; Urban, rural, suburban; Grades K-12) • Averages from an aggregate of 12 survey items show PEEC dose accounts for 19% of variance in Overall Educator Practice • Also at ΔR2≥ .10: Educator Engagement/ Growth; Use of Local Places for Teaching; Student Engagement in Learning; Student Civic Engagement; Student Time Spent Outdoors, Student Stewardship Behavior; Community Civic Engagement; Community planning/decision making process • Approx 200 more educator survey responses to be added in 04-06 (2 more programs, approx. 20 more schools)
Survey Analysis Provides Finer-grained View of Cultural “Tipping” at Haley Elementary, 2006 • Pre-post measures spanning 3 years show large, significant effects • Dose-response calculation for aggregate of the pre- and post responses shows CO-SEED dose accounts for 45% of variance in Overall Educator Practice Change • Dose-response calculation for post-only responses shows CO-SEED dose accounts for 12% of variance in same outcome • Comparing zero dose only: in 2003 Mean=1.9, n=20, SD=.51; by 2006 Mean =2.8, n=4, SD=.60. • “Tipped-but-not-completely” still need some mechanism to sustain
Reflections on UsingDose-Response Measurement Strategy • Single measurement event (without giving up possibility for pre-post comparisons, either of independent groups or matched pairs) • Conducive to both aggregation and disaggregation (across time and across programs) • Effect size is readily apparent (ΔR2 or percent variance) • Requires another level of understanding of statistics for users of the evaluation data • Be open to site-specific construction of dose
First Grade Academic Achievementas a Function of CO-SEED/ Community-Based Units(Young Achievers School, 2005) Design: • Principal says “One thing we know is that kids’ writing is much more interesting, complex, and detailed if they’ve had rich experience…The current first grade has about a third of the kids who didn’t have Kindergarten here and in general it is breathtaking the difference in the academic achievement. Our Kindergarten has the strongest place-based education in the school, especially with language development.” First grade is also strong. • 3 measures (Direct Reading Assessment, TERC Math, YA Writing Assessment) tracked in YA’s assessment database • Compared 1st graders with one v. two years of exposure to strong PBE teachers Young Achievers School, Jamaica Plain, MA
First Grade Academic Achievementas a Function of CO-SEED/ Community-Based Units(Young Achievers School, 2005) Findings: • 1st graders w/ more place-based education outperformed peers on all measures Young Achievers School, Jamaica Plain, MA
Effects of CO-SEED onStandardized Test Scores (MCAS) at theBeebe Health & Environmental Magnet School(Beebe School, 2005, Massachusetts) • CO-SEED worked with Beebe 1999-2003, helped secure CSR funding to continue work 2002-2005 • Several lines of evidence suggest that the environmental theme has become embedded in the school culture • Before analyzing MCAS scores, we predicted that Beebe would deviate from the typical pattern and increase performance relative to district and/or state in the following content areas: • Math (mostly near 3rd and 4th grade) • English Language Arts – Writing • Life Science • Earth Science Design: Beebe School, Malden, MA
Effects of CO-SEED onStandardized Test Scores (MCAS) at theBeebe Health & Environmental Magnet School(Beebe School, 2005, Massachusetts) • Typical pattern: State performs highest, then Beebe, then district Findings: Beebe School, Malden, MA
Effects of CO-SEED onStandardized Test Scores (MCAS) at theBeebe Health & Environmental Magnet School(Beebe School, 2005, Massachusetts) • Only a few deviations from the typical pattern (6th & 8th grade Math, 8th grade Life & Earth science) Findings: Beebe School, Malden, MA
Effects of CO-SEED onStandardized Test Scores (MCAS) at theBeebe Health & Environmental Magnet School(Beebe School, 2005, Massachusetts) • Analysis mildly supported the prediction for two areas (Math & Earth Science) • Analysis strongly supported the prediction in one area (Life Science) • Analysis did not support the prediction in one area (Writing, the typical pattern persisted in both grades 4 and 7) • Future prediction: Strongest results will continue to show up in the upper grades (i.e. where students have the highest cumulative dose of the environmental/ place-based theme integration) Findings: Beebe School, Malden, MA
References • Coleman, J. S., Campbell, E. Q., Hobson, C. J., McPartland, J., Mood, A. M., Weinfield, F. D., & York, R. L. (1966). Equality of educational opportunity. U.S. Government Printing Office, Washington, DC. • Duffin, M., Powers, A. L., Tremblay, G., & PEER Associates. (2004). Place-based Education Evaluation Collaborative: Report on cross-program research andother program evaluation activities, 2003-2004.Retrieved October 6, 2004 from http://www.peecworks.org/PEEC/PEEC_Reports/S0019440A. • Duffin, M., & PEER Associates. (2006). Portrait of an urban elementary school: Place-based education, school culture, and leadership; An evaluation of Project CO-SEED at the Dennis C. Haley Elementary School, 2003-2006. Retrieved October 6, 2006 from http://www.peecworks.org/PEEC/PEEC_Reports/S00FE7771-0100A9B6 • Marzano, R. J. (2003). What works in schools: Translating research into action. Association for Supervision and Curriculum Development, Alexandria, VA. • Wang, G., Zheng, Z., Heath, G., Macera, C., Pratt, M., & Buchner, D. (2002). Economic burden of cardiovascular disease associated with excess body weight in U.S. adults. American Journal of Preventive Medicine, 23 (1), 1-6.
Using a Dose-Response Analysis Strategy to Measure Outcomes of Place-based Education Prepared by: Michael Duffin, PEER Associates, Inc. With support from: the Place-based Education Evaluation Collaborative (PEEC) For presentation at: the NAAEE Pre-conference Research Symposium October 10, 2006 For more details, see www.PEECworks.orgorwww.PromiseOfPlace.org
Program “Response” (Measures of Intended Outcome) Lower Higher Less More Program “Dose” (Exposure + Implementation) Using a Dose-Response Analysis Strategy to Measure Outcomes of Place-based Education Summary notes from October 10, 2006 presentation at NAAEE pre-conference symposium by Michael Duffin, PEER Associates Duffin, M., Powers, A. L., Tremblay, G., & PEER Associates. (2004). Place-based Education Evaluation Collaborative: Report on cross-program research and other program evaluation activities, 2003-2004. Retrieved October 6, 2004 from http://www.peecworks.org/PEEC/PEEC_Reports/S0019440A. Duffin, M., & PEER Associates. (2006). Portrait of an urban elementary school: Place-based education, school culture, and leadership; An evaluation of Project CO-SEED at the Dennis C. Haley Elementary School, 2003-2006. Retrieved October 6, 2006 from http://www.peecworks.org/PEEC/PEEC_Reports/S00FE7771-0100A9B6 For more details and related information, see www.PEECworks.orgorwww.PromiseOfPlace.org