Extra temperature-digit persistence was uncertain in three of four US bird samples in 1985-2024

We tested whether five-unit temperature reports persist more when a bird survey route keeps the same observer. We used 61,064 consecutive-year pairs from accepted US Breeding Bird Survey runs in East and West samples in 1985-2004 and 2005-2024. Four linear probability models controlled temperature bins, date, state and year. Extra persistence estimates were +10.67, -0.27, +12.15 and +10.75 percentage points. Only the early eastern interval was above zero after correction for four tests. The other intervals included zero. The test did not show a clear common rise with observer retention. Observer changes were uncommon and not random. Stable instruments, weather, selection and data entry can affect the result. This is a report-quality audit, not an estimate of climate effects on birds or actual temperature error. The package has a selected panel, source hashes, all results, code, numerical checks and a figure.

observation bias · bird surveys · weather · temperature · digit preference · observer retention · null result