The people who count things: volunteer surveys and what they are worth

This works through volunteer recording in the order the parts actually depend on each other.

The datasets have no professional equivalent

National breeding bird surveys, butterfly transects and plant atlases run on volunteer effort measured in hundreds of thousands of hours annually. No research budget anywhere could fund the same spatial and temporal coverage, which is why these schemes underpin national indicators. Policy targets on biodiversity in several countries are measured directly against volunteer-collected data.

Their real value is length rather than size, because a trend only separates from year-to-year noise after a couple of decades, which is why a scheme started this year cannot substitute for one that has run since the 1970s however generously it is funded.

Method discipline is what makes it science

A butterfly transect works because the same route is walked in the same way under specified weather conditions, year after year. That consistency is what allows a change in count to be interpreted as a change in population rather than a change in observation. Schemes therefore spend more effort on protocol training than on species identification, which surprises most new recorders.

The detail that matters: handovers are the weak point, since a new recorder taking on a long-running transect can introduce a step in the data that looks exactly like a population change, which is why the better schemes overlap the outgoing and incoming recorder for a season.

Effort has to be modelled explicitly

If more people record in a warm spring, counts rise without any change in the underlying population. Modern analyses correct for recorder effort, detection probability and spatial bias, which is why raw totals are rarely published.

In practice, opportunistic records without effort information are still useful for distribution but much weaker for trends. The most under-supplied data of all is the zero, because a completed visit that recorded nothing is exactly as informative as one that recorded fifty, and almost nobody submits a blank form since it does not feel like a result.

Where volunteer data is weakest

Coverage is biased toward accessible, populated and attractive places, and toward species that are easy to identify. Nocturnal, marine, subterranean and taxonomically difficult groups are consistently under-recorded. This is not a criticism of the schemes so much as a description of where professional survey money should be going.

The correction most schemes use is to allocate squares rather than let recorders choose them, which sends people to dull, distant and inconvenient places and is the reason national surveys can say something about a country rather than about its car parks.

Starting is easier than it looks

Most schemes are designed for beginners, with a single square or transect and a short annual commitment. Identification support through local recording groups is generally generous, and being wrong is expected and corrected.

The consistent finding from participants is that it changes how you look at a place you already knew well. The commitment is smaller than people assume and more rigid than they expect, since a breeding bird square is commonly two early-morning visits a year that have to fall inside specified windows and conditions, and it is that inflexibility rather than the identification that defeats new recorders.

A record only becomes data once somebody verifies it

Most national schemes route submissions to a county or taxon-specific verifier, who checks the identification against the date, the location and the plausibility of that species being there at all. Records carrying a photograph, a grid reference of stated precision and a named recorder survive that process; a bare species name dropped on a map frequently does not.

On the ground, unverified records are usually retained but flagged, and analysts exclude them, so a large share of casual app submissions never reaches the trend they were meant to inform. Attaching a photograph and a line on how you identified it costs seconds and is the difference between a sighting and a usable observation.

The takeaway

The graph in the government biodiversity report was drawn by people walking the same line every summer for thirty years.

Go slowly enough to notice, and most of this stops being advice and starts being obvious.