Mainstreaming structured analysis of web contents in post-Soviet area studies

Find the needle, characterise the haystack, and build community

Giorgio Comai (OBCT/CCI)

The project: Text as data & data in the text

Studying conflicts in post-Soviet spaces through structured analysis of textual contents available on-line

tadadit.xyz

Funding and disclaimers

This project is carried out with the support of the Italian Ministry of Foreign Affairs and International Cooperation under art. 23 bis, D.P.R. 18/1967. All opinions expressed within the scope of this project represent the opinion of their author and not those of the Ministry.

“Le posizioni contenute nel presente report sono espressione esclusivamente degli autori e non rappresentano necessariamente le posizioni del Ministero degli Affari Esteri e della Cooperazione Internazionale”

Starting point

In post-Soviet area studies:

  • so much of our research process is based on online sources
  • our encounter with online sources is overwhelmingly serendipitous
  • it’s all fine, but we’re missing out

What are the benefits of a structured approach?

  • find the needle
  • characterise the haystack (or put the needle in context)
  • validate, dismiss, correct assumptions more efficiently
  • case selection, timeframe selection
  • be in the position to describe what you did, and, if you didn’t find what you were looking for, move on

Structured ≠ quantitative

  • “structured” does not imply quantitative approaches
  • “structured” does not imply epistemological re-thinking

Examples

  • discoursive:
    • has Crimea been prominent in official or media discourse before 2014?
    • when did this obsession with Russophobia, denazification, etc. start?
  • finding data/evidence:
    • all sentences in Transnistria’s news that include “pension” AND either digits, percent, or rouble

Who’s using structured approaches?

  • Mostly people for whom the method is central to their research:
    • content analysis
    • often discourse analysis
    • quantitatively-oriented researchers (text as data)
  • their methods are often difficult to access for researchers with a different background

What are the obstacles to wider uptake?

  • technical complexity
  • costs
  • licensing issues
  • limited availability of ready-to-use corpora

How can we overcome these obstacles?

  • technical complexity -> community
  • costs -> community
  • licensing issues -> community
  • limited availability of ready-to-use corpora -> community

Easier access

Pre-processed corpora

Licensing

  1. interface: just a custom search engine?
  2. share it within research networks?
  3. share recipe to replicate

Finally, if you do the fancy structured analysis thing, think of the readers who usually don’t

  • “I did the fancy thing”
  • “if you’re also able to do fancy things, here’s how you can replicate what I did”

Better:

  • with one click and no technical skills, you can see what I found in its original context (example)
  • even better, you can test alternative hypotheses yourself

Why bother?

  • maybe it’s not really useful
  • obviously, there’s not a sensible incentive structure
  • why bother, really?
  • if we think of ourselves as a community of scholars, maybe this can work and enhance our research workflows