News flash: citizen scientists are trainable

20 11 2014

Not to be smug about it, but SEANET has been doing citizen science since before citizen science was cool. Or at least, since before it was widely accepted as a viable and valid method of collecting large scale data sets. For a long time, the critique from the science establishment was that lay people without formal science training could not match the quality of data generated by trained scientists. Over time, more and more citizen science projects have taken off and produced high quality data sets on a scale impossible for scientists and their exhausted grad student minions to replicate. Now, in the face of these data, many former naysayers have had to admit that citizen science plays a role nothing else can. Professional scientists needn’t have feared; the rise of citizen science can, in many cases, free them up to do things besides constantly collecting data themselves. Study design, data analysis, publication of results, all these things remain in their purview.

As these views have evolved, it has remained critical to ensure that citizen scientists, who often collect data with little to no direct supervision, are actually doing a good job. In some projects, that requires intensive training before joining up–learning how to identify organisms, or take measurements with accuracy and precision, for example. These trainings were often in person and labor intensive, and sometimes also cost prohibitive, so many programs began looking for alternative methods of training, whether with print materials or mult-media ones.

The slick and flashy app interface of the Outsmart cit sci project.

The slick and flashy app interface of the Outsmart cit sci project.

A new study out in PLOS one looked into the relative effectveness of print vs. video training for volunteers monitoring invasive plant species in Massachusetts. They determined that video training out performs plain print resources (photos and text descriptions of plants), but, perhaps more importantly, rivals in person training in preparing volunteers who can correctly identify invasive species. Moreover, when volunteers did incorrectly identify species, they tended to be plants that were challenging for all volunteers, regardless of how they were trained. Identifying exotic honeysuckles, for instance, is hard for almost everyone, while multiflora rose was identified by nearly 100% of volunteers.

For SEANET, we have chosen to put more work in on the back end of data collection. Our volunteers are not required to have any knowledge of species identification when they join (though they tend to acquire it if they stick around); instead, every bird must be photographed so that it can be reviewed for accuracy as to species. For a program of SEANET’s relatively modest size, this works well, and I am able to review everyone’s reports individually. But on a scaled up project–something like ebird for instance–the program leaders must rely on the volunteer’s abilities and experience, and only reports that seem strange or outlandish are challenged and followed up on. If SEANET were to get so big as that, we’d have to modify things.

One thing the invasive plant project studied in this article has that I do indeed covet is a very slick and attractive interface in a smartphone app. We’ve talked about having such a thing for SEANET for a long time, but as of now, it’s not in the cards. But you can all gaze in wondering admiration at this plant project’s interface, and, if you are located in the northeast, you might even consider participating. If, of course, it wouldn’t cut into your Seanetting.