The Wardle Test for a #socialmedia #selfie effect in science

2046472-367fff8e

‘And I am immortal’ (through social media).
Connor MacLeod (The Highlander).

A recent paper in the journal Ideas in Ecology and Evolution inspired me to rethink/temper my optimism in social media as a panacea for effective scientific communication. The running title of the paper, how to tweet your way to honour and glory, by David Wardle captures several primary concerns with altmetrics as a tool to estimate merit, value, or even global reach. We are discussing these ideas at NCEAS today, and as a heuristic, I prepared the following deckumentary (commentary + slide deck). The strengths and limitations of social media as a tool to communicate science are explored.  Several basic solutions are proposed. However, there is an incredible opportunity here to more throughly examine how we handle social media as a tool and evaluate its capacity for effective outreach.

One of the highlights proposed in the article that I really enjoyed but want to emphasize more directly here is the test of a particular potential limitation – non-independence of outreach from the social media stream of the producer.  I propose we should entitle the test developed The Wardle Test for a social-media selfie effect in science.

The social-media selfie effect workflow

  1. Select a set of products with different authors but from a similar outlet (i.e. a journal).
  2. Structure sampling of products to ensure reproducibility (i.e. regular, random, or random-stratified sampling from the outlet), and ensure author-identities are unique in each instance.
  3. Record altmetric scores reported for each product.
  4. Capture twitter-stream for each product.
  5. Assign tweets to product producer (rule: personal twitter account matches first author or organization such as lab) or other (potentially independent twitter account).
  6. Contrast altmetric scores between products tweeted by producers relative to others.

Fantastic idea as a proxy for the positive and negative ‘echo-chamber’ effect discussed widely online. We need an r-script to scrape a larger set of products and associated accounts!

Then, can can calculate not only this social-media selfie effect but also explore some of the contemporary analytical solutions produced online by many ‘influence’ indices including diversifying the signal analysis, weighting (often by audience), and normalization.

The ‘quickening’ of social media amplification is perhaps not immortal, but it is a challenge and thus opportunity for scientific communicators and critical citizens to better validate and use this effect appropriately.

637b8a68953b08d898f2f564c9486dfd