Twitter Political Ideology Project

Background

The goal of this project is to determine the political ideology of psychology and biology academics on Twitter, and explore whether there are differences in the social networks of academic on the basis of political ideology.

Data Collection

Social Psychologists

Names of social psychologists were collected from the Member Directory of the Society for Personality and Social Psychology (SPSP) using the search bar. Every letter of the alphabet was entered into the search bar for first name and last name (e.g., First Name: A, Last Name: A, First Name: B, Last Name: B, etc). The first 200 names displayed (the maximum number displayed) were copied.

Data collection date: 13 October 2020
Twitter extraction date: 17 October 2020

Sample size:

• Number extracted from SPSP: 7448
• After removing people who did not specify country: 7406 (-42)
• After removing duplicates: 5117 (-2289)
• Matching Twitter names: 2756 (-2361)

Evolutionary Biologists

Names of evolutionary biologists were collected from the Member Directory of the Evolution Directory (Evoldir).

Data collection date: 18 November 2019
Twitter extraction date: 17 October 2020

Sample size:

• Number extracted from Evoldir: 10749
• Matching Twitter names: 5186 (-5563)

Evolutionary Psychologists

Names of evolutionary psychologists were collected from the Member Directory of the Human Behavior & Evolution Society (HBES).

Data collection date: 18 November 2019
Twitter extraction date: 17 October 2020

Sample size:

• Number extracted from HBES: 579
• Matching Twitter names: 336 (-243)

R Script Summaries

1. data_cleaning.R

This script takes the raw names, countries, and text copied from the databases and converts it into tidy datasets that are easier to work with.

2. extract_user_bios.R

In this script, the names extracted from the databases is searched on Twitter. Any matches are saved, and Twitter bios are searched to determine if they contain any keywords that might indicate that they are an academic (e.g., academia, PhD, researcher). Other info extracted includes the number of friends, followers, Twitter location, account creation date, and number of twitter posts. Their likely gender (based on historical name data) is added. The Twitter API only permits 900 requests per 15 minutes, so the script pauses once this limit is reached and starts again when the limit resets. Where there are multiple results, these are saved as separate rows. The data are saved as csv files and google sheet files, ready to be manually rated using a Shiny app.

3. rate_evo_psychs

This is a Shiny App that allows users to manually rate which of multiple Twitter users extracted in the script above, if any, is an evolutionary psychologist. Users are displayed in a table, together with their Twitter name and Twitter bio. The user selects the correct user, or skips if none appear to be academics. Results are saved in a googlesheets file, which can then be downloaded.

4. add_location.R

Twitter locations, where they exist, can include countries, cities, or suburbs. In this script the Google Maps API is used to standardize location by extracting latitude, longitude, and country from location data.

5. match_users.R

To reduce the likelihood that differences in social networks or political ideology of evolutionary psychologists, social psychologists, and evolutionary biologists are due to differences in other variables, social psychologists and evolutionary biologists were matched to the completed and rated list of evolutionary psychologists using optimal pair matching. They were matched by country, likely gender, number of friends, number of followers, number of twitter posts, and number of retweets.

At this stage the list of social psychologists and evolutionary biologists has not yet been rated, so not all the Twitter users in this list will actually be academics. To account for this, 5 users from each list were paired with the evo psych list to ensure that at least one of the users will actually be an academic. These are rated in the Shiny app below.

6. rate_evol_biol & rate_social_psychs

These two Shiny apps rate evolutionary biologists and social psychologists (similar to rate_evo_psychs). From the script above, 5 potential social psychs and evol biols were matched with each evo psych. This app allows you to manually rate which of the 5 options (from closest match to least close match), if any, is a social psych or evo biologist.