On the reliability of N-mixture models for count data

RT @n_a_gilbert: Barker et al. (2018): On the reliability of N‐mixture models for count data "Our message is that fundamentally count data…

RT @n_a_gilbert: Barker et al. (2018): On the reliability of N‐mixture models for count data "Our message is that fundamentally count data…

Barker et al. (2018): On the reliability of N‐mixture models for count data "Our message is that fundamentally count data under imperfect detection can only be reliably used as indices." #365papers https://t.co/bvNP9glb96 https://t.co/qulUaqxAxg

As a result: 1) we don't know how models respond to trends in detectability; 2) inference regarding necessary effort could be optimistic; and 3) some may conclude uncorrected counts models and detectability models like N-mixture are equivalent e.g. https:/