Within-collection tests found no clear excess of missing catch entries in flooded ECN beetle traps in England and Scotland, 1993-2015

We tested whether flooded pitfall traps had no catch entry more often than unflagged traps from the same collection. We used ECN beetle data in England and Scotland in two periods from 1993 to 2015. Only 667 of 5,879 selected collections had both flood states. Mean within-collection differences were -2.28, 0.00, -0.82 and -0.24 percentage points. Three four-test adjusted intervals included zero. The fourth sample had zero contrast in every collection, so its cluster interval was not estimable. Crude missing-entry rates were higher in flooded traps, but they target a different comparison. Explicit no-beetle codes were more common in flooded traps in a secondary check. A catch entry, including a no-beetle code, does not establish a valid catch. Selection of mixed collections, small year groups, local trap differences and reporting practice limit inference. This study tests observation bias. It does not estimate beetle loss, a causal rain effect, climate change or pollution. The package includes new code, source hashes, aggregate evidence and a raw rebuild recipe.

observation bias · carabid beetles · flooding · pitfall traps · missing records · null result