Age-related disgust responses to signs of disease

ABSTRACT

Previous studies found similarities in adults’ disgust responses to benign (e.g. obesity) and actual disease signs (e.g. influenza). However, limited research has compared visual (i.e. benign and actual) to cognitive (i.e. disease label) disease cues in different age groups. The current study investigated disgust responses across middle childhood (7–9 years), late childhood (10–12 years), adolescence (13–17 years), and adulthood (18+ years). Participants viewed individuals representing a benign visual disease (obese), sick-looking (staphylococcus), sick-label (cold/flu), and healthy condition. Disgust-related outcomes were: (1) avoidance, or contact level with apparel the individual was said to have worn, (2) disgust facial reactions, and (3) a combination of (1) and (2). Avoidance was greater for the sick-looking and sick-label than the healthy and obese conditions. For facial reaction and combination outcomes, middle childhood participants responded with greater disgust to the sick-looking than the healthy condition, while late childhood participants expressed stronger disgust towards the sick-looking and obese conditions than the healthy condition. Adolescents and adults exhibited stronger disgust towards sick-label and sick-looking than obese and healthy conditions. Results suggest visual cues are central to children’s disgust responses whereas adolescents and adult responses considered cognitive cues.

KEYWORDS:

• Disease avoidance
• age-related disgust
• disease dues
• emotion

Disgust has been argued to have evolved as a disease avoidance mechanism (Oaten et al., 2009). However, infectious diseases can have inconsistent presentations, with some diseases being asymptomatic (i.e. no visual cues) and some non-infectious conditions appearing infectious. Consequently, disgust should elicit avoidance of individuals with any visual atypicality that could be interpreted as a sign of disease, such as a port wine stain or an obese individual, and of those labelled with a disease, such as influenza (flu), even in the absence of visual indicators of disease (Oaten et al., 2009, 2011). Although researchers have previously examined the role of disgust in the avoidance of disease cues in adults (Ryan et al., 2012), research in child and adolescent populations is lacking. The current study compares disgust responding to visual signs of disease, in the form of both an individual with a visual atypicality (i.e. obesity – a benign disease cue) and an individual who has visual signs of an infectious disease (i.e. staphylococcus infection), as well as to a disease label (i.e. cold/flu) within middle childhood, late childhood, adolescence, and adulthood.

Disgust

Disgust is a core emotion argued to have evolved to protect organisms from disease threats with its physiological (e.g. nausea), qualia/subjective experience (e.g. revulsion), behavioural (e.g. avoidance), and expressive (e.g. disgust face) components (Oaten et al., 2009). There is also evidence the behavioural components’ avoidance occurs in response to potential and actual infectious individuals (Ryan et al., 2012). Thus, disgust can be viewed as a protective response to objects and people that may pose a disease threat. However, non-disease elicitors, such as moral violations, have also been associated with disgust (e.g. Rozin et al., 2000; Tybur et al., 2009). This suggests that disgust may have evolved as a disease threat avoidance system, with non-disease related disgust elicitors being a preadaptation (e.g. a new system building upon an old system), where the disease avoidance system is utilised for non-disease cues such as moral violation being avoided similarly to an infected individual (Oaten et al., 2009; Rozin et al., 2000).

However, the disgust mechanism is prone to errors. Oaten et al. (2011) argued that the disease avoidance aspect of disgust could represent a mechanism for stigmatisation of atypical individuals. As many contagious diseases have limited obvious cues (e.g. cytomegalovirus), and some non-contagious conditions look contagious (e.g. acne), the disgust system is likely to result in false positives, signalling a risk of infection where one is absent, and false negatives, not responding when a risk is present. Disease labels generate further complexity as a disease label should elicit an avoidance response despite the absence of visual disease signs as a conceptual understanding of the label should produce thoughts of infection and disease (Oaten et al., 2009). Therefore, it is often better for individuals to assume a disease threat even if the evidence is limited or uncertain, as explained by error management theory, which posits that humans and other animals are biased towards making certain errors that favour fitness (Haselton & Buss, 2009). As detailed by Oaten et al. (2011), it is safer to assume individuals possessing a visual atypicality or disease label have an infectious disease and avoid them. This facet of the disgust mechanism is a crucial component for explaining the functional basis of avoidance, since the primary goal of the system is to protect an individual from potential disease threats, while also providing an explanation as to why individuals with a non-infectious visual atypicality will be avoided. The functional interpretation of disgust incorporates both specific cues to disgust but also can operate on a spectrum of “normality”, where deviations from this perceived normal range can elicit disgust, suggesting a potentially flexible and adaptive system responsive to a variety of threats.

