No Evidence That Omission and Confirmation Biases Affect the Perception and Recall of Vaccine- Related Information
Total Page:16
File Type:pdf, Size:1020Kb
1 NO EVIDENCE THAT OMISSION AND 2 CONFIRMATION BIASES AFFECT THE 3 PERCEPTION AND RECALL OF VACCINE- 4 RELATED INFORMATION 5 6 This is a preliminary draft 7 8 Ángel V. Jiménez 1 2 3 *, Alex Mesoudi 3 & Jamshid J. Tehrani 1 2 9 1 Durham University, UK, Centre for the Coevolution of Biology and Culture, Department of 10 Anthropology. 11 2 Durham University, UK, Conspiracy Theories in Health Special Interest Group, Wolfson Research 12 Institute for Health and Wellbeing. 13 3 Human Behaviour and Cultural Evolution Group, Department of Biosciences, University of Exeter, 14 UK 15 *Correspondence should be addressed to Ángel V. Jiménez: [email protected] 16 ABSTRACT 17 Despite the spectacular success of vaccines in preventing infectious diseases, fears 18 about their safety and other anti-vaccination claims are widespread. To better understand 19 how such fears and claims persist and spread, we must understand how they are perceived 20 and recalled. One influence on the perception and recall of vaccination-related information 21 might be universal cognitive biases acting against vaccination. An omission bias describes 22 the tendency to perceive as worse, and recall better, bad outcomes resulting from 23 commissions (e.g. vaccine side effects) compared to the same bad outcomes resulting from 24 omissions (e.g. symptoms of vaccine preventable diseases). Another factor influencing the 25 perception and recall of vaccination-related information might be people’s attitudes 26 towards vaccines. A confirmation bias would mean that pre-existing pro-vaccination 27 attitudes positively predict perceptions of severity and recall of symptoms of vaccine 28 preventable diseases and negatively predict perceptions of severity and recall of vaccine 29 side effects. To test for these hypothesized biases, 202 female participants aged 18-60 30 (M=38.15, SD=10.37) completed an online experiment with a between-subjects 31 experimental design. Participants imagined that they had a 1-year old child who suffered 32 from either vaccine side effects (Commission Condition) or symptoms of a vaccine- 33 preventable disease (Omission Condition). They then rated a list of symptoms/side effects 34 for their perceived severity on a 7-point Likert scale. Finally, they completed a surprise recall 35 test in which they recalled the symptoms/side effects previously rated. An additional scale 36 was used to measure their attitudes towards vaccines. Contrary to the hypotheses, 37 perceptions of severity and the recall of symptoms/side effects were not associated with 38 experimental condition, failing to support the omission bias, nor did they interact with 39 attitudes towards vaccines, failing to support the confirmation bias. This raises the 40 possibility that the spread of anti-vaccination claims might not be explained by these 41 particular universal cognitive biases. 42 KEY WORDS 43 Omission bias, Confirmation Bias, Recall Bias, Antivaxx, Cultural Evolution, Evolutionary Psychology. 44 1.- INTRODUCTION 45 Ever since vaccination was developed in the late 18th century, it has been 46 accompanied by scepticism and opposition. Early resistance to the first vaccine, against 47 smallpox, was motivated by fears towards a practice that conflicted with folk intuitions and 48 religious beliefs (Arnold, 1993; Baldwin, 1999; Durbach, 2005). Smallpox vaccination 49 entailed the introduction of an alien substance into the human bloodstream, which was 50 seen as a violation of bodily purity and as the cause of, rather than the cure for, diseases 51 (Durbach, 2005). Moreover, the mechanism by which vaccination against smallpox worked 52 was counterintuitive as it implied that the intentional infection with one disease (cowpox) 53 could prevent the future development of a related but different disease (smallpox) 54 (Baldwin, 1999). Furthermore, smallpox vaccination involved the mixture of a substance 55 coming from a cow with human blood, which was considered anti-natural and anti-Christian 56 in Europe (Baldwin, 1999) and was also disturbing for Hindus in colonial India (Arnold, 57 1993). 58 More recently, specific vaccines (e.g. the MMR vaccine) have been the targets of 59 opposition and have often been falsely linked to idiopathic medical conditions such as 60 autism, neurological disorders and allergies (Burgess, Burgess, & Leask, 2006; Leask, 61 Chapman, & Robbins, 2010; Offit, 2008). Anti-vaccination messages recurrently emerge and 62 spread through the mass media (Larson, Wilson, Hanley, Parys, & Paterson, 2014; Leask & 63 Chapman, 1998), social media (Salathé & Khandelwal, 2011) and on anti-vaccination 64 websites (Bean, 2011; Davies, Chapman, & Leask, 2002; Kata, 2010; Wolfe, Sharp, & Lipsky, 65 2002; Zimmerman et al., 2005). This spread of anti-vaccination messages affects people’s 66 decisions regarding whether to vaccinate or not (Betsch, Renkewitz, Betsch, & Ulshöfer, 67 2010; Jolley & Douglas, 2014). As a consequence of the erosion of herd immunity and the 68 formation of clusters of unvaccinated individuals (Salathé & Bonhoeffer, 2008), outbreaks of 69 vaccine-preventable infectious diseases still occur in developed countries, sometimes 70 leading to fatal consequences. 