The effect of falsely balanced reporting of the autism–vaccine controversy on vaccine safety perceptions and behavioral intentions
- 1Department of Communication, Cornell University, 336 Kennedy Hall, Ithaca, NY 14850 and 2Department of Communication, George Mason University, MS 3D6 4400 University Drive, Fairfax, VA 22030, USA.
- ↵*Correspondence to: G. Dixon. E-mail: gnd5{at}cornell.edu
- Received April 15, 2012.
- Accepted October 25, 2012.
Abstract
Controversy surrounding an autism–vaccine link has elicited considerable news media attention. Despite being widely discredited, research suggests that journalists report this controversy by presenting claims both for and against a link in a relatively ‘balanced’ fashion. To investigate how this reporting style influences judgments of vaccine risk, we randomly assigned 320 undergraduate participants to read a news article presenting either claims both for/against an autism–vaccine link, link claims only, no-link claims only or non-health-related information. Participants who read the balanced article were less certain that vaccines are safe, more likely to believe experts were less certain that vaccines are safe and less likely to have their future children vaccinated. Results suggest that balancing conflicting views of the autism–vaccine controversy may lead readers to erroneously infer the state of expert knowledge regarding vaccine safety and negatively impact vaccine intentions.
Introduction
Although vaccination is considered one of the greatest public health achievements of the last century [1, 2], controversy surrounding a link between vaccines and autism has elicited considerable media attention and public concern [3]. Despite a strong medical and scientific consensus backed by rigorous epidemiological studies indicating no link between autism and vaccines [4–7], research suggests that journalists in the United Kingdom and United States often report this controversy by presenting claims both for and against a link in a relatively ‘balanced’ fashion [3, 8, 9]. In some cases, so-called ‘falsely balanced’ reporting fails to mention which claim is supported by a scientific consensus [10].
With news media as an important source of health- and vaccine-related information [11, 12], health officials have argued that coverage of the autism–vaccine controversy increased public uncertainty about vaccine safety and decreased vaccine uptake [13–15]. Studies partially support this argument, finding that following media reports of the autism–vaccine controversy, public confidence around vaccination declined [8], while selective non-receipt of vaccination increased in some localities [16, 17]. Focus group interviews also point out that news coverage of the controversy potentially increased audience uncertainty regarding vaccine safety [9, 18, 19]. Despite these findings, the aforementioned studies either do not measure participants’ media exposure or use experimental methods to establish a causal association between media exposure and vaccine beliefs and decision making. Using experimental methods, we extend this research by examining how specific forms of news messages on the autism–vaccine controversy (i.e. balance) influence readers’ judgments of vaccine safety and intentions to vaccinate—factors that are highly predictive of vaccination uptake [20–23].
Methods
An experiment was administered to undergraduate students (N = 320; M = 19.9 years; SD = 1.56; 67% female) recruited from an online database at a mid-sized university in northeastern United States. Undergraduate students were chosen as a convenience sample. The study received ethical approval from the Institutional Review Board.
Procedure
Participants were randomly assigned to read one of four real news articles from prominent newspapers and television stations published online and in print no >1 year prior to data collection. The news articles were chosen through online searches using terms such as ‘autism and vaccines’ and ‘autism vaccine controversy’. Aside from selecting articles based on their presentation style (i.e. balanced, link and no link), we preferred recently published articles (no later than 1 year prior to our experiment) from prominent news media with high readership. Articles were standardized using identical font/size and removing source identification. The four articles were coded as either ‘no-link’ (presenting only the perspective that vaccines do not cause autism), ‘link’ (presenting only the perspective that vaccines could cause autism), ‘falsely balanced’ (presenting both perspectives without mention that a scientific/medical consensus rejects a link) or ‘control’ (non-health-related article).
After completing a questionnaire requesting demographic data (age and sex) and their intention to have children, participants were given their assigned news article to read. After completing the reading, participants answered a questionnaire assessing their evaluation of how their respective article presented the autism–vaccine link (the manipulation check); how certain they are about vaccine safety and how certain they think scientists are about vaccine safety and intention to vaccinate oneself and their future children.
Participants were debriefed and provided with a vaccine information sheet upon conclusion of the experiment.
Dependent variables
Certainty was measured using a six-point certainty scale (1 = very uncertain; 6 = very certain), adapted from Corbett and Durfee [24]. Participants’ ‘personal’ certainty about vaccine safety consisted of the following single item:
-
(i) Please indicate how certain ‘you’ are that vaccines are safe to receive.
