Recorded Western Bluebird success changed in opposite directions for all and repeated nest boxes, 2007–2011 to 2021–2025

We tested whether changes in the recorded nest-box sample affected a nesting-success comparison in NestWatch version 6. We used attempts with valid positive clutch counts in the United States and Canada. We compared 2007–2011 with 2021–2025, accepted valid fledgling counts, and allowed each missing count to be a success or a failure. In an exploratory check of five species, Western Bluebird showed opposite directions for two samples. With equal weight per recorded location, the all-box change was between −7.29 and −2.64 percentage points. At 277 locations recorded in both periods, it was between +2.57 and +10.18 points. These are sharp bounds for the selected records, not confidence intervals. One early source account supplied 11,158 attempts and no late records. Without that source, the all-box change allowed both signs, from −1.31 to +4.80 points. Moving the early period to 2012–2016 also removed the strict sign from the repeated-box result. The five-species mean allowed both directions. Thus source coverage can reverse a recorded success comparison. The data do not establish a wild-population trend. The package contains original code, aggregate results, source hashes, figures, exact arithmetic checks, and instructions to retrieve the licensed source. A clean run matched all 14 numerical outputs.

biodiversity · nesting success · NestWatch · Western Bluebird · sample selection · missing outcomes · open data · reproducibility