71 The recurrence and potency of social fears about the safety of vaccines and the 72 widespread diffusion of anti-vaccination information, despite the spectacular contribution of 73 vaccines in preventing diseases, calls for a scientific explanation. Miton and Mercier (2015) 74 recently argued that two characteristics of human cognition - disgust towards potential 75 contaminants and an omission bias - make vaccination counterintuitive. This counter- 76 intuitiveness makes anti-vaccination messages “cognitively attractive” and likely to spread 77 (see Boudry, Blancke, & Pigliucci, 2014 for a general discussion on how scientific and 78 pseudoscientific ideas spread). In the present article we empirically test the role of one of 79 these cognitive factors, the omission bias, i.e. the tendency to consider bad outcomes 80 resulting from a commission (e.g. side effects of a vaccine) as worse that the same bad 81 outcomes resulting from an omission (e.g. symptoms of a vaccine-preventable disease). 82 While there is some evidence for the existence of omission bias in the context of vaccination 83 (see next section), our study is novel in extending the scope of application of the omission 84 bias to perceptions of severity of vaccine-related symptoms and side effects and their recall. 85 Consequently, we predict that people should rate as more severe, and retain longer in 86 memory, vaccine side effects over symptoms of a vaccine-preventable disease, where the 87 symptoms and the side effects are otherwise identical. These perceptual and memory 88 effects relating to vaccine-related information (symptoms and side effects) would play a key 89 role in the retention and transmission of anti-vaccination messages and opinions. Our study 90 is one of the first experiments exploring the effects of cognitive biases in the recall of 91 vaccine-related information (see also Altay & Mercier, 2018; Jiménez, Stubbersfield, & 92 Tehrani, 2018). 93 1.1.- The Omission Bias in Vaccine Decisions 94 A substantial body of literature maintains that humans have a general tendency to 95 consider the bad outcomes resulting from an action (commission) as worse than the same 96 bad outcomes resulting from a lack of action (omission), even when the bad outcomes 97 resulting from an omission affect a greater number of people or have a higher probability of 98 occurrence (e.g. Spranca, Minsk, & Baron, 1991). The basic paradigm for studying the 99 omission bias applied to vaccine decisions was developed by Ritov and Baron (1990) . 100 Participants are prompted to imagine a scenario in which an infectious disease kills 10 out of 101 10,000 non-vaccinated children but a new vaccine confers immunity against this disease. 102 However, the side effects of this vaccine can result in death with a certain probability. 103 Participants have to decide whether or not to vaccinate their child when the probability of 104 death by the vaccine side effects is 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 out of 10,000 vaccinated 105 children. The omission bias is detected when participants prefer not to vaccinate when the 106 probability of dying from vaccine side effects is lower than the probability of dying from the 107 vaccine-preventable infectious disease. Multiple experiments have found that many people 108 prefer not to vaccinate themselves or their children when the risks of suffering from a 109 vaccine-preventable disease are higher than the risks of suffering from a vaccine reaction, 110 where the severity of the disease and the vaccine reaction are equal (Asch et al., 1994; 111 Baron, 1992; Baron & Ritov, 1994, 2004; Connolly & Reb, 2003; DiBonaventura & Chapman, 112 2008; Meszaros et al., 1996; Ritov & Baron, 1990; Wroe, Turner, & Salkovskis, 2004). 113 Importantly , the omission bias in hypothetical vaccine decisions positively predicts actual 114 vaccine behaviour in both retrospective (Asch et al., 1994; Meszaros et al., 1996; Wroe et 115 al., 2004) and prospective studies (DiBonaventura & Chapman, 2008). 116 One of the cognitive mechanisms that might explain the omission bias is anticipated 117 regret. Some authors (e.g. Asch et al., 1994) have suggested that bad outcomes resulting 118 from actions elicit considerably more regret than the same bad outcomes resulting from a 119 lack of action. The higher level of regret after bad outcomes