Participants’ perception of scientists’ certainty about vaccine safety was measured using the following single item:
-
(ii) Please indicate how certain you believe ‘scientists’ are that vaccines are safe to receive.
Behavioral intention was measured using a six-point Likert agreement scale (1 = strongly disagree; 6 = strongly agree) adapted from Fishbein [23]. Self-intention to be vaccinated consisted of the following single item:
-
(i) I intend to receive any vaccination that my health care provider recommends in the future.
Intention to vaccinate ones’ future children consisted of the following single item:
-
(ii) If I were to have children in the future, I will make sure they receive any vaccination that my health care provider recommends.
We measured the dependent variables using a one-way analysis of variance (ANOVA), with group as the independent variable. Post hoc analyses were performed to measure differences between groups for each dependent variable.
Manipulation check
To establish measurement validity, the manipulation check assessed whether participants exposed to a particular article actually perceived the article as presenting link only claims, no-link only claims or balanced claims (there was no check for the control article). Using a six-point Likert agreement scale, participants responded to what extent they thought their article:
-
Presents only the perspective that vaccines do not cause autism.
-
Presents only the perspective that vaccines could cause autism.
-
Presents both perspectives stating (1) vaccines could cause autism and (2) vaccines do not cause autism.
Articles were not quantitatively coded in terms of the percentage of each one devoted to link and no-link perspectives. Instead, we relied on participant perceptions that our ‘balanced’ article did, in fact, appear to reflect both link and no-link positions more so compared with participants receiving link only and no-link only articles.
Results
A series of ANOVAs found no differences among groups in terms of age (P = 0.315) and gender (P = 0.527). In addition, 94% of participants reported that they intend to have one or more children in the future.
Manipulation check
The manipulation check assessed whether participants who read a particular article actually perceived it as link, no-link or balanced. Participants across the three groups included in the manipulation check (the control group was not included) rated their respective article in agreement with our coding (Table I).
Multiple comparison table of the manipulation check
Personal certainty of vaccine safety
A significant effect was observed, F(3, 314) = 5.536, P < 0.001, η2 = 0.05. Using a Bonferroni post hoc comparison test, participants who read the falsely balanced article reported lower certainty scores that vaccines are safe (M = 4.08, SD = 1.13) than participants in the no-link (M = 4.67, SD = 0.94), P < 0.01 and the control group (M = 4.61, SD = 1.07), P < 0.01, but not the link group (M = 4.375, SD = 0.93). In terms of percent, participants in the balanced group reported certainty scores that were on average 13% lower than those in the no-link group and 12% lower than those in the control group (Fig. 1).
Participants’ personal certainty that vaccines are safe. A six-point certainty scale (1 = very uncertain; 6 = very certain) adapted from Corbett and Durfee [24] was used. The x-axis denotes condition and y-axis denotes mean certainty rating.
Perception of scientists’ certainty of vaccine safety
A significant effect was observed, F(3, 315) = 5.452, P < 0.001, η2 = 0.049. A Games–Howell post hoc comparison test revealed that participants in the falsely balanced group reported that scientists were less certain of vaccine safety (M = 4.59, SD = 1.16) than those in the no-link group (M = 5.19, SD = 0.83), P < 0.01, but not the link or control group. In terms of percent, participants in the balanced group reported certainty scores that were on average 12% lower than those in the no-link group (Fig. 2).
Participants’ perception of scientists’ certainty that vaccines are safe. A six-point certainty scale (1 = very uncertain; 6 = very certain) adapted from Corbett and Durfee [24] was used. The x-axis denotes condition and y-axis denotes mean certainty rating.
Behavioral intentions
A significant effect was observed for intention to vaccinate future children, F(3, 315) = 3.023, P < 0.05, η2 = 0.028. Using a Bonferroni post hoc comparison test, participants in the falsely balanced group reported a lower intention (−8%) to have future children vaccinated (M = 4.35, SD = 1.01) than those in the control group (M = 4.79, SD = 0.95), P < 0.05; however, the difference between falsely balanced and no-link; and falsely balanced and link was not significant (Fig. 3). There was no significant difference in intentions to vaccinate oneself among groups.