Disgust-driven avoidance of individuals who are perceived to pose a disease risk, regardless of actual contagion risk, has been well documented in adults. Research shows that adults treat those with a non-infectious condition based on visual cues involving atypical body form similarly to those with an infectious disease. For instance, Ryan et al. (2012) had participants interact with objects they were informed had been used by healthy controls, individuals who were sick, or individuals who had a port wine stain. Participants showed similar avoidance responses towards the person with the port wine stain and the sick individual displaying infectious disease cues. Relatedly, as obesity has been labelled as a disease in contemporary Western societies (Allison et al., 2008), although non-infectious, it specifically is expected to elicit avoidance similar to an actual disease cue. For example, Tapp et al. (2020) conducted a study where participants reported their induced emotion and willingness for contact when shown images of a person with the disease label “flu”, a “false alarm” (obese person), a negatively valanced, non-disease control (person with facial bruising), and a healthy control. Participants responded with a similar level of avoidance of objects they believed had been handled by the obese individual and the individual with the flu. Both conditions elicited greater avoidance than the healthy control. This aligns with other studies showing pathogen disgust sensitivity, more so than moral disgust, predicted obesity stigma (Lieberman et al., 2012) and that pathogen threat salience is highly associated with negative concerns towards obese individuals (Park et al., 2007). This supports the contention that non-contagious disease cues produce levels of disgust similar to contagious disease cues.

Whilst previous studies, demonstrate that atypical non-contagious disease cues may be treated like a disease, they do not clearly distinguish between visual (e.g. obesity) and cognitive (e.g. disease label) disease signs. For instance, the disease conditions in previous studies (Ryan et al., 2012) have focused on the flu as the disease sign. Although the flu may involve some visual signs of disease (e.g. red nose from sneezing), this is arguably not a strong visual disgust elicitor compared to other diseases (e.g. skin sores from staphylococcus infection) and relies on the conceptual understanding of the flu. As humans rely more heavily on visual cues than conceptual ones when making health judgements (Oaten et al., 2011), a stronger visual disease sign could be used to examine whether a non-infectious visual disease sign (e.g. obesity) is treated similarly to a visual contagious disease sign (e.g. staphylococcus). Furthermore, distinguishing between purely visual disease signs and disease labels (e.g. the flu with no visual disease signs) may enable a better understanding of the link between disgust and disease avoidance, with stronger reliance on, and disgust response to, visual disease cues than cognitive labels anticipated. As well as clarifying this distinction amongst adults, the current study also aimed to add to the limited knowledge regarding the extent to which these different cues might elicit disgust in younger age groups.

Disgust and disease avoidance responses in children

Research examining disgust has been mixed about what point in development this emotion emerges. Although studies demonstrate disgust emerges around 7 years of age (Brown & Harris, 2012), some studies show disgust to be present much earlier than this, by 2–3 years (e.g. Stevenson et al., 2010). For instance, studies show children can match a disgusted facial expression to an appropriate scenario (e.g. seeing rotten food) at around 3 years of age (Widen & Russell, 2013). However, there are criticisms of those observing disgust at earlier ages. For example, even though children can match a product to a facial reaction, could mean they have seen others make that face in similar situations and does not necessarily mean they would have the disgust feeling. Although Stevenson et al. (2010) observed avoidance of disgust objects before the age of 7, Rottman (2014) noted that the children below this age made disgust facial reactions in response to different types of disgust elicitors (e.g. maggots) only 20% of the time, increasing to 40% by 6.8 years. Therefore, although children younger than seven displayed behavioural avoidance in response to the elicitors, this avoidance was often not accompanied by a disgust expression, making it difficult to determine if disgust was playing a role. As avoidance can be due to a variety of factors (e.g. social expectations, fear) this raises some concerns when determining whether disgust is motivating the behaviour. Nonetheless, it appears that at around 7 years of age, disgust is increasingly likely to prompt avoidant behaviour.

