A Letter from Craig Moncreiff
A personal introduction to the work and an invitation to help shape it.
Use your Planning Workspace goals as end-of-period targets. Follow 2,000 synthetic students as participation and progress change over time.
Run the selected plan alongside the same starting school with no new intervention expansion.
Existing supports continue in both paths. Students can improve, plateau, or lose progress.
Bands show random variation under fixed assumptions—not a confidence interval or a validated forecast. The baseline is the starting condition, not an intervention. A gold target line marks the deadline goal, not a forced trajectory.
Save the current result as A. Change the actions or behavior settings, run again, and compare A with the new result B.
Enter real follow-up counts when available. This records observations and calculates candidate response rates; it does not prove causation or automatically retrain the model.
No observations recorded. Live Planner / Homework Habit integration is not connected in this build.
| Active | Outcome | Current | District | County | State | Your Goal | Model Result | Status |
|---|
| Staff Resource | Modeled | Maximum | Status |
|---|
| Specific Action | Current Reach | Recommended Reach | Recommended Level | Details |
|---|
A personal introduction to the work and an invitation to help shape it.
How real school data, research, simulation, intervention logic, and future program learning fit together.
Decision Lab begins with two different kinds of evidence: official information about the school you actually lead, and national longitudinal research that helps describe how student conditions relate over time. The two principal public authorities are the California Department of Education and the U.S. Department of Education's National Center for Education Statistics. The simulation layer is kept separate from both.
The current research foundation includes the NCES Education Longitudinal Study of 2002 (ELS:2002), a nationally representative longitudinal high-school study used in model development and calibration. It helps Decision Lab represent relationships among student conditions that a school-level dashboard cannot show by itself.
School identity and available baseline and benchmark values are drawn from California Department of Education DataQuest and official downloadable public data files. The selected school's real public baseline is the starting point for the scenario.
The Dashboard is California's public accountability and continuous-improvement system. It provides the official reporting context for measures that include chronic absenteeism, College/Career Indicator, graduation, and suspension. Decision Lab uses CDE's official public data files as baseline inputs; it does not treat Dashboard reporting as a Decision Lab prediction.
The campus baseline is assembled from specific CDE public files—not from a generic statewide average.
Official data tells Decision Lab where your school is. Research helps describe the relationships underneath those outcomes. Decision Lab then models what could happen if you change goals, interventions, reach, overlap or staff constraints.
The two numbers answer two different questions. Year 1 shows the immediate annual value of the modeled attendance recovery plus the first cohort of reconnected learners. Steady state shows what the annual value can look like once two returning reconnection cohorts overlap.
Attendance Recovery + First-Year Reconnection Value.
Attendance Recovery + two overlapping reconnection cohorts.
The typical modeled reconnected learner returns around the beginning of junior year and remains enrolled for about two years. In Year 2, the prior year's returning cohort is still enrolled when the next cohort returns. Once that pattern repeats, the school carries approximately two overlapping reconnection cohorts in a normal year.