Participants’ mean intention rating to have their future children receive any vaccine their healthcare provider recommends. A six-point agreement scale (1 = strongly disagree; 6 = strongly agree) was used. The x-axis denotes condition and y-axis denotes mean intention rating.
Discussion
This study suggests that falsely balanced reporting of an alleged autism–vaccine link heightens readers’ uncertainty regarding vaccine safety and lowers their intention to vaccinate their future children. Whereas similar research has found that one-sided, anti-vaccine-only website messages lower individuals’ vaccine intentions [25], our results suggest that two-sided news messages with claims both for/against an autism–vaccine link can have a similar effect. Although we observed a medium effect size for certainty perceptions and a small effect size for child intention, the effect of falsely balanced media coverage could influence enough people to withhold childhood scheduled immunizations, which could increase the risk of a vaccine-preventable disease outbreak. This is important since these outbreaks can originate and spread from a small cluster of unvaccinated people [26]. However, we acknowledge that this article represents a preliminary study and more research should be conducted in this area. The falsely balanced article, moreover, produced a stronger effect than the ‘link-only’ article. This finding may be due to two-sided presentations eliciting a stronger perception that experts (both scientists and medical experts) are divided over the autism–vaccine link than ‘link-only’ presentations. Perceiving a divide may be a driving force in uncertainty perceptions and decreased intention, and we urge more research in this area—specifically, whether including a statement (within a balanced article) that a scientific/medical consensus overwhelmingly rejects an autism–vaccine link increases readers’ certainty of vaccine safety and vaccine intentions.
This study has several limitations. Since participants were undergraduate students, further research could include parents as a sample population of current vaccine decision makers: individuals who arguably have a more vested interest in issues of vaccine safety and thus represent a more generalizable sample. At the same time, we view the present study as the first step in a larger program of research that explores the potential impact of mediated coverage of the autism–vaccine controversy on perceptions of vaccine benefit/safety; views on an autism–vaccine link; immunization-related intentions and other factors. The fact that we still found a meaningful, albeit small, effect among a sample for whom child rearing is arguably a more distant consideration is indicative of an interesting phenomena that requires additional research.
Additionally, the effect of one-shot exposure to a news article might not be as long lasting as cumulative exposure over a longer period of time and, as observed by focus group studies [19, 27], the effect on individuals might instead lead them to seek more information from their physician and/or friends and family members. Longitudinal measurement of participants’ vaccine beliefs, intentions and exposure to media coverage of the autism–vaccine controversy could expand on this issue. We also recognize that participants were not measured for pre-existing certainty perceptions and vaccine intentions. A pre-test/post-test experimental design could assess those measures; however, we reasoned that asking questions about vaccine intentions and beliefs prior to treatment exposure could prime participants’ responses.
An additional limitation concerns our use of single item measures instead of multi-item scales for certainty and intention. A multi-item scale for certainty and intention measures could potentially increase the study’s internal validity, and we encourage future research in this area to explore using multi-item measures.
Overall, this study raises important ethical considerations, in that the type of news reporting about health risk may influence people’s decisions on important health-related action that affects not just themselves, but their offspring and the larger public. Recent data from the Centers for Disease Control and Prevention show a nationwide increase in the non-medical exemption rate for school immunizations from 2005 to 2011 [28]. Even in areas with high vaccination rates, vaccine preventable disease outbreaks can occur among clusters of unvaccinated children who received these exemptions [26]. Understanding why people choose to forgo vaccinations for themselves and their children is a chief concern among public health officials. While this study demonstrates a potential association between falsely balanced reporting of the autism–vaccine controversy and heightened uncertainty of vaccine safety and lowered intentions to vaccinate, we recommend that health officials not only focus on vaccine education but also open up discussions with journalists on the appropriateness of balancing viewpoints with significantly uneven evidentiary support. As a journalistic norm, balance is often used for objective, fair and impartial reporting, as well as highlighting controversial news that sells [29]. We suggest that balance can be achieved without being a false balance if journalists provide complete information that puts the autism–vaccine controversy in the appropriate context—that a scientific and medical consensus overwhelmingly rejects a link between vaccines and autism. Although not explored in this study, previous work suggests that adding context about the strength of evidence between two conflicting views could have several positive outcomes, such as increased certainty about the state of evidence [24] and a stronger view that the scientists involved are trustworthy [30]. Future work addressing this area could help lay the groundwork for improved journalism standards for reporting on health controversies.
- © The Author 2012. Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com
