Part 11 · How robust is your finding?
Module 26 · How do you assess robustness?
Would the finding survive a small nudge?
- Explain what robustness means for a CNA finding.
- Run frscored_cna() across a range of thresholds.
- Judge whether a finding is steady enough to brief.
A result you plan to act on should not fall apart the moment the data or the thresholds wobble a little. Robustness asks exactly that. The CNA ecosystem has a dedicated companion package for it called frscore, which scores how "fit-robust" your models are across a whole range of consistency and coverage thresholds, so you don't have to do it by hand and lose an afternoon.
# install once install.packages("frscore") library(frscore) # re-analyse across many thresholds and score robustness in one call fr <- frscored_cna(immunization, fit.range = c(1, 0.7), granularity = 0.1) fr
Higher fit-robustness means a steadier finding
frscored_cna() runs CNA again and again across a grid of thresholds, here from 1.0 down to 0.7 in steps of 0.1, and rewards the models that keep showing up. A pattern that reappears no matter how you turn the dials is one you can trust enough to put in front of a director without sweating.
Think about confidence.
- Would your finding survive a small change in the data or thresholds?
- How robust does a result need to be before you'd put it in front of a director?