Data-Driven Dialogue Predicting, Exploring, and Explaining Data Hygiene Elementary School Data-Driven Dialogue Predicting, Exploring, and Explaining Data September 6, 2012

Outcomes & Process What predictions do we have about our data? What observations and trends appear from the data? What are our top priorities from the data? What are our root causes?

Why This Process? Education Accountability Act of 2009 (S.B. 09-163) requires all schools and districts to submit Unified Improvement Plans (UIP) to CDE to be posted at www.schoolview.org Unified Improvement Planning is a collaborative process where all staff have input into the school plan

Data-Driven Dialogue Steps Outcomes & Process Data-Driven Dialogue Steps Step 1 – Predict (Activate & Engage) Step 2 – Explore (Explore & Discover) Step 3 – Explain (Organize & Integrate) Step 4 – Take Action

Outcomes & Process Data Team Roles RECORDER MATERIALS MANAGER Check with each team member before recording MATERIALS MANAGER Organize data & charts for viewing, recording PROCESS CHECKER Monitor for balanced participation ENVRIONMENTAL ENGINEER Organize physical arrangement for all to view

Step 1: Predict The purpose: To activate interest and bring out our prior knowledge, preconceptions, and assumptions regarding the data with which we are about to work. Prediction allows dialogue participants to share the frame of reference through which they view the world and lays the foundation for collaborative inquiry. The steps include: Clarify the questions that can be answered by the data Make predictions about data Identify assumptions behind each prediction   Prediction Sentence Starters: I predict . . . I expect to see . . . I anticipate . . . Assumption Questions: Why did I make that prediction? What is the thinking behind my prediction? What do I know that leads me to make that prediction? What experiences do I have that are consistent with my prediction? Quantify observations – describe with numeric measures or attributes

Step 1: Predict ~5 minutes for Predict Data Sources: TCAP Reading, Writing and Math: Executive Summary – % P/A each grade (overall & subgroups) Instructional Summary – sub-contents (ex. Vocab, fiction) Scores compared to all schools in the District and also to Colorado Growth Percentiles – overall and for subgroups

Step 1:Predict Predictions Assumptions Quantify observations – describe with numeric measures or attributes

Step 2: Explore The purpose: Generate priority observations or fact statements about the data that reflect the best thinking of the group. The steps include: Interact with the data (highlighting, creating graphical representations, reorganizing) Look for patterns, trends, things that pop out Brainstorm a list of facts (observations) Prioritize observations Turn observations into priority performance challenges Avoid: Statements that use the word “because” or that attempt to identify the causes of data trends. Sentence starters: It appears . . . I see that . . . It seems . . . The data shows . . . Quantify observations – describe with numeric measures or attributes

Step 2: Explore ~25 minutes for Explore Data sources (one at a time): TCAP Reading, Writing and Math: Executive Summary – % P/A each grade (overall & subgroups) Instructional Summary – sub-contents (ex. Vocab, fiction) Scores compared to all schools in the District and also to Colorado Growth Percentiles – overall and for subgroups

Step 2: Explore The purpose: Generate priority observations or fact statements about the data that reflect the best thinking of the group. The steps include: Interact with the data (highlighting, creating graphical representations, reorganizing) Look for patterns, trends, things that pop out Brainstorm a list of facts (observations) Prioritize observations Turn observations into priority performance challenges Avoid: Statements that use the word “because” or that attempt to identify the causes of data trends. Sentence starters: It appears . . . I see that . . . It seems . . . The data shows . . . Identify no more than 2 priority needs, and write each one on a separate poster

Share Group Findings Share the trends you saw in your data Share the Priority Performance Challenges Summary of the data where there are challenges For example: Persistent low performance among English Language Learners in reading across all standards and grades. For the past three years, English Language Learners have had median growth percentiles below 30 in all content areas, substantially below the minimum state expectation of 55. Identify no more than 2 priority needs, and write each one on a separate poster

Step 3: Explain The Purpose: Generate theories of causation, keeping multiple voices in the dialogue. Deepen thinking to get to the best explanations and identify additional data to use to validate the best theories. The steps include: Generate questions about observations Brainstorm explanations Categorize/classify brainstormed explanations Narrow (based on criteria) Prioritize Get to root causes Validate with other data Guiding Questions: What explains our observations about out data? What might have caused the patterns we see in the data? Is this our best thinking? How can we narrow our explanations? What additional data sources will we explore to validate our explanation? 2 posters – one for each priority need statement. Generate explanations & possible causes, categorize, and search for root cause. The “5 Whys” protocol is used to drill down and identify the true root cause.

The school should have control over the Root Causes A cause is a “root cause” if: The problem would not have occurred if the cause had not been present The problem will not reoccur if the cause is dissolved Correction of the cause will not lead to the same or similar problems The school should have control over the root cause 

Root Cause Examples Non-Examples Why Non-Examples? Student attributes (poverty level) Parent education & involvement Student motivation Why Non-Examples? Schools do not have control over these causes

Root Cause Examples The school does not provide additional support/interventions for students performing at the unsatisfactory level Lack of clear expectations for tier 1 instruction in math. Lack of intervention tools and strategies for math. Limited English language development. Inconsistency in instruction in the area of language development. Low expectations for all subgroups. Low expectations for IEP students.

The “5 Whys” Protocol (Explanation) Getting to Root Causes The “5 Whys” Protocol (Explanation) Proposed Cause:____________ Why? 4. Why? • Because…. • Because…. 2. Why? 5. Why? Why? • Because….

5 Why Example ELL students are not engaged in learning in the core content classes. Why? Because… Core curriculum is not accessible to ELL students. ELL students’ English skills are not proficient enough to participate in discussions, ask questions, and comprehend core content. There is inconsistent English language support for students in core content classes. Lack of implementation of INSIDE and EDGE ELL curriculum as parallel support for ELL students in core content classes.

Step 3: Explain 5 minutes for explanations 15 minutes for root cause identification Each group will chart explanations and identified root causes

Step 3: Explain The Purpose: Generate theories of causation, keeping multiple voices in the dialogue. Deepen thinking to get to the best explanations and identify additional data to use to validate the best theories. The steps include: Generate questions about observations Brainstorm explanations Categorize/classify brainstormed explanations Narrow (based on criteria) Prioritize Get to root causes Validate with other data Guiding Questions: What explains our observations about out data? What might have caused the patterns we see in the data? Is this our best thinking? How can we narrow our explanations? What additional data sources will we explore to validate our explanation?

Share Group Work Trends Explanations Root causes

Next Steps Refine root causes Validate root causes (does our other data support the root cause) Action Planning