Evidence over intuition
Anchor choices in experiments, counterfactual logic, and model-based forecasts—not anecdotes.
Decision Science
How humans and algorithms make, justify, and optimize decisions under complexity.
See Core PrinciplesFrom judgment to measurable impact
PrimeStata integrates cognitive psychology, behavioral economics, and statistical modeling to improve decisions where stakes are high and data are messy. Work draws on research and applied practice across universities and national research institutions (Boston University, NYU, University of Akron) and industry settings (Google People Analytics, NCCER, and consulting).
Core decision principles
Calibrate, don’t guess
Use probabilistic thinking, confidence training, and post-decision review to reduce over/underconfidence.
Bias-aware processes
Structure decisions (checklists, blinded reviews, thresholds) to limit confirmation and selection bias.
Interpretability first
Favor models leaders can explain (GLMs, causal diagrams, uplift) before adding complexity.
Methods we use
Experiments & quasi-experiments
A/B and multivariate tests, difference-in-differences, synthetic controls, and randomized encouragement designs.
- Power & sample planning
- Pre-registration & analysis plans
- Heterogeneous treatment effects
Statistical modeling
Multiple & logistic regression, standardization, mediation/moderation, mixed models, survival and hazard analysis.
- Partitioning explained vs. residual variance
- Diagnostics: assumptions & robustness
- Cross-validation & error decomposition
Decision analytics
Expected value & risk, utility curves, scenario trees, and portfolio-style trade-off analyses for strategy.
- Cost-benefit thresholds
- Uplift models for targeting
- Post-decision audits & learning
Human + AI alignment
Designing workflows where model output improves—not replaces—judgment; trust calibration and transparency.
- Model cards & decision logs
- SHAP/ICE for explainability
- Guardrails & escalation paths
Where it’s applied
Leadership & talent
Selecting and developing leaders with structured interviews, validated rubrics, and adverse-impact checks.
Growth & product
Pricing, funnel experiments, and causal lifts tied to revenue, retention, and unit economics.
Operations & safety
Policy pilots and decision thresholds that balance risk, throughput, and wellbeing outcomes.
Operator-ready artifacts
Decision briefs
One-pager: question, design, signal, uncertainty, and recommendation with next actions.
Experiment playbooks
Templates for hypotheses, metrics, guardrails, and rollout criteria leaders can reuse.
Calibration reports
Brier/Log scores and reliability plots to track forecast skill and improve future calls.
Related flagship path
This research most directly supports Data Science work where experimentation, causal structure, and interpretable models need to move from theory into operational decisions.
Collaborate on Decision Science
Partner with PrimeStata to design credible tests, build interpretable models, and install decision processes that scale.