Characterizing the Prevalence of Obesity Misinformation, Factual Content, Stigma, and Positivity on the Social Media Platform Reddit Between 2011 and 2019: Infodemiology Study

Misinformation Bigram Stigma Disgust
DOI: 10.2196/36729 Publication Date: 2022-09-07T10:38:32Z
ABSTRACT
Reddit is a popular social media platform that has faced scrutiny for inflammatory language against those with obesity, yet there been no comprehensive analysis of its obesity-related content.We aimed to quantify the presence 4 types content on (misinformation, facts, stigma, and positivity) identify psycholinguistic features may be enriched within each one.All sentences (N=764,179) containing "obese" or "obesity" from top-level comments (n=689,447) made non-age-restricted subreddits (ie, smaller communities Reddit) between 2011 2019 contained one series keywords were evaluated. Four common natural processing extracted: bigram term frequency-inverse document frequency, word embeddings derived Bidirectional Encoder Representations Transformers, sentiment Valence Aware Dictionary Sentiment Reasoning, Linguistic Inquiry Word Count Program. These used train an Extreme Gradient Boosting machine learning classifier label sentence as 1 categories other. Two-part hurdle models semicontinuous data (which use logistic regression assess odds 0 result linear continuous data) evaluate whether select presented differently in misinformation (compared facts) stigma positivity).After removing ambiguous sentences, 0.47% (3610/764,179) labeled misinformation, 1.88% (14,366/764,179) 1.94% (14,799/764,179) positivity, 8.93% (68,276/764,179) facts. Each category had markers distinguished it other well external corpus. For example, higher average percent negations (β=3.71, 95% CI 3.53-3.90; P<.001) but lower number words >6 letters (β=-1.47, -1.85 -1.10; relative Stigma proportion swear (β=1.83, 1.62-2.04; first-person singular pronouns (β=-5.30, -5.44 -5.16; positivity.There are distinct properties can leveraged rapidly deleterious minimal human intervention provide insights into how population perceives patients obesity. Future work should these shared across languages platforms.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (59)
CITATIONS (7)