TY - GEN
T1 - US news and social media framing around vaping
AU - Chen, Keyu
AU - Babaeianjelodar, Marzieh
AU - Shi, Yiwen
AU - Aanegola, Rohan
AU - Cheung, Lam Yin
AU - Nakov, Preslav Ivanov
AU - Yadav, Shweta
AU - Bancroft, Angus
AU - KhudaBukhsh, Ashiqur R.
AU - De Choudhury, Munmun
AU - Altice, Frederick L.
AU - Kumar, Navin
PY - 2023/2/11
Y1 - 2023/2/11
N2 - In this paper, we investigate how vaping is framed differently (2008–2021) between US news and social media. We analyze 15,711 news articles and 1,231,379 Facebook posts about vaping to study the differences in framing between media varieties. We use word embeddings to provide two-dimensional visualizations of the semantic changes around vaping for news and for social media. We detail that news media framing of vaping shifted over time in line with emergent regulatory trends, such as; flavored vaping bans, with little discussion around vaping as a smoking cessation tool. We found that social media discussions were far more varied, with transitions toward vaping both as a public health harm and as a smoking cessation tool. Our cloze test, dynamic topic models, and question answering showed similar patterns, where social media, but not news media, characterizes vaping as combustible cigarette substitute. We use n-grams and LDA topic models to detail that social media data first centered on vaping as a smoking cessation tool, and in 2019 moved toward narratives around vaping regulation, similar to news media frames. Overall, social media tracks the evolution of vaping as a social practice, while news media reflects more risk based concerns. A strength of our work is how the different techniques we have applied validate each other.
AB - In this paper, we investigate how vaping is framed differently (2008–2021) between US news and social media. We analyze 15,711 news articles and 1,231,379 Facebook posts about vaping to study the differences in framing between media varieties. We use word embeddings to provide two-dimensional visualizations of the semantic changes around vaping for news and for social media. We detail that news media framing of vaping shifted over time in line with emergent regulatory trends, such as; flavored vaping bans, with little discussion around vaping as a smoking cessation tool. We found that social media discussions were far more varied, with transitions toward vaping both as a public health harm and as a smoking cessation tool. Our cloze test, dynamic topic models, and question answering showed similar patterns, where social media, but not news media, characterizes vaping as combustible cigarette substitute. We use n-grams and LDA topic models to detail that social media data first centered on vaping as a smoking cessation tool, and in 2019 moved toward narratives around vaping regulation, similar to news media frames. Overall, social media tracks the evolution of vaping as a social practice, while news media reflects more risk based concerns. A strength of our work is how the different techniques we have applied validate each other.
KW - harm reduction
KW - social media
KW - vaping
UR - https://osf.io/kuxjp/overview
UR - https://www.scopus.com/pages/publications/85150872695
U2 - 10.1007/978-3-031-26303-3_17
DO - 10.1007/978-3-031-26303-3_17
M3 - Conference contribution
AN - SCOPUS:85150872695
SN - 9783031263026
VL - 13831
T3 - Lecture Notes in Computer Science (LNCS)
SP - 188
EP - 199
BT - Computational Data and Social Networks
A2 - Dinh, Thang N.
A2 - Li, Minming
PB - Springer
T2 - 11th International Conference on Computational Data and Social Networks, CSoNet 2022
Y2 - 5 December 2022 through 7 December 2022
ER -