Increasing accessibility of big data and proliferation of social media platforms like Facebook and Twitter have led to an exponential growth in studies drawing on these data sets, and progressively linked to the EU and its member states. However, currently few studies report their methods in any useful detail and reporting practices can be slipshod. Poor visualisation of data can be misleading in terms of the volume or significance of data covered and can distort the validity and meaningfulness of the findings. To address the bias problem, we use three collection strategies: gathering data using hashtags; extracting data from the random stream; and collecting from users known to be discussing the debate. We combine automated and manual processes. We argue that research designs employing multiple collection strategies also have an important role to play in contextualizing the wider relevance of research results.