Furthermore, the changes in disgust responding may be due to differences in learning regarding what is disgusting. Research indicates that children can develop disgust and fear beliefs towards unfamiliar animals after receiving disgust-related information about them, leading to avoidance behaviours and fear (Muris et al., 2009, 2012). Parents also tend to amplify their vocalizations and facial expressions of disgust when they, alongside their child, come across a source of adult disgust (Muris et al., 2013; Oaten et al., 2014). Even though young children might struggle with facial recognition of disgust (Widen & Russell, 2013), they can still perceive the negative nature of the expression. However, it should be acknowledged there also appears to be a “preparedness” regarding what is found to be disgusting. Evidence from Stevenson et al. (2014) showed children were more attracted to images adults deemed disgusting or scary when accompanied by related sounds than when paired with unrelated negative sounds, suggesting a generic inclination in children to align their attention with combined visual and auditory cues of disgust or fear (Stevenson et al., 2014). Nonetheless, some of the changes to disgust throughout childhood may indicate learnt responding to specific objects. This is particularly relevant for the more complex and abstract cues as disgust responses to these appear later in development (Stevenson et al., 2010).

Although limited research has investigated whether disgust elicitors produce different disgust-related responses at different ages, some patterns have been observed. In Stevenson et al.’s (2010) research, concrete and abstract disgust elicitors were considered, with concrete disgust elicitors driven by visual cues based on the object’s appearance and abstract disgust elicitors driven by meaning-based labels. Results showed that disgust in response to concrete elicitors emerged first, starting at about 2.5 years but getting stronger and more consistent until 6.8 years. By contrast, disgust in response to abstract elicitors emerged later, starting at 6.8 years but appearing more consistently at 10.1 years and more so at 14.3 years. This overall pattern of findings suggests that disgust responses to disgust elicitors that require more conceptual understanding only start to appear later in development, around middle childhood however this understanding is inconsistent at first and then improves throughout the lifespan.

This emergence of disgust in response to more conceptual disgust elicitors is relevant to understanding children’s responses to potentially infectious people. As disgust is proposed to be the emotion associated with a disease avoidance system, a relationship should exist between disease-related knowledge and disgust-related behaviours. Research suggests that children under the age of 7 overgeneralise what is contagious (Bares & Gelman, 2008). Further, children’s understanding of contamination and contagion appears to improve throughout the lifespan, with older children having a better understanding of what is contagious compared to younger children (Bares & Gelman, 2008; DeJesus et al., 2021). This understanding extends to avoidant behaviour. Blacker and LoBue (2016) invited 4- to 7-year-old participants to play with two confederates, one of whom was labelled “sick”. Children avoided proximity to, and contact with, the sick confederate and her toys, but only 6- and 7-year-olds chose to play near the non-sick confederate over the sick confederate at a rate significantly above chance. Although this suggests an improvement in children’s understanding of contamination at around 6 years of age, they remain worse than adolescents and adults at differentiating between contaminants that would likely cause harm (e.g. a cough) and those that would not (e.g. a leaf; Apicella et al., 2018). Relatedly, Rottman et al. (2020) examined individuals’ responses to dirtiness, demonstrating that, regardless of cause of dirtiness, biases towards trusting those who were clean were evident for both adults and children. However, children were more explicit with their bias, with the research demonstrating that children are less concerned about regulating their biases across all three studies. With respect to disease cues, this suggests children should be poor at distinguishing between visual atypicalities that represent benign disease cues, such as obesity, and actual disease cues, such as visual indicators of an infection, perceiving these conditions as equally threatening. By contrast, adolescents and adults should have an improved ability to distinguish between these threats.

Research on children’s responses to benign visual disease cues is limited, with most studies focusing on responses to obese children. For instance, Kornilaki and Cheng (2019) found a negative association between body size and popularity, with neglected and rejected children having the highest BMI. However, adults appear more accepting of obese individuals compared to children, demonstrating less avoidance and negative emotions towards them (Latner et al., 2005). Klaczynski (2008) observed that 7- and 10-year-olds perceived items that had been handled by obese children, compared to non-obese children, more negatively and as more likely to take on negative properties and would avoid interacting with them. This is consistent with the perspective that young children see the object as contaminated by the obese child. However, this study did not explicitly test whether the obese condition cued similar responses to sick-looking or sick-label conditions. Accordingly, the current study builds upon this research by comparing disgust responses of participants of different ages to individuals with a non-contagious visual disease sign to two different disease-related conditions – i.e. a disease label and a visual sign of disease.

