A comparison of three methods to determine the subject matter in textual data

This study compares three different methods commonly employed for the determination and interpretation of the subject matter of large corpuses of textual data. The methods reviewed are: (1) topic modeling, (2) community or group detection, and (3) cluster analysis of semantic networks. Two different datasets related to health topics were gathered from Twitter posts to compare the methods. The first dataset includes 16,138 original tweets concerning HIV pre-exposure prophylaxis (PrEP) from April 3, 2019 to April 3, 2020. The second dataset is comprised of 12,613 tweets about childhood vaccination from July 1, 2018 to October 15, 2018. Our findings suggest that the separate "topics" suggested by semantic networks (community detection) and/or cluster analysis (Ward's method) are more clearly identified than the topic modeling results. Topic modeling produced more subjects, but these tended to overlap. This study offers a better understanding of how results may vary based on method to determine subject matter chosen.

Parental COVID-19 Vaccine Hesitancy in the United States

Objective:

Little is known about parents’ willingness to vaccinate their children against COVID-19. We assessed the prevalence of vaccine hesitancy among parents with a child or adolescent aged 12-15 years, examined predictors of parents’ COVID-19 vaccine hesitancy, their reasons for resisting a pediatric COVID-19 vaccine, and the correlation between parents’ intentions to vaccinate their child and the acceptance of a vaccine for themselves.

Methods:

We conducted a national online survey of 637 parents of a child or adolescent aged 12-15 years in March 2021, before COVID-19 vaccines had been approved for this age group. We assessed univariate predictors of vaccine hesitancy, and we used logistic regression analysis to assess independent effects of variables on vaccine hesitancy.

Results:

Nearly one-third (28.9%; 95% CI, 25.5%-32.5%) of respondents reported pediatric vaccine hesitancy. Vaccine-hesitant parents were less knowledgeable about vaccines, more accepting of vaccine conspiracies, and less worried about COVID-19 risks to their child’s health than vaccine-accepting parents were. Vaccine hesitancy was higher among female (vs male), single (vs married/living as married), older (vs younger), low income (vs high income), non–college graduates (vs college graduates), and Republican (vs Democrat) parents. The primary concerns expressed by vaccine-hesitant parents pertained to vaccine safety rather than vaccine effectiveness. One-quarter of vaccine-hesitant parents preferred that their child obtain immunity through infection rather than vaccination. Non–vaccine-hesitant parents’ reasons for vaccinating focused on protecting the health of their child and others. Childhood COVID-19 vaccine acceptance was strongly associated with parents’ intentions to get the vaccine for themselves.

Conclusion:

A messaging strategy for effective public health interventions that includes educating the public about vaccination, countering misinformation about vaccine development and safety, and stressing the safety of approved COVID-19 vaccines may boost vaccine acceptance among vaccine-hesitant parents.

Conservative Media Use and Childhood COVID-19 Vaccine Information: A Test of the Contradictory Health Information Processing Model

The Contradictory Health Information Processing (CHIP) model explains individuals' processing of conflicting health claims. Tests of the model, while highly supportive, have been experimental and have relied upon low-familiar topics. Accordingly, a survey of parents with a child aged <12 years (N = 510) was conducted to test the application of the CHIP model to the controversial issue of childhood COVID-19 vaccination; such a vaccine had not yet been approved for this age group at the time of the survey. As hypothesized, reliance upon conservative news was associated with the perception that media information contradicted official guidance to vaccinate children, which led to issue uncertainty. Issue uncertainty prompted negative appraisals and decision uncertainty. Specifically, decision uncertainty partially mediated the effect of issue uncertainty on negative appraisals of vaccination, which in turn aroused threat emotions. However, threat emotions did not predict information-seeking, as specified in the model. This result may have been due to respondents having already decided to vaccinate their child, or not - a reflection of the partisan nature of the topic and the extensive coverage it had received. Theoretical and practical implications are discussed.

Predictors of intention to vaccinate against COVID-19: Results of a nationwide survey

BACKGROUND

Public polling indicates that vaccine uptake will be suboptimal when COVID-19 vaccines become available. Formative research seeking an understanding of weak vaccination intentions is urgently needed.

METHODS

Nationwide online survey of 804 U.S. English-speaking adults. Compensated participants were recruited from the U.S. through an internet survey panel of 2.5 million residents developed by a commercial survey firm. Recruitment was based on quota sampling to produce a U.S. Census-matched sample representative of the nation with regard to region of residence, sex, and age.

RESULTS

COVID-19 vaccination intentions were weak, with 14.8% of respondents being unlikely to get vaccinated and another 23.0% unsure. Intent to vaccinate was highest for men, older people, individuals who identified as white and non-Hispanic, the affluent and college-educated, Democrats, those who were married or partnered, people with pre-existing medical conditions, and those vaccinated against influenza during the 2019-2020 flu season. In a multiple linear regression, significant predictors of vaccination intent were general vaccine knowledge (β = 0.311, p < .001), rejection of vaccine conspiracies (β = -0.117, p = .003), perceived severity of COVID-19 (β = 0.273, p < .001), influenza vaccine uptake (β = 0.178, p < .001), having ≥ 5 pre-existing conditions (β = 0.098, p = .003), being male (β = 0.119, p < .001), household income of ≥ $120,000 (β = 0.110, p = .004), identifying as a Democrat (β = 0.075, p < .029), and not relying upon social media for virus information (β = -0.090, p 〈002). Intent to vaccinate was lower for Fox News (57.3%) than CNN/MSNBC viewers (76.4%) (χ2(1) = 12.68, p < .001). Political party differences in threat appraisals and vaccine conspiracy beliefs are described.

CONCLUSIONS

Demographic characteristics, vaccine knowledge, perceived vulnerability to COVID-19, risk factors for COVID-19, and politics likely contribute to vaccination hesitancy.

Relationship of people's sources of health information and political ideology with acceptance of conspiratorial beliefs about vaccines

BACKGROUND

Conspiracies about vaccination are prevalent. We assessed how the health information sources people rely upon and their political ideologies are associated with acceptance of vaccine conspiracies.

METHODS

Online survey (N = 599) on Amazon's Mechanical Turk crowdsource platform. Hypotheses were tested via structural equation modeling.

RESULTS

Acceptance of vaccine conspiracy beliefs was associated positively with greater reliance on social media for health information (coef. = 0.42, p < .001), inversely related to use of medical websites (coef. = -0.21, p < .001), and not significantly related to use of providers for health information (coef. = -0.13, p = .061). In addition, liberal political orientation was negatively associated with acceptance of vaccine conspiracies (coef. = -0.29, p < .001).

CONCLUSIONS

An understanding of vaccine conspiracy acceptance requires a consideration of people's health information sources. The greater susceptibility of political conservatives to conspiracy beliefs extends to the topic of vaccination.