Training: Data quality for Citizen Science, UCL

As part of our Citizen Science for All Talks (link to Citizen Science For All blog post please), on 25th May UCL organised a webinar on Data Quality for Citizen Science.

Our Citizen Science For All Talks are open to all audiences and involve 2 or 3 guest speakers giving a short talk on some aspect of citizen science. In this case, they were:

• Frank Ostermann
• Peter Mooney

Data quality is an often contested topic, linked to accuracy and completeness of science, and open to different interpretations from various stakeholders! It is also a mature topic, with an enormous list of publications available in Google Scholar. Data quality also is not quite the same thing as accuracy.

Citizen science often creates large datasets which professional scientists would be unlikely to produce on their own. Some stakeholders worry that citizen generated data may not be of high quality, so Peter Mooney told us about the guidelines he would like to draw up to produce data that is both useful and usable. Different stakeholders will have different requirements regarding data. Neither professionals nor amateurs have a monopoly on good data – and in fact there is plenty of room to blur the lines between the two.

Frank Ostermann told us that data quality may be to do with coverage or with representativeness. He showed us a slide that definitely highlights how many different ways data can be accurate, or not!

In fact, in some citizen science projects, he told us that sometimes there is more concern about stakeholders’ uncertainty about data quality than about the data itself! He asks if there is a way to standardise quality checks, but cautions against too much standardisation: citizen scientists often think of completely new things, and there is a great deal of diversity in what they produce, which could be lost if everything has to be done in one set way. He also pointed out that not everyone has the same access to resources or digital tools – “the digital divide”. Therefore, if there is a lack of observation where we would expect one, is this due to nobody choosing to report, or that nobody was able to report?

We thank Peter and Frank for their interesting talks. You can catch up on our livetweets here, or watch these talks on YouTube here.