Data Collection and Quality Control of Citizen-Science Studies
A main concern of citizen-science studies is whether the collected data are reliable and comparable to professional studies. Four main aspects need to be considered in order to ensure or improve data quality: (1) preparation of easy and straightforward protocols, (2) training of volunteers, (3) in situ supervision by professional participation, and (4) validation of data and samples (modifi from Bonney et al. 2009). Of the citizen-science studies on marine litter examined, 55 % included at least one of these steps (e.g. Rosevelt et al. 2013; Anderson and Alford 2014; Gago et al. 2014).
Preparation of Easy and Straight-Forward Protocols
The studies that took measures to guarantee data quality, provided standardized protocols, guidelines and datasheets (e.g. Ribic 1998; Gago et al. 2014). In order to create clear protocols, some studies needed to adjust the sampling target to be easily identifi by citizen scientists. For instance, Ribic et al. (2010) found that citizen scientists occasionally missed small pieces of debris (no specifi size range was mentioned) in a monitoring program for beach litter. As a consequence of that observation, Hidalgo-Ruz and Thiel (2013) in a study focusing exclusively on small-plastic debris, decided to sample only items larger than 1 mm, which can be identifi by the naked eye after sieving of sand. Once the sampling target is determined, the marine litter items likely to be found by citizen scientists can be photographed and included in preparatory materials. Photographs of marine litter items were used in 15 % of the studies (e.g. Moore et al. 2009; Anderson and Alford 2014).
Training of Volunteers
Data quality can also be improved by volunteer training (e.g. Storrier and McGlashan 2006; Smith et al. 2014). Indeed, 38 % of the citizen science studies examined here included a degree of training or preparation of the volunteers. Training could consist of a one-hour classroom preparation (e.g. Smith et al. 2014) or a brief introduction in the field just before the sampling activity (e.g. Moore et al. 2009). For instance, a study on ghost fishing by derelict crab traps (Anderson and Alford 2014) was preceded by a training period. Furthermore, during one study year, participants were asked to take photos of every trap and to identify the organisms in the traps. These photos were later examined by professional scientists who confirmed that the data recorded for each trap were accurate.