Written by: Josh Rosenberg
Primary Source: Joshua M. Rosenberg – December 9, 2016
Over the past year, I did research on how teachers use Twitter across our graduate degree program for teachers here at MSU, and State Educational Twitter Hashtags, like #miched and #nced.
Since looking at State Educational Twitter Hashtags for the first six months of 2016, we collected data for the 2015-2016 school year.
An idea was to look at the data from a ten-thousand foot view using a very simple and maybe useful approach using the Linguistic Inquiry and Word Count software for analyzing texts.
After I loaded the tweets(1.3 million over the 2015-2016 school year) into LIWC, and a few moments (okay, minutes) later, out popped a spreadsheet with the tweets as well as the proportion of each response determined with LIWC to be in one of seven “main” categories I chose to focus on out of the approximately 60 or so categories available:
- Social: Family and friend-related text
- Drives: Affiliation, achievement, and “power”-related text
- Cognitive Processes: Insight and reasoning-related text
- Affect: Insight, causation, and tentativrelated text
- Time Focus: Time-related (i.e., future) text
- Perceptual Processes: Seeing, hearing, and feeling-related text
- Biological Processes: Health-related text
Here is what the results look like in terms of the patterns over the year.
As a ten-thousand foot view, what – if anything – can this tell us about what teachers talk about on Twitter? Well, teachers are not talking about Biological Processes (except for the biology teachers) or Perceptual Processes very much. It looks like there are a high proportion, and slight increases as the year goes in tweets LIWC determined relate to the Social and Drives categories, and a high proportion, and slight decreases as the year goes in, in tweets that relate to Cognitive Processes
My two cents? I’m not sure how well it worked.
I wanted to think about use of LIWC and how it relates to other approaches to examining text, and this was a good sandbox to try it out.
And, there are finer-grained categories (i.e., Drives includes “affiliation”, “achievement”, and “risk” sub-categories) that may be worth a look at. But, I think this is a case (like most) when diving into the tweets (and actually looking at their content) is critical. But, this could (maybe) give insight into what to look for, in terms of the categories, and maybe even where to look, in terms of focusing on tweets LIWC determined relate to one or more of the categories.
Latest posts by Josh Rosenberg (see all)
- A Shiny interactive web application to quantify how robust inferences are to potential sources of bias (sensitivity analysis) - January 19, 2018
- Outcomes from a self-generated utility value intervention in science (in IJER) - December 30, 2017
- Review of ‘What’s Worth Teaching: Rethinking Curriculum in the Age of Technology’ - November 7, 2017