The current study

To test disgust in response to different disease conditions within different developmental groups, a mixed-design, experimental study was conducted. Four age groups were included: middle childhood (7–9 years); late childhood (10–12 years); adolescence (13–17 years); and adulthood (18+ years). Participants were shown photographs of individuals representing one of four disease cue conditions: an individual who looked obese (benign visual disease cue); a sick-looking individual (visual infectious disease cue; staphylococcus infection); a healthy-looking individual with a contagious label (disease label; a cold/flu); and a healthy individual. Behavioural responses, in the form of participants’ willingness to have increasing levels of contact with an item (e.g. sock, shirt) described as previously used/worn by the individual in the image, were assessed. Lower levels of willingness to have contact reflected a higher level of avoidance, a behavioural indicator of disgust. Facial responses to the images were also video-recorded, with indicators of disgust coded. It was hypothesised that:

(1) For all age groups, the healthy condition would produce the lowest level of disgust compared to the conditions involving visual cues of disease (obese and sick-looking conditions) and the sick-label condition.

(2) For the middle and late childhood age groups, the conditions involving both benign and infectious visual cues of disease (obese and sick-looking conditions) would produce higher levels of disgust compared to the sick-label condition.

(3) For the adult and adolescent age groups, the condition involving visual cues of an infectious disease (sick-looking condition) would produce greater disgust compared to the conditions involving benign disease cues (obese) and the sick-label condition.

Method

Participants

A total of 120 participants were included in the study, collected from 2017 to the beginning of 2020. Sample size was determined based on previous studies (e.g. Oaten et al., 2014; Stevenson et al., 2010, 2014) and feasibility. Participants aged 7–16 years were recruited using advertisements in school newsletters, letterbox drops, a broadcast email to staff and students at [redacted for blind review] University, and flyers around the University campus. As the study was conducted in-person, the parents of participants 16 years and under were remunerated with a retail gift card for bringing their child to campus. Participants aged 17 years and older were first year psychology students at [redacted for blind review] University and received partial credit for participating. Ethical approval was obtained from the participating University’s Human Ethics Committee.

Participants were divided into four age categories. Groupings were middle childhood (age range = 7 years, 0 months to 9 years, 11 months; M_age = 7.87, SD = 0.83; 10 female and 5 male), late childhood (age range = 10 years, 0 months to 12 years, 11 months; M_age = 10.89, SD = 0.99; 12 female and 7 male), adolescence (age range = 13 years, 0 months to 17 years, 11 months; M_age = 14.96, SD = 1.63; 16 female and 8 male), and adulthood (age range = 18 years, 0 months to 44 years, 11 months; M_age = 21.97, SD = 6.70; 44 female and 18 male). Chi-squared analysis indicated age groups had a similar gender make-up, χ²(3) = 0.48, p = .922. More than 74% of participants reported they were White, with the next highest being Asian at 5% and Hispanic at 4.2%. See supplementary materials for further breakdown.

Materials

Three models were recruited for the sick-looking, sick-label, and healthy conditions. For the obese condition, three photographs were sourced online due to model unavailability. All models were White males, clean-shaven, and wore a blue polo shirt for the photographs.

Procedure

Participants were presented with 16 × 11.5-inch images of a head-and-shoulders shot of the models on a 24-inch screen. In the sick-label condition, participants were told either that the person in the picture had the flu (i.e. adults) or a cold (i.e. children and adolescents) when the picture was taken. Cold was used for children because, although both adults and children see colds and flus as similar, or even identical, children understand colds as an illness but limited understanding of the flu (Sigelman, 2012). Further, adults distinguish between colds and flu but see colds as less serious or even harmless compared to the flu (Garcia et al., 2020), with children perceiving colds as a similarly serious as adults’ perception of the flu (Sigelman, 2012).

