I recently distributed a survey for some research I did via social media. I used Twitter and online forums to ask people to complete the survey, and learned some very valuable lessons along the way.
Approach the gatekeeper first. Online forums and Twitter chats have a moderator or organiser who coordinates and oversees that particular community. It’s important to always check with this gatekeeper for permission to post first – don’t just spam forums with links or jump onto hashtags in the middle of Twitter chats. Apart from the obvious risk of coming across as as rude or irrelevant and therefore reducing the chances of anyone completing the survey, it’s also a requirement of the BERA (2018) education research guidelines. Following such guidelines and conducting research in an ethical way improves the image of research for the whole community. As a researcher, it also demonstrates duty of care for participants.
Distributing online surveys via social media introduces sample bias. Not every teacher uses social media in a professional capacity, and so using social media as the distribution method immediately skews the sample of respondents to those who do, or who are members of a particular online community. These respondents may not be representative of the teacher population as a whole.
Distributing surveys via social media produces a sample of respondents which cannot be generalised to a whole population. The teachers who kindly completed my survey produced data about their plans and intentions for teaching physical computing activities. This sample of teachers was obtained by opportunistic sampling (also known as convenience sampling) where they self-identified that they would like to take part. This meant that the sample was not representative of the general population of teachers as a whole and it was impossible to generalise the results with any confidence. I also asked respondents to retweet the survey on Twitter, which was a form of snowball sampling , and another non-probability technique which limited generalising the results.
There were some advantages to recruiting participants via Twitter and online forums. It was a relatively time-efficient way of gathering some data for analysis and gave an initial indication of possible trends in planning and delivering physical computing activities. In order to gather more rigorous data from a large population, however, I would look at other ways of obtaining the data for future studies.
Recruiting educational research participants via social media is still an emerging data collection method, and there are many ethical issues to consider. Gelinas et al (2017) provide examples of these in a medical context and it would be useful to cross-reference these into educational research.
Computation X 