Uses and Limitations of Growth Data Jonathan Dings Boulder Valley School District June 22, 2009 jonathan.dings@bvsd.org

Overview • Setting Meaningful Goals for Continuous Improvement • Metrics that Incentivize Productive Responses • Contextualizing Growth Data to Provoke Thoughtful Use • Avoiding Over-Interpretation

Continuous Improvement Model Plan of Action Assess Current Reality Vision Set SMART Goals Why Are We Getting These Results? • Achievement • Equity • Climate Theory of Action If ….Then….and X will result BayCES, 2002

Assessing Our Results ■ What are the trends over time? ■ What are the trends for groups? ■ How did overall achievement change, compared to the same students’ achievement last year? ■ How does the change compare to a standards- based or norms-based idea of a year’s growth? ■ How does the change differ by groups? ■ How do achievement data compare with school climate, parent, staff and NCG survey results?

Colorado Student Assessment Program •Grades 3-10 Read/Write/Math; Sci at 5,8,10 •Based on Content Standards •Standardized Administration to Nearly All Students, with Accommodations Available •Selected- and Open-Response Questions •CSAP Development Process and Statistical Evidence Support Making Inferences from Results •Growth Data More Relevant for Many Purposes than Achievement

A Relatively New Metric: Colorado’s Growth Model ■ Easier to use and understand than the previous methods ■ Supported by a useful video ■ CDE graphs available online by district and school ■ Current year results arrived at a later than ideal time for full use

Colorado Median Growth Percentile ■ Amount of gain is statistically compared across students starting from the same point (“academic peers”) ■ Growth percentile assigned: percent of “academic peers” making that amount of gain or less ■ Percentiles range from 1 to 99, can be compared with state median of 50 ■ Accountability Measure: the middle growth %-ile

Several View of BVSD Writing Data Colorado’s Growth Model ■ District ■ Grade ■ Race/Ethnicity

District Median Growth %-ile Overall Eng Home La ELL Meal Assist Special Ed

Productive Response to Median Growth Percentile ■ Users are far less dismissive of gaps in growth than in achievement ■ Because a student’s growth is not readily predicted, raising median growth requires group improvement ■ Reflection on what went well for students with high growth

School and Teacher Use ■ Fair comparison across content areas ■ Encouraged for use in school goal setting, especially around equity ■ Published at student level in Student Information System (Infinite Campus), contextualized with grades, demographics, simple gain

School Results, vs. District

Student Roster in IC

Teacher level results

Avoiding Over-Interpretation ■ Contextualize with Error Estimates ■ Examine Multiple Content Areas and Years of Results for Each Student ■ Consider Likely Response of Parents, Having Seen Percentiles only for Achievement, not Growth

About this copy

Uses and Limitations of Growth Data — CCSSO, June 2009.

This is an image-based copy of the original PowerPoint slides; builds and animations are flattened.