Disease cue was a within-subject factor – all participants saw a model for each of the disease-related conditions. Participants saw one picture at a time and each picture was paired with one of four items (i.e. hat, socks, gloves, or shirt). The pairing of each item with each image was counterbalanced, with order presentation of the image/item pairing also counterbalanced. For each item, participants were offered increasing levels of contact, first being asked if they were willing to touch the item, then hold it, then put it against their chest, and, finally, put it on, or put their hand inside it (for case of sock so participants did not have to take off their shoes and socks). When participants indicated they would not interact with the item, the next picture-item pair was presented. Participants’ responses were video-recorded.

Coding

Two coders, blind to the study aims, coded whether participants completed the requested interaction in response to the images/item pairings, as well as participants’ disgust facial responses for all videos. Coders were unaware of the image/item pair the participant was viewing as the camera was angled at the participant rather than the screen with the images. To obtain avoidance behaviour scores, responses were coded for each potential level of contact the participant had with the item. Responses were coded as “N” (0) if a participant refused to engage with the item and “Y” (1) if the participant had contact with the object. Consequently, higher scores indicated more willingness to interact with the item (i.e. lower disgust-related (avoidance) responding). A total avoidance score for each condition was calculated by summing the score for each stage (touch, hold, place on chest, put on) therefore ranged from 0 (i.e. would not touch the item) to 4 (i.e. highest level of interaction).

For facial reaction, coders used Ekmans’ Facial Expressions of Emotions coding system (Young et al., 2002) to establish facial reaction recognition. For the disgust face, AU9, AU10, and AU25 were the primary focuses in the training, however, tongue protrusion was also included as this is considered an additional disgust face. As participants were asked to have increasing levels of contact with the items, coders were required to code facial expressions at each stage of contact (i.e. when touching, holding, putting against their chest, and putting on the item). An “N” (1) was noted for the absence of any facial expression and a “Y” (0) for the presence of a facial expression. Although all facial reactions were coded to maintain blind coding, only the disgust coding was included in the analyses. Scores for each stage were summed with 0 indicating disgust face before item offered to 5 no disgust face. As avoidance and disgust facial expression are two representations of disgust (i.e. behavioural and expressive components, respectively), an overall score was created for when either an avoidant behaviour or a disgust facial reaction was first observed. If either a disgust face or avoidance behaviour occurred at a particular step, it was considered a disgust-related response. If a disgust face occurred in response to the image alone before any interaction was offered, a score of 0 was given. Scores ranged from 0 disgust response at early stage to 5 no disgust response. Cohen’s kappa indicated a high level of consistency across the raters for avoidance (κ = .98, p < .001) and facial reactions (κ = .89, p < .001). See supplementary material for further details.

Analysis

Linear mixed-effects models were run through SPSS version 25, one for each of the dependent variables (avoidance, facial disgust, and combined disgust). The independent variables were age group (i.e. middle childhood, late childhood, adolescence, and adulthood) and disease cue (i.e. obese, sick-looking, sick-label, and healthy). For any significant main effects or interactions, pairwise comparisons with Sidak corrections were examined.

Results

Avoidance

The linear mixed effect model with avoidance as the dependent variable demonstrated a significant main effect for disease cue, F(3, 348) = 9.08, p < .001, but non-significant results for the main effect of age group, F(3, 116) = 2.33, p = .078, and the interaction of age group by disease cue, F(9, 348) = 1.61, p = .112. Pairwise comparisons for the disease cue main effect were run with a Sidak correction. Results indicated significantly greater avoidance in the sick-label condition (M = 1.66, SD = 1.70) compared to both the healthy (M = 0.88, SD = 1.37, p < .001), and obese conditions (M = 0.80, SD = 1.26, p < .001), as well as greater avoidance in the sick-looking (M = 1.35, SD = 1.62) compared to obese condition (M = 0.80, SD = 1.26, p = .049).

Disgust facial reactions

There were five participants who had data that was considered missing for various reasons. Two did not touch any item for any condition with no facial reaction. One in the 7–9-year-old group and one in the 10–12 age group. A further three could not be observed due to being unable to see their faces such as they moved their head out of the line of sight of the camera at that moment. One was an adult in the sick-label condition, another was a 7–9-year-old participant in sick-looking and healthy condition, and finally another 7–9-year-old in the obese, sick-looking and sick-label.

