The diversity and evolution of ecological and environmental citizen science

Abstract

Citizen science-the involvement of volunteers in data collection, analysis and interpretation-simultaneously supports research and public engagement with science, and its profile is rapidly rising. Citizen science represents a diverse range of approaches, but until now this diversity has not been quantitatively explored. We conducted a systematic internet search and discovered 509 environmental and ecological citizen science projects. We scored each project for 32 attributes based on publicly obtainable information and used multiple factor analysis to summarise this variation to assess citizen science approaches. We found that projects varied according to their methodological approach from 'mass participation' (e.g. easy participation by anyone anywhere) to 'systematic monitoring' (e.g. trained volunteers repeatedly sampling at specific locations). They also varied in complexity from approaches that are 'simple' to those that are 'elaborate' (e.g. provide lots of support to gather rich, detailed datasets). There was a separate cluster of entirely computer-based projects but, in general, we found that the range of citizen science projects in ecology and the environment showed continuous variation and cannot be neatly categorised into distinct types of activity. While the diversity of projects begun in each time period (pre 1990, 1990-99, 2000-09 and 2010-13) has not increased, we found that projects tended to have become increasingly different from each other as time progressed (possibly due to changing opportunities, including technological innovation). Most projects were still active so consequently we found that the overall diversity of active projects (available for participation) increased as time progressed. Overall, understanding the landscape of citizen science in ecology and the environment (and its change over time) is valuable because it informs the comparative evaluation of the 'success' of different citizen science approaches. Comparative evaluation provides an evidence-base to inform the future development of citizen science activities.

Fig 1. The rate of increase in the cumulative number of ecological and environmental citizen science projects as revealed by a systematic search.

Black vertical lines indicate the estimated years when the rate of increase changed, as indicated by segmented regression, with 95% confidence intervals around these estimates indicated by grey rectangles (see Table B in S1 File).

Fig 2. The landscape of citizen science described by a Multi-Factor Analysis (MFA) of 32 attributes of 509 citizen science projects in environmental and ecological science.

The majority of projects (coloured points) are best described according to their position on the first two multivariate axes (the methodological approach: x axis, and the complexity of the approach: y axis). There is a separate cluster of computer-based projects (light grey points, labelled ‘v’) explained by their position on the third axis. The quarters of the plot and the grey points are labelled i-v for reference to Fig 3A & 3B. Points are coloured with the hue-chroma- luminesence colour scheme to avoid perceptual artefacts from the rainbow colour scheme [61].

Fig 3. Changes in the landscape of citizen science over time, showing the emerging diversity (projects started in each period) and accumulated diversity (projects active in each period).

A shifting focus of project instigation (A, C-F, I), despite a lack of consistent increase in citizen science science diversity (G) has resulted in the accumulated diversity of citizen science over time (B, H). Coloured bars (A,B) represent angular position (see Fig 2 for details). Bar width is relative to number of projects, and black lines indicate position of cartesian axes (with the four quarters and the separate cluster labelled i-v). The shifting focus over time (C-F) is revealed by kernels (dark, mid and light grey indicating 5, 50 and 90% kernels, respectively) and angular position (individual projects projected as black points on the circle; the large point and arcs showing angular mean ± deviation), which are summarised for emerging (G) and accumulated (H) diversity. Projects are becoming increasingly different over time as indicated by proportion of kernel overlap and Watson U² test for differences in angular distributions (I: *<0.05; **<0.01).