A defined protocol. Published either way.
Funders avoid this field because most work in it cannot say what would prove it wrong. Every StrayPaw study is falsifiable, measured against established methods, and published either way.
Seven stages, in this order.
The order matters more than any individual stage. A hypothesis written after the data is not a hypothesis, and a baseline taken after the intervention is not a baseline, both are the ordinary way this field produces numbers that cannot be relied on.
A specific, answerable question tied to a named geography, not a theme. Registered before fieldwork begins, with the geography, timeframe and executing partner fixed.
A falsifiable statement with a direction and a magnitude, plus the null it is tested against. If no result could disprove it, it is not a hypothesis and does not proceed.
Sampling frame, unit of analysis, sample size and the reference method it will be compared against. Chosen before data collection, and published with the protocol.
The pre-intervention measurement. Without it there is no denominator, and coverage, the number that decides whether anything worked, cannot be computed at all.
Delivered by a named executing organisation against a defined protocol, with every action recorded to the individual animal rather than to a daily total.
Re-measurement using the same method as the baseline, at a stated interval, reported against indicators defined in advance.
Results released whatever they show, including the protocol, the analysis and the limitations. A null result is published on the same terms as a positive one.
Established methods, not invented ones.
Nothing in the StrayPaw method set is novel, and that is deliberate. Free-roaming dog population estimation has a literature; a reviewer should be able to check our approach against it rather than take our word for anything.
Mark–resight along fixed transects, the standard approach for estimating free-roaming dog populations. Individual identification comes from coat markings and, where animals are chipped, from the code itself.
Repeated counts along the same routes at the same time of day, giving a comparable index over time even where an absolute population estimate is not affordable.
Sterilisation and vaccination status resolved to the animal rather than aggregated to a programme total, so coverage is counted rather than estimated.
Community-reported data compared against a gold-standard count on the same ground, reporting agreement and error rather than asserting the cheaper method works.
Who approves it, and who checks it.
Studies involving animal handling run under the ethics approval of the executing partner or an academic collaborator, to the standards their institution requires. StrayPaw does not self-certify this.
Every study has a named principal investigator accountable for the design and the analysis, and a named executing organisation accountable for the fieldwork.
Animal records carry no personal data about reporters beyond what is needed to contact them. Precise locations are restricted, because publishing exact positions of street animals creates a welfare risk.
The funder, the executing partner and any commercial relationship are stated on the study record. StrayPaw's own interest in the result is part of that disclosure.
Null results, published the same way.
A funder’s real risk in this sector is not that the work fails. It is that failure goes unreported, so the next programme repeats it. Publishing a null result is the cheapest thing that breaks that cycle.
Design, hypothesis and analysis plan fixed in advance and public.
A null result is released on the same terms as a positive one.
What the study cannot show is part of the study.
Aggregated openly; individual records available for verification.