A linear mixed effect model examining disgust facial reaction as the dependent variable showed both the main effect for disease cue, F(3, 338.98) = 21.09, p < .001, and the interaction of age group by disease cue, F(9, 338.37) = 2.67 p < .01, to be significant. However, the main effect for age group was non-significant, F(3, 114.92) = 0.72, p = .518. Pairwise comparisons with a Sidak correction were conducted to compare each disease cue condition within each age group (see Figure 1). For middle childhood, disgust expressions occurred earlier for the sick-looking compared to the healthy condition, p = .048. For late childhood, earlier disgust reactions were observed for the obese, p = .001, and sick-looking conditions, p < .001, compared to the healthy condition. For adolescents, disgust responses were earlier for the sick-label, p = .005, and sick-looking, p = .002, compared to the healthy and obese conditions. Lastly, for adults, earlier onset of disgust was observed for the sick-label, p < .001, and sick-looking conditions, p < .001, compared to the healthy and obese conditions. No other significant differences were reported across age groups.

Figure 1. Disgust score for the facial reaction measure.

Composite disgust score

The linear mixed effect model run using the composite disgust score as the dependent variable revealed a non-significant main effect for age group F(3,116) = 1.38, p = .253, but a significant main effect for disease cue F(3,348) = 23.39, p < .001, as well as a significant interaction of age group by disease cue F(9,348) = 2.76 p = .004. Pairwise comparisons with a Sidak correction were conducted to compare each disease cue condition within each age group (see Figure 2). For middle childhood, participants expressed more disgust for sick-looking than healthy images, p = .028. In late childhood, disgust was greater for the sick-looking and obese conditions than the healthy condition, p < .001 and p = .001, respectively. For adolescents, disgust was higher for the sick-looking and sick-label conditions than the obese condition, p = .004 and p = .006, respectively, as well as the healthy condition, p = .002 and p = .003, respectively. Adults mirrored adolescents, showing more disgust for the sick-label and sick-looking conditions than the healthy and obese conditions, all ps < .001. No other significant differences were observed across the age groups.

Figure 2. Disgust score for the composite outcome measure.

Discussion

This study investigated age-related disgust responses to different disease cues, comparing visual disease cues, both benign and actual, to a disease label and a healthy. We predicted that participants in middle and late childhood would respond with a stronger disgust response towards visual cues of disease, whether real or benign, compared to a sick-label. Further, it was predicted that adolescents and adults would have a stronger disgust response to the sick-looking individual than the sick-label and obese individual. For all age groups, disgust was expected to be lowest in response to the healthy model. The main results of the study were that, first, for the facial reaction and composite disgust outcome measures, children tended to rely on visual cues of disease (i.e. greater disgust to the sick-looking than the healthy condition), while adolescents and adults relied on visual and cognitive cues (i.e. greater disgust to the sick-looking and sick-label than the healthy condition). Second, these age by condition interactions were found for the disgust facial reaction and composite outcome measures, but not for avoidance behaviour alone. Third, except for the late childhood groups’ facial reactions and composite scores, disgust responses to the obese condition did not differ from the healthy condition. As responses to the obese condition were unique it will be discussed separately.

Age-related differences in disgust responses to disease cues

For disgust facial reactions and the composite disgust outcomes, children relied on the actual sign of disease (e.g. sick-looking cues) as this differed from the healthy condition, whereas their responses to the more ambiguous conditions (i.e. disease label) did not differ from either the sick-looking or healthy conditions. Adolescents and adults, however, had the strongest disgust response to the sick-looking and sick-label conditions, both stronger than the healthy condition. Therefore, hypothesis 1 was partially supported as, for the childhood age groups, only the sick-looking condition produced more disgust than the healthy condition. However, the results for adolescents and adults were in line with the hypothesis, with both sick-looking and sick-label conditions producing more disgust than the healthy condition. However, it was also found that the sick-looking and sick-label conditions did not significantly differ in their responses for any age groups, with this pattern of results not supporting hypotheses 2 and 3.

Nonetheless, these results support previous disease knowledge studies as research demonstrated children have an incomplete understanding of diseases compared to adolescents and adults (Apicella et al., 2018; DeJesus et al., 2021). Therefore, the current study’s finding that children’s disgust response to the sick-label condition were similar to a healthy condition, whereas adolescents’ and adults’ disgust response to the sick-label were different to the healthy condition, suggests that children may lack a complete understanding of what the label means while adults perceive the label as a disease threat.

