An online cross-sectional survey was conducted using Qualtrics. This research was approved by the IRB at the University of North Dakota (UND), which considered the study exempt (IRB # IRB005352).
Sample
Students, staffs, and faculty over the age of 18 at UND were invited to participate in the survey. The enrollment began in October 2022, the last participant answered in November 2022, and the online survey was closed in December 2022. In total, 1,043 participants accessed the survey, among them, 1,038 were eligible, indicating that they were an on-campus student, faculty or staff member at UND. We observed a response rate of 7.90% (1,038/13,143) among eligible individuals invited to participate. Non-responses may be due to various factors including time restraints or disinterest [27, 28]. Distance students and remote employees were not eligible. In total, 914 participants completed the survey and were used in the final data analysis.
Procedure
All students, faculty, and staff at UND were recruited via mass email, which contained a link to access the Qualtrics survey. Those who access the survey link were first asked to read and agree to a consent form. All participants who consented and reported that they were a student, faculty, or staff at UND were eligible to proceed. Participants were then asked to answer demographic questions before answering questions regarding their vaccination status, beliefs, and attitudes about vaccination regarding the following routine vaccines: COVID-19, influenza, HPV, MenB, MenACWY, Tdap, MMR, varicella, shingles, Hepatitis A, and Hepatitis B. The survey took approximately 15 min. After completing the survey, participants received a $10 gift card in exchange for their participation in the survey.
Measures
The survey questionnaire included questions measuring participants’ trust in relevant stakeholders, vaccine knowledge, information sources, vaccine confidence, and sociodemographic and health characteristics. Questions included multiple response options adapted by scales used in previous literature and previously piloted iterations of survey questionnaire [29]–[36]. The reliability statistics for constructed variables were found to be satisfactory (see the Variable Construction section below). The pilot-testing of the survey questionnaire among a small group of undergraduate students demonstrated that there was little to no confusion or questions regarding the questionnaire.
Outcome (vaccine confidence)
Vaccine confidence.
Confidence in vaccine information sources was measured by asking the following question: “Overall, how confident are you that you could get advice or information about vaccines if you needed it?” Reponses ranged from 1 (“not at all”) to 5 (“a lot”) [35].
Independent variables (Trust, Knowledge, and information source)
Trust. Trust in vaccine information sources was measured by asking the following question: “In general, how much would you trust information about vaccines from each of the following?” Five sources of trust were provided: (1) “A doctor” (2), “Family or friends” (3), “Government health agencies” (4), “Charitable organizations” (5), “Religious organizations and leaders” [29, 35]. Response options for each source of trust ranged from 1 (“not at all”) to 4 (“a lot”).
Vaccine knowledge
Questions seeking individual knowledge of vaccines and the diseases they prevent included seven true-or-false questions about the influenza (flu) vaccine (e.g., “The influenza vaccine is only minimally effective”) [31, 32], six true-or-false questions about the HPV vaccine (e.g., “Men cannot get HPV”) [33], three true-or-false questions about the Tdap vaccine (e.g., “Adults should get a tetanus booster every 10 years”) [30], and three true-or-false questions about the shingles vaccine (e.g., “The chance of developing shingles increases with age”). Three response options were provided for each question: “True”, “False”, or “I don’t know” [34]. Detailed information about the creation of vaccine knowledge variable is presented under the results section below.
Information sources
To measure individuals’ general information sources in terms of vaccine information, the following question was asked: “Imagine that you had a strong need to get information about vaccines. Where would you go first?” [35]. Response options included (1) “Books” (2), “Brochures, pamphlets, etc.” (3), “Vaccine organizations” (4), “Family” (5), “Friend/Co-worker” (6), “Doctor or health care provider” (7), “Internet” (8), “Library” (9), “Magazines” (10), “Newspapers” (11), “Telephone information number” (12), “Complementary, alternative, or unconventional practitioner” (13), “Other (please specify).
Another question was asked to identify individuals’ Internet sources regarding vaccine information: “If you search the Internet for vaccine information, please rank the following from 1 to 12, with 1 being your first place to look for information. Twelve Internet sources included: (1) “Blogs” (2), “Facebook” (3), “Twitter” (4), “Instagram” (5), “Tik Tok” (6), “YouTube” (7), “TV news website (NBC, FOX, CNN, etc.)” (8), “Newspaper/magazine website (Wall Street Journal, The Atlantic, etc.)” (9), “Government websites that end in .gov” (10), “Websites of nonprofits that end in .org” (11), “Websites of universities that end in .edu”, and (12) “Other online sources”. Detailed information about the creation of information source variables is presented under the results section below.
Covariates
Covariates regarding demographic characteristics and socioeconomic status included [35]: gender, race (white vs. non-white), political ideology (liberals, neutral, vs. non-liberals), education level (some college or more vs. lower), marital status (married vs. single), general health (good vs. poor), religious ideology (atheist vs. non-atheist), age, and insurance status (insured vs. not insured) [36]. Original questions had multiple response options, for example, race included “White”, “American Indian/Alaska Native”, “Black or African American”, “Asian or Pacific Islander”, “Hispanic, Latinx or Spanish”, “Mexican, Mexican American, Chicano”, “Puerto Rican”, “Cuban”, “Another Hispanic, Latinx, or Spanish origin” but were merged into top responses.
Variable construction
For the vaccine knowledge variable, responses were recoded depending on the correct answer for a question. For all questions “I don’t know” was coded as “incorrect”. For each set of knowledge questions (i.e., flu, shingles, tetanus, and HPV), “correct” answers were totaled to create a knowledge index for each set. Once all correct answers between all sets of knowledge questions were totaled, a final knowledge index was created to determine the level of vaccine knowledge. Higher index scores indicated higher knowledge.
For information source variables, the top ranked answers from respondents were used to create three dichotomous variables. First, 55.5% of respondents selected “doctor or health care provider” (out of 13 response options) as their answer to the following question – “Imagine that you had a strong need to get information about vaccines. Where would you go first?”, which led to the creation of expert primary source variable (yes/no). Second, 32.8% of respondents selected “Internet” (out of 13 response options) as their answer to the aforementioned question, which led to the creation of Internet primary source variable (yes/no). Third, 49.3% of respondents selected “Government websites that end in .gov” (out of 12 response options) as their answer to the following question – “If you search the Internet for vaccine information, please rank the following from 1 to 12, with 1 being your first place to look for information.” This led to the construction of government primary source variable (yes/no). These three information source variables are reflective of popular vaccine information sources available to the general public [13, 37].
Data analyses
First, descriptive statistics were obtained. Next, logistic regressions estimated regression coefficients (B) and 95% confident intervals (CI) for associations between independent variables – trust in five sources: (1) doctors (2), family/friends (3), government health agencies (4), charitable organizations, and (5) religious organizations, vaccine knowledge and three information sources: (1) expert primary source (2), Internet primary source, and (3) government primary source – and the dependent variable, vaccine confidence, accounting for gender, race, political ideology, education level, marital status, self-reported health status, religion, age, and insurance status. Given the low number of missing variables, a complete-case analysis was deemed appropriate. As such, missing and incomplete responses were excluded.