Compare abundance trends without assuming constant detection

Define the ecological target

An observed count is affected by abundance and by how likely individuals are to be detected. Observer skill, vegetation, weather, activity and survey effort can change that probability. A repeated count protocol is useful, but unchanged written instructions do not prove unchanged detection.

Specify the observation model

In the simplified model, each individual is detected with the same supplied probability p, with no duplicate counts or false positives. Dividing count by p provides a point adjustment for missed individuals. The detection probability is provided externally, not estimated from the count; the exercise therefore isolates the logic of a correction rather than implementing a full abundance model.

Work through the sampling logic

The reference count is 40 with supplied detection 0.5, corresponding to adjusted abundance 80. A current count of 40 at detection 0.25 corresponds to 160. Equal observed counts therefore need not imply equal abundance. Conversely, observed counts can fall while adjusted abundance remains constant if detection also falls.

Challenge the inference

Uncertainty in p can dominate an adjusted estimate, especially at low detection. Heterogeneous detection, changing spatial coverage and count saturation violate the common-p model. Repeated visits, distance measurements, capture histories or validated observation methods can provide information about detection, but each method carries its own assumptions.

Design a defensible survey

A monitoring report should show raw counts, effort, detection evidence and adjusted estimates separately. A correction must not be treated as an automatic improvement merely because it produces a larger number. If detection is unknown, report the range of interpretations supported by plausible scenarios instead of claiming a precise population trend.

Worked original packet

Reference adjusted abundance is 40/0.5=80. Current count 40 at p=0.25 gives 160, an adjusted ratio 2 even though the raw count ratio is 1. At current count 20 and p=0.25, adjusted abundance is 80, matching the reference despite a 50% raw-count decline.

Investigation

Hold the current count 40 and vary p from 0.25 to 1. Then compare count 20 at p 0.25 with the reference. State which detection information is assumed rather than measured.

Changed-case transfer

Current count 60 has independently supplied p=0.75. Compare raw and adjusted ratios to the reference. What additional uncertainty should a real report retain?

Feedback

Adjusted abundance 80; adjusted ratio 1 versus raw count ratio 1.5. Actual estimation uncertainty in p, count variability and model assumptions must be retained.

Sources and evidence boundary

Original instructional prose, values and exercises. Context checked 30 October 2026: primary agency source. All observations are synthetic. No organism collection, trapping, tagging, release or unsafe-site access is requested; a real study requires qualified design, permission and supervision. Public feedback is practice, not independently verified mastery.