Further, these findings bear some similarity to previous research regarding disgust responses to sick individuals. Previous research has demonstrated both adults (Ryan et al., 2012; Tapp et al., 2020) and children (Blacker & LoBue, 2016) responded with more disgust to a sick than healthy individual in experimental settings avoiding an object previously used by them (Ryan et al., 2012), avoiding proximity to the person themselves (Blacker & LoBue, 2016), and in self-report measures (Tapp et al., 2020). The current study affirms these findings as a sick-looking individual consistently produced one of the strongest disgust responses in all age groups. However, this study contributes to the research literature by showing how a cognitive cue in the form of a label was similar to the healthy condition for the childhood age groups, but adolescents and adults were different in their disgust responding to these conditions.

The study also supports theories that argue for different responses between visual and cognitive cues in relation to disgust as a disease avoidance mechanism. As humans generally rely more on visual cues, they should find visual disease cues more disgusting than cognitive cues but, should still find the cognitive cues disgusting if they understand the meaning of them (Oaten et al., 2011). In the present study, this nuanced prediction was not observed as cognitive cues either produced the same levels of disgust as a visual cue, as in the adolescent and adult groups, or produced equivalent disgust responses to the healthy condition, as in the childhood groups.

Avoidance responses across conditions

One unexpected result of the current study was that there were no age-related differences to the disease cues for the avoidance outcome. Although avoidance was higher for sick-looking and sick-label than the healthy individual, providing some support for a disease avoidance theory of disgust (Oaten et al., 2009), there were no observed age-related differences in this pattern, meaning the results were not consistent with the current hypotheses 2 and 3. It was expected that for all types of disgust responses (i.e. facial reactions, avoidance, and the combined outcome), adolescents and adults would respond more strongly to cognitive cues and children would respond more strongly to visual cues of disease (DeJesus et al., 2021; Oaten et al., 2011). Relatedly, Rottman et al.’s (2020) research revealed that children, but not adults, explicitly displayed biases towards unclean individuals (another visual cue), indicating early establishment of biases towards such cues and age as a potentially relevant factor. Contrary to these findings, however, in the current study, an interaction of age and disease cue was not observed in relation to avoidance.

One possible explanation for why an age-related interaction with the conditions was not observed for the avoidance outcome but was observed for facial reactions and the combined disgust measure, could be related to the controllability of the responses. A facial response may align more with an emotional response, being more instantaneous and harder to control, while the avoidance response may reflect a more controlled behavioural response. As demonstrated by electromyographic studies that presented adult participants with various photos that elicited emotions while requesting them to inhibit their responses (Kappas et al., 2000), facial reactions were still observed. This can be contrasted with regulation of behaviour where studies have demonstrated that even children can regulate their behavioural responses from 7 years of age (Montroy et al., 2016). This might explain why an interaction of age and condition was observed for facial reactions but not for behavioural avoidance.

Disgust response to the benign disease cue

Another observation in the current study was the lack of differences in disgust responses to the benign disease cue condition, specifically between the obese and healthy conditions. The only instance where this difference was found was in the late childhood age group, with this group showing greater disgust facial reactions and combined disgust towards the obese than the healthy model. Therefore, regarding the benign disease cue, only the late childhood age group’s responses were in line with hypothesis 1. The expectation for differences between the obese and healthy conditions was based on previous studies that have shown disgust and disease avoidance to be associated with obesity. For instance, Tapp et al. (2020) had adults rate images of individuals with either influenza, obesity, facial bruising, or a healthy appearance for disgust and willingness for increasing physical proximity (e.g. sitting beside to sexual intercourse). The results showed higher disgust ratings for the obese and sick conditions, and participants were less willing to have intimate physical contact with an obese individual compared to a healthy condition. Similarly, a study with children who were offered a drink made by either an obese or non-obese child demonstrated children treated the drink made by the obese child as contaminated (Klaczynski, 2008).

Consequently, this study’s lack of differences in disgust responses to the healthy and obese models has important implications for a model of stigmatisation that argues disgust is the underlying mechanism for some forms of stigmatisation (Oaten et al., 2011). This model posits that the avoidance of individuals with visual atypicalities may be due to the triggering of a disgust mechanism designed for disease avoidance. It is argued that a false alarm is occurring related to avoiding disease threats. Therefore, all age groups should have a greater disgust response to the obese condition than the healthy one. This effect should be stronger with children, as their disease knowledge is incomplete and will often overgeneralise (Bares & Gelman, 2008; DeJesus et al., 2021). This was only observed in the late childhood age group and so does not explain responses to the obese condition in the middle childhood, adolescent, and adult age groups.

