The Adventures of Stat-Woman
Smart statistical decisions across the combination product battlefield
Non-normal data. Borrowed sample sizes. An outlier nobody wants to explain. Statistics shows up at every stage of a combination product program, and there's rarely a Stat-Signal telling you which tool to reach for.
Follow Stat-Woman and her trusted advisor, Mr. Ledger, through seven patrols of Variance City, one for each stage of the product lifecycle. Along the way, they take on The Rogue Distribution: the everyday statistical traps that quietly turn solid programs into regulatory questions.
Seven parts. About 75 minutes of reading. One full product lifecycle.
Why suit up for this series
Match the tool to the phase
What works in a feasibility sandbox can sink you in Design Verification. See how the right method, confidence level, and acceptance criteria shift as your program matures.
Build a story that survives review
Reviewers don't ask whether a result is interesting. They ask whether it's justified. Learn how to make every statistical decision traceable before anyone else asks "why?"
Keep it defensible for the long haul
Launch isn't the finish line. Carry your development assumptions into post-market monitoring so an old issue in a new disguise doesn't catch your team off guard.
Choose your patrol
Each part stands on its own, but the story builds. New to Variance City? Start with Part 1.
Part 1: The origin story
Smart Statistical Decisions Across the Combination Product Battlefield
Stat-Woman opens her toolbelt for a first patrol of the entire lifecycle, from sandbox experiments to the final battle of validation. The lesson that sets up everything else: having every tool matters less than knowing when to use each one.
Villains: The whole Rogue Distribution makes an entrance
Part 2: Planning
The Stat-Cave: Why the Best Statistical Decisions Start Before the Data Does
Before a single unit is built, the questions you define decide whether your answers will hold up. A phase-by-phase look at planning, from feasibility through lifecycle monitoring, and why the right question beats the right test.
Villain: Captain Assumption
Part 3: Exploration
Lessons from the Training Grounds: Exploratory Statistics, Non-Normal Data, and Preparing for the Worst
When in doubt, graph it out. Picking the visual that answers your actual question, treating non-normal data as information rather than a problem, and untangling TMV, MSA, and GR&R once and for all.
Villain: The Confounder
Part 4: Design Verification
Into the Field: Building a Design Verification Strategy That Holds Up Under Fire
The safety net comes down. Choosing among confidence intervals, tolerance intervals, capability, and reliability testing, justifying sample sizes that aren't just inherited, and knowing when an outlier is noise versus a warning.
Villains: The Outlier, with Powerless close behind
Part 5: Validation and regulatory review
Proving It Twice: Statistical Rigor from Validation Through Regulatory Review
You proved it once. Now prove it to someone whose job is not to believe you. Validation across its three distinct fronts, what real pre-specification looks like, and one simple test for spotting a midstream rationalization.
Villain: Dr. Data himself
Part 6: Launch and lifecycle
Staying Sharp in Variance City: Statistical Continuity From Launch to Lifecycle
The team disperses and the protocols get archived, but the patrol continues. Why treating each phase as a statistical reset creates blind spots, and how development data becomes the lens for every post-market decision.
Villains: Miss-Interpretation and Powerless
Part 7: Season finale
The Return of Bias: The Villain Who Never Really Left
Mr. Ledger finally opens his notebook. Every battle so far gets a second look, and a pattern emerges that no single study could have revealed. Best read after the other six.
Villain: Bias, the one behind them all
Need a Mr. Ledger?
If your team is weighing a sample size, an acceptance criterion, or an outlier that won't go away, we're always glad to talk it through.
Talk with our team