For centuries, much scientific observation was carried out by amateurs: clergy recording the weather, farmers noting the first swallows of spring. Professional science later pushed such contributions to the margins. Today they are returning under the name citizen science, in which members of the public collect or analyse data for research projects, from counting birds to classifying galaxies in telescope images.
The advantages are scale and reach. Volunteers can gather observations across whole continents and over many years, something no research team could afford. Online projects allow thousands of people to examine images that would take a single scientist a lifetime. The disadvantage most often raised is quality: untrained volunteers may misidentify species or record data inconsistently.
Projects have developed ways to manage this. Simple tasks, clear instructions and the use of several independent volunteers for the same item allow errors to be detected and averaged out. Studies comparing volunteer and expert data have often found them close enough to be useful, especially when the design accounts for known weaknesses. Citizen science works best, in other words, when it is designed around what volunteers can do reliably.