This week I’m residing in London for the Annual Meeting of the International Communication Association, meeting friends and colleagues and looking at some new developments in social media research. After an extensive ice cream tasting yesterday, today’s first sessions started off quite well.
My personal highlight was a panel on methodological challenges (yep, big issue for me, too) and the opprtunities that „big data“ offer for insights on what people do online and why, comprising high quality presentations on detecting influencers and conducting sentiment analsis as well as insights from Facebook Research. These are my take home-messages (for my dissertation and beyond):
- „Big data“ isn’t everything. Or so it seems when even Facebook is not solely relying on obejctive behavioral masses of data for business development. Long live the survey.
- EdgeRank is evil ;) – this is something that’s been know in social media marketing/PR for some time now but today we were provided with a few examples of how Facebook algorithms also affect data collection.
- Online environments change, including privacy settings, profile layouts, rating features, hashtag incorporations – you name it. Scientific research on social media mechanisms and underlying psychological and communicative processes is – once published – supposed to be citeable for more than a few months. Thus, more consideration needs to be given to the issue of constant changes when addressing topics that are closely associated with certain features of the platforms in question. Yup, I remember not being amused when the Timeline was introduced during my data collection.
- New technologies and analytical methods combined with limited ressources encourage a tendency to search for one’s car key under the light. The panel’s respondent gave a polite reminder that it would be far more fruitful to spend some time on working out ways to look for the car key where it’s been lost.
- Finally, a nice insight into Sentistrength, maybe useful for sentiment analysis.