This document discusses key issues researchers face when conducting social science research using social media data. It addresses questions around appropriate quantitative and qualitative methods for analyzing different social media like Facebook and YouTube. It also discusses ethical issues around permission, anonymity, and credibility of social media data. Researchers must determine the best methods for new data sources while addressing validity, generalizability, and how well existing human research paradigms apply to publicly available online data.
Potential of AI (Generative AI) in Business: Learnings and Insights
Social media research issues
1. Information Studies
Social media and social science
research: What are the key issues
researchers face?
Mike Thelwall
Statistical Cybermetrics Research Group
University of Wolverhampton, UK
2. 1a. Quant: Exploiting Big Data
Social media can generate big data relevant
to social science research
Can new types of big data methods give new
insights into social media?
E.g., webometrics
Can old methods from existing fields be re-
purposed for social media?
E.g., variants of social network analysis for social
network sites?
E.g., sentiment analysis for social media texts?
3. #oscars
% matching posts
Proportion of tweets
mentioning the Oscars
9 Feb 2010 Date and time 9 Mar 2010
Av. +ve sentime
Subj. Sentiment strength
Just subj.
Increase in –ve sentiment strength Av. -ve sentime
Just subj.
9 Feb 2010 Date and time 9 Mar 2010
4. Big Data transformation example:
Large networks can sometimes be
produced and drawn with 30 minute
and computer software rather than
taking a small research project
5. 1b. Identifying appropriate
qualitative research methods
What is the best way to investigate
Facebook?
YouTube?
Can existing methods be repurposed?
e.g., virtual ethnography, blogs-as-diaries
Are completely new methods needed?
6. 2. Ethics
Does the human research paradigm fit
the web data – e.g., tweets?
Or documentary research?
Ask permission from everybody or
nobody to research their public web
postings?
Anonymise or give credit?
7. Wolverhampton Riots
YouTube video commenters
Names removed
Ages kept
Both could be discovered
Red arrows = happy replies, black = angry replies
8. Trialog 3. Lack of established
methods to analyse new social media
Increased uncertainty about methods
validity
Lack of precedent to justify methods
Increased attention on methods validity
Weaker data, if web content is analysed
Loss of control over sample selection,
compared to surveys and interviews
Potential lack of face validity
E.g., trolls
9. Network of links between
university web sites
5 per country
Asian and European nations
But how useful are links
as evidence of connections
such as collaboration?
10. Summary
Qualitative and qualitative research
issues
Which methods to use
How should they be used/modified?
What are their limitations?
What are the key ethical issues that
arise with methods?
Is there a quality issue with social
media research?