The findings of the current study, and the differences observed between them and previous results, therefore have implications for understanding why certain visual atypicalities might trigger a disgust response. In particular, the results suggest that disgust may not be a normative response to obesity but, rather, that learning might have a role to play. Previous researchers have acknowledged that disgust responses have a learning component (Muris et al., 2009, 2012; Oaten et al., 2014). As, increasingly, there have been ongoing attempts to reduce weight stigma in society (e.g. Selensky & Carels, 2021), as well as a shift away from framing obesity as a disease (Penney & Kirk, 2015). It may be that, compared to earlier work in the field, the results of the current study reflect a reduction in the likelihood of learning to equate obesity with disgust. Relatedly, these changing attitudes towards obesity would be particularly relevant if considered in the light of disgust responses that are due to moral disgust. As Rozin (1999) suggests, there is a link between disgust and moral values, with the internalisation of a behaviour being considered “immoral” harnesses the disgust system to reject that behaviour (e.g. moral vegetarians find eating meat disgusting compared to health vegetarians). With societal shifts away from weight stigma, obesity may be less likely to be viewed as immoral, resulting in few differences between the obese and healthy conditions across ages. While a learning explanation does seem relevant to the findings of the current study, it is still important to note, however, that other recent studies have found a link between obesity and disgust (e.g. Tapp et al., 2020). Consequently, further research is needed to replicate the findings of the current study and investigate the mechanisms that underlie responses to obesity.

Strengths, limitations, and future research directions

The current study has several strengths. First, the presentation of stimuli was counterbalanced for order of presentation and different photographic models. As multiple models were included, any effect of the individual person was controlled for and changing presentation order reduced recency or primacy effects. Second, disgust behaviour scores were coded by coders blind to the aims of the study, thus minimising experimenter biases.

However, limitations of the study also need to be considered. First, the current study’s sample may not be generalisable to various cultures. Most participants stated their ethnicity as White, with other ethnicities making up only 15.43% of the sample (see supplementary material). As disgust can have a moral or social component (e.g. Oaten et al., 2011), and moral or societal views about disease can differ across cultures, future research with more culturally diverse samples is warranted. Relatedly, most of the adult sample was between the ages of 18 and 29 and so may not reflect older adults’ disgust responses to disease cues. Most of the participants were female and, ass disgust sensitivity is higher in females (e.g. Curtis et al., 2004), the results may not generalise to males. Future studies should aim to recruit a sample with a more even gender distribution.

Finally, the study may have been limited by its use of obesity as a benign disease cue. Obesity was the focus here, given the more extensively documented link to disgust and disease avoidance in different age groups (e.g. Klaczynski, 2008; Tapp et al., 2020) compared to other benign conditions (e.g. facial disfigurement). Therefore, the current study’s aim was not to amplify weight stigma but rather test a theory that relates to stigmatisation (i.e. Oaten et al., 2009, 2011). However, future research using an alternative (or additional) benign disease cue (e.g. a facial disfigurement) could help determine whether responses to the benign cue in the current study are due to the specific condition (i.e. obesity) or benign disease cues universally.

Conclusion

The study aimed to examine responses to actual and benign disease cues and investigate potential age-related differences. Middle to late childhood participants relied more on visual disease cues, while adolescents and adults incorporated cognitive knowledge about disease as demonstrated by the facial reactions and composite scores. Stronger avoidance behaviours were observed for actual disease threats, suggesting that the perceived realness of the disease cue influences avoidance. Interestingly, no age-related differences in avoidance were observed, indicating that it may not solely reflect feelings of disgust but rather a controllable response. These findings have important implications for understanding the triggers of the disgust mechanism in children, adolescents, and adults regarding disease threats.

Informed consent

Informed consent was obtained from all participants included in the study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data is not available due to ethical constraints.

Additional information

Funding

This work was supported by Australian Research Council: [Grant Number DP150101166] and Australian Government Research Training Program Scholarship.