Fighting the Database & Winning: Cohort Data & Disaggregation Best Practices Christopher M. Johnson, PhD, Renton Technical College Michelle Andreas, EdD, Washington State Board of Community & Technical Colleges
Overview Discuss and survey issues with getting relevant data to analyze AtD strategies – Similarities and differences between US States Show you our solution – Take it for a test drive How we got there and why! – Share best practices 2/4/2010Chris Johnson & Michelle Andreas2
Quick Questions How many of you are institutional researchers? How many of you are satisfied with your current data system? How many of you spend enough quality time doing useful analysis? Instead, how many of you spend more time fighting the database than getting useful data out of it? 3Chris Johnson & Michelle Andreas2/4/2010
Issues with Your Data Please mix with others from other states and other colleges as much as possible With your neighbors, please brainstorm: – Personnel Issues – Technical Issues – Design Issues – Communication Issues 4Chris Johnson & Michelle Andreas2/4/2010
Detailed Issues – The Survey Please fill out the detailed survey sheet – please add to web survey now or later Feel free to talk about it with your neighbors, but please fill out according to your college and state specifically Questions for everyone – Are there any strong disagreements from the survey – not an issue in your state or with your college? – Are there any agreements from the survey – big issues in your state or with your college? 5Chris Johnson & Michelle Andreas2/4/2010
The Positives in Washington All 34 colleges use the same system – A lot of people know a lot of things – A puzzle, but someone often has a piece or two The State Board puts out an easy-to-use database that answers a lot of questions – Which is good if both the college and the State have the same question – Not as good if your question is different – Unfortunately, tracking cohorts (in general) was not a State question The Washington Student Achievement Model – Milestones between start and finish – Success tracking – Integrated into the State databases 6Chris Johnson & Michelle Andreas2/4/2010
So what did we do at Renton Tech? The “Madrona” System Features – Pre-designed Excel Pivot Tables and Charts – Deployed on a Campus Network – Currently, 25% of the campus is connected (includes all full time faculty, exec cabinet, admin) – Tech lite 7Chris Johnson & Michelle Andreas2/4/2010
Why Excel Pivot Tables? Easy to set up – easy to filter Easy to modify and adapt Allow easy drill down or disaggregation Everyone has Excel, so you don’t need to run around and install software Most people use Excel Pivot tables are intuitive Try it out for yourself! – – Interactive Cohort Retention Report (in MS Excel 2007) – or ask Chris to copy the file to your computer) Interactive Cohort Retention Report 8Chris Johnson & Michelle Andreas2/4/2010
Solving the Problem – Communication & Resources Integrity Task Force – our strategy – Find out who knows what on and off campus Look at every resource available – Data dictionary, any available diagrams Map it out yourself Establish good relationships with IT personnel on & off campus 9Chris Johnson & Michelle Andreas2/4/2010
Alignment with State Resources Chris Johnson & Michelle Andreas2/4/2010
Solving the Problem – Systematic Extraction, Query, Report Phases Develop Extraction & Report Phases First Develop Summary Fields and Preserve Detail Fields (for drill down) Use lookup tables to compensate for quirky data design (or lack of any design) Get the data out to as many as possible as fast as possible (file server & network access) 11Chris Johnson & Michelle Andreas2/4/2010
Put out the reports Run reports past the Data Team for feedback Get the reports out - fast. Get feedback from campus. Take two Alka Seltzer and grow a thick skin Hit back on the pot shots – If they say it’s wrong, make them prove it! Establish an open door policy for anyone to challenge (with source documents) Clean up the mistakes – design & run integrity tests Use legacy system for real time reporting to give you time to develop Get everything else out on the data warehouse 12Chris Johnson & Michelle Andreas2/4/2010
Provide Training & Support Open door and quick response to faculty and staff Negotiate important research questions from “fishing trips” Refresh and Review Once a quarter department meetings Faculty data groups in addition to Data Team Creates a feedback loop 13Chris Johnson & Michelle Andreas2/4/2010
Don’t do this alone! Partner with other colleges in your state! Partner with your college system’s governing body! Give away your data system to your partners (Code sharing & joint development) Don’t fall in love with your system, your technology, and/or your methods! – Someone else does some things better – You do some things better than someone else Don’t design a big system with the goal to flip the switch when it’s done! (e.g., dashboards, SharePoint) Instead, do Rapid Prototyping – Madrona is rapid prototyping, nothing more 14Chris Johnson & Michelle Andreas2/4/2010
Remember No one cares one iota about your queries or your data warehouse or your cool technology They only care about the reports and information They can’t wait two or more years till your system is done! By getting everyone involved ASAP, it’s everyone’s system and not your baby! 15Chris Johnson & Michelle Andreas2/4/2010
The Future & the Buzz in Washington 34 colleges in Washington State have the same challenge (same legacy system) Partnerships – Washington State Board of Community & Technical Colleges – 5 Colleges: Data Mart Project Washington Association of Research Practitioners (WARP) – Changing from a code dictation, top down entity to a knowledge sharing body – Collaboration site being set up – knowledge sharing 16Chris Johnson & Michelle Andreas2/4/2010
Thank you! Any questions? Detailed Technical Questions at Institutional Research and Effectiveness Focus Area Networking Session – 3:15 pm today Chris Johnson & Michelle Andreas172/4/2010