I wrote a series entitled "Why is a Statistician Invited to the Room Late?" This is really saying why is collaboration between scientists and statisticians sub-optimal. I wrote and posted 9 reasons. Here I add reasons why this collaboration matters along with links to the 9 articles that address why this collaboration is sub-optimal. This sub-optimality can lead to a host of issues: from poor decisions to information that is collected but not fully utilized, which ultimately impacts real patients. A problem is fixed by recognizing the problem and then understanding the reasons why the problem exists. Now it is time for each of us (scientists and quantitative scientists) to reflect and move this collaboration towards optimality.

This post shows, both intuitively and mathematically, why adjusting for baseline using ANCOVA is always more efficient than analyzing change from baseline.

A reflection on how drug development and applied statistics have transformed unrealized human potential into longer, fuller lives.

Investment decisions in drug development require different statistical frameworks than regulatory submissions—applying regulatory standards (multiplicity control, single pre-specified estimands) to staged investment decisions cedes control to the most risk-averse stakeholder and kills promising drugs.

Rigorous empirical research using 100 clinical datasets proved we could save 20% on trial costs, but the statistical community dismissed it as irrelevant—here's why 'statistics on statistics' remains valid.

The most effective means of building an exceptional team is through strategic relationship-based hiring rather than reactive resume reviews and references. It uses the field of statistics as an example how.

This essay shows the roles to engage to combat this never-ending concern.

Pharmaceutical statistical organizations have become so focused on regulatory process that they've lost the ability to actually analyze data, and it's setting up failures no one sees coming.

Imagine Billie Bob Thornton portraying a biotech CEO—now imagine learning about drug development risk along the way.

This essay argues that introductory statistics should prioritize numerical literacy and interpretation over mathematical and computational procedures.

Read with last week’s post.

When we have time to think, mathematics can clarify choices that instinct alone cannot resolve.

The history of creatinine clearance shows that scientific progress comes less from refining formulas than from improving the biological quantities they are meant to estimate. A model is useful for conveniently estimating the underlying biology.

Its luck was revealed in childhood; its power lies in knowing where unequal allocation helps, and where it cannot.

This blog shows how to reason quantitatively about FDA’s proposed policy change instead of defaulting to either alarm or celebration.

From crime fiction to sufficient statistics, why every layer of modeling necessarily blurs information rather than creating it.

What a few college football games reveal about weighing evidence, applying models, and making informed decisions — lessons that extend far beyond the field.

Statistics did not disappear; one of its creations has become powerful, and with that power comes responsibility we are currently failing to take seriously.

A guide to avoiding wasted potential in clinical trials by leveraging repeated measures and integrating pharmacology and biological insights into analysis.

Thanksgiving stories lead to a broader lesson: knowing all possibilities allows one to apply six critical levers for smarter study design.