Twitter Population Sample Bias and its impact on predictive outcomes

Filho, Renato Miranda; Almeida, Jussara M.; Pappa, Gisele L.

doi:10.1145/2808797.2809328

Cited by 26 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Our study provides some reason for hope: After controlling for demographics, social media's association with vote choice and political attitudes greatly declines. This suggests social media data could be used for studying public opinion and forecasting if the data is appropriately weighted using demographics (which some work has already begun to (Filho et al, 2015)) and political attitudes or adjusted using an approach such as multilevel regression and post-stratification (Wang et al, 2014). 7 Although social media data provides numerous opportunities for political science, it is vital to remember that Twitter and Facebook are not representative of the general population.…”

Section: Resultsmentioning

confidence: 99%

Twitter and Facebook are not representative of the general population: Political attitudes and demographics of British social media users

Mellon¹,

Prosser

2017

Research & Politics

313

187

View full text Add to dashboard Cite

A growing social science literature has used Twitter and Facebook to study political and social phenomena including for election forecasting and tracking political conversations. This research note uses a nationally representative probability sample of the British population to examine how Twitter and Facebook users differ from the general population in terms of demographics, political attitudes and political behaviour. We find that Twitter and Facebook users differ substantially from the general population on many politically relevant dimensions including vote choice, turnout, age, gender, and education. On average social media users are younger and better educated than non-users, and they are more liberal and pay more attention to politics. Despite paying more attention to politics, social media users are less likely to vote than non-users, but they are more likely to support the left leaning Labour Party when they do vote. However, we show that these apparent differences mostly arise due to the demographic composition of social media users. After controlling for age, gender, and education, no statistically significant differences arise between social media users and non-users on political attention, values or political behaviour.

show abstract

Section: Resultsmentioning

confidence: 99%

Twitter and Facebook are not representative of the general population: Political attitudes and demographics of British social media users

Mellon¹,

Prosser

2017

Research & Politics

313

187

View full text Add to dashboard Cite

show abstract

“…In the same research it has been detected that Twitter users differ from the general population in terms of demographics, political attitudes and political behaviour. Since biases in these relevant dimensions can be significant, a possible solution for using social media data for studying public opinion and forecasting elections requires the data to be appropriately weighted using demographics, as suggested by [22]. Future research should be seek to overcome some of these limitations and integrate the temporal component into unified framework proposed.…”

Section: Discussionmentioning

confidence: 99%

Geographies of Twitter debates

Gobbo

Fontanella

et al. 2021

J Comput Soc Sc

View full text Add to dashboard Cite

Over the last years, the prodigious success of online social media sites has marked a shift in the way people connect and share information. Coincident with this trend is the proliferation of location-aware devices and the consequent emergence of user-generated geospatial data. From a social scientific perspective, these location data are of incredible value as it can be mined to provide researchers with useful information about activities and opinions across time and space. However, the utilization of geo-located data is a challenging task, both in terms of data management and in terms of knowledge production, which requires a holistic approach. In this paper, we implement an integrated knowledge discovery in cyberspace framework for retrieving, processing and interpreting Twitter geolocated data for the discovery and classification of the latent opinion in user-generated debates on the internet. Text mining techniques, supervised machine learning algorithms and a cluster spatial detection technique are the building blocks of our research framework. As real-word example, we focus on Twitter conversations about Brexit, posted on Uk during the 13 months before the Brexit day. The experimental results, based on various analysis of Brexit-related tweets, demonstrate that different spatial patterns can be identified, clearly distinguishing pro- and anti-Brexit enclaves and delineating interesting Brexit geographies.

show abstract

“…The use of the public Twitter streaming API to collect data pre-filtered only for tweets including latitude and longitude coordinates represent a subset of all tweets posted during the time frame of the study. There exists the potential for sampling bias associated with different Twitter APIs that are not representative of all Twitter data (e.g., Firehose data), and data filtered only for geocoded data may omit many conversations from collegeaged populations about topics, such as smoking, which may be linked to college-related user groups (see "Limitations" section for more details) (50). Though resulting in a much smaller volume of data, our approach nevertheless allows for detection of tweets in specific geospatial bounds at the high resolution of latitude and longitude coordinates in the state of California.…”

Section: Data Collectionmentioning

confidence: 99%

“…This method of data collection may have introduced bias in the types of tweets collected, thereby limiting the generalizability of findings as the majority of Twitter users do not geolocate their posts. Potential sampling biases for Twitter include oversampling for certain geographic areas (e.g., there are a higher number of U.S. Twitter users than other countries), filtering for specific features (e.g., language, location), and the limitations of the Twitter public streaming API (used in this study) in lieu of other data collection approaches (e.g., Twitter REST and SEARCH APIs) (50). Future studies should examine the use of multiple Twitter APIs to generate a more representative Twitter dataset (including different strategies for filtering, demographic characterization, and purposeful user sampling) and compliment findings with other traditional sources of data (e.g., survey data, focus groups, clinical records, etc.)…”

Section: Limitationsmentioning

confidence: 99%

Characterizing Self-Reported Tobacco, Vaping, and Marijuana-Related Tweets Geolocated for California College Campuses

Cuomo

Purushothaman

Li³

et al. 2021

Front. Public Health

View full text Add to dashboard Cite

Introduction: College-aged youth are active on social media yet smoking-related social media engagement in these populations has not been thoroughly investigated. We sought to conduct an exploratory infoveillance study focused on geolocated data to characterize smoking-related tweets originating from California 4-year colleges on Twitter.Methods: Tweets from 2015 to 2019 with geospatial coordinates in CA college campuses containing smoking-related keywords were collected from the Twitter API stream and manually annotated for discussions about smoking product type, sentiment, and behavior.Results: Out of all tweets detected with smoking-related behavior, 46.7% related to tobacco use, 50.0% to marijuana, and 7.3% to vaping. Of these tweets, 46.1% reported first-person use or second-hand observation of smoking behavior. Out of 962 tweets with user sentiment, the majority (67.6%) were positive, ranging from 55.0% for California State University, Long Beach to 95.8% for California State University, Los Angeles.Discussion: We detected reporting of first- and second-hand smoking behavior on CA college campuses representing possible violation of campus smoking bans. The majority of tweets expressed positive sentiment about smoking behaviors, though there was appreciable variability between college campuses. This suggests that anti-smoking outreach should be tailored to the unique student populations of these college communities.Conclusion: Among tweets about smoking from California colleges, high levels of positive sentiment suggest that the campus climate may be less receptive to anti-smoking messages or adherence to campus smoking bans. Further research should investigate the degree to which this varies by campuses over time and following implementation of bans including validating using other sources of data.

show abstract

Twitter Population Sample Bias and its impact on predictive outcomes

Cited by 26 publications

References 8 publications

Twitter and Facebook are not representative of the general population: Political attitudes and demographics of British social media users

Twitter and Facebook are not representative of the general population: Political attitudes and demographics of British social media users

Geographies of Twitter debates

Characterizing Self-Reported Tobacco, Vaping, and Marijuana-Related Tweets Geolocated for California College Campuses

Contact Info

Product

Resources

About