Introduction
Individuals' conceptions about older people, old age and their own ageing are always embedded in historical, cultural and social contexts (Kornadt et al., Reference Kornadt, Kessler, Wurm, Bowen, Gabrian and Klusmann2019). While the specific features ascribed to older adults change over time, a recent study (Ng et al., Reference Ng, Chow and Yang2021) reveals that societal narratives surrounding ageing have become more negative over the last centuries across multiple media genres. Of the institutions reflecting and producing narratives on ageing, mass media remains the most important (Rozanova et al., Reference Rozanova, Northcott and McDaniel2006; Loos and Ivan, Reference Loos, Ivan, Ayalon and Tesch-Römer2018). The media have a gatekeeping role able to actively shape public reality (Evans, Reference Evans1997). When stereotypical representations of older adults are repeatedly selected by the media, they contribute to a normalisation of stereotypes, potentially causing negative effects on people's wellbeing (Levy, Reference Levy2009; Meisner and Levy, Reference Meisner, Levy, Bengtson and Settersten2016).
During the COVID-19 pandemic, media stories shone a spotlight on old age and older adults in the face of their increased risk of severe illness, hospitalisation and death (e.g. Centers for Disease Control and Prevention, 2021). The media also focused on measures designed to contain the disease in the older population, such as bans on visitors to long-term care facilities (Martins Van Jaarsveld, Reference Martins Van Jaarsveld2020). The literature provides preliminary evidence that this increased coverage in the mass media was accompanied by a (re-)emergence of negative stereotypes about older age (Bravo-Segal and Villar, Reference Bravo-Segal and Villar2020; Allen and Ayalon, Reference Allen and Ayalon2021; Allen et al., Reference Allen, Odziemczyk, Perek-Białas and Ayalon2021; Köttl et al., Reference Köttl, Tatzer and Ayalon2022; Lagacé et al., Reference Lagacé, Doucet, Dangoisse and Bergeron2021; Morgan et al., Reference Morgan, Wiles, Williams and Gott2021; Zhang and Liu, Reference Zhang and Liu2021; Yang et al., Reference Yang, Au, Wong and Louin press). However, these studies primarily relied on published written materials. Despite the power of visual imagery to shape individual and societal views on ageing, to date, no study has systematically investigated the visual representation of older adults in the media. Specifically, images are the primary access point to published texts and therefore have a significant influence on how topics are framed in the media (Garcia and Stark, Reference Garcia and Stark1991). According to framing theory (Entman, Reference Entman1993), journalists highlight certain aspects of the issues they are covering, thereby promoting a particular interpretation. Thus, the media framing of a global crisis such as the COVID-19 pandemic may have a significant impact on the public's responses to and understanding of the pandemic. Visual content is processed faster, demands a smaller cognitive load and is easier to recall than the written content of articles (Dan, Reference Dan2018). Thus, in overwhelming times (such as the COVID-19 pandemic), visual frames exert a particular power. Further, due to the unconscious processing of visual content, stereotypes – including those about age – are less obvious than when they appear in verbal content (Messaris and Abraham, Reference Messaris, Abraham, Reese, Gandy and Grant2001).
With this in mind, this study is the first to examine how older people were portrayed visually in the German mass media during the first year of the COVID-19 pandemic. Specifically, a quantitative content analysis of cover photos on the four most popular online news sites in Germany published during the first phase of the pandemic was conducted to understand the visual portrayal of older adults. Visual content analysis is a popular method for examining the visual imagery of older people in the media (e.g. Simcock and Linn, Reference Simcock and Lynn2006; Roy and Harwood, Reference Roy and Harwood1997; Kessler et al., Reference Kessler, Schwender and Bowen2010; Kessler and Schwender, Reference Kessler and Schwender2012; Xu, Reference Xu2022). Online media publications were chosen as they have gained enormous significance in recent years, particularly during the pandemic, as demonstrated by the rise in Germany of subscriptions to the digital editions of news magazines such as Der Spiegel and Bild (Newman et al., Reference Newman, Fletcher, Schulz, Andi, Robertson and Nielsen2021). Germany has one of the oldest populations in the world; in Germany, the highest percentage of COVID-19-related deaths occurred among individuals aged 80+ (Federal Statistical Office of Germany, 2021). The pandemic itself may have also directly impacted journalistic practices regarding older people: while interest in visually mapping the lived reality of life during the COVID-19 pandemic may have increased, measures such as physical distancing concomitantly hampered photographing people, thus limiting the image material available for news stories.
The portrayal of older adults in the German media is in line with its Western counterparts: they are hardly visible, especially women and the oldest-old (Röhr-Sendlmeier and Ueing, Reference Röhr-Sendlmeier and Ueing2004; Hoppe et al., Reference Hoppe, Tischer, Philippsen and Hartmann-Tews2016; Myrczik et al., Reference Myrczik, Bowen, Franke, Taeuber and Kessler2022), while one-sided negative images (Ueltzhöffer, Reference Ueltzhöffer1992; Femers, Reference Femers2007; Wangler, Reference Wangler, Schwender, Hoffmann and Reißmann2013) are increasingly complemented by extraordinarily positive portrayals in advertisements and television (Kessler et al., Reference Kessler, Rakoczy and Staudinger2004; Röhr-Sendlmeier and Ueing, Reference Röhr-Sendlmeier and Ueing2004; Kessler et al., Reference Kessler, Schwender and Bowen2010; Wangler, Reference Wangler2012; Hoppe et al., Reference Hoppe, Tischer, Philippsen and Hartmann-Tews2016), even among those individuals experiencing dementia (Kessler and Schwender, Reference Kessler and Schwender2012). The period between 1 January 2020 and 31 January 2021 was chosen for this study, to coincide with the beginnings of global coverage of the pandemic. One year later, infection numbers were decreasing after the spikes in mid-December 2020, which coincided with the second wave of the COVID-19 pandemic in Germany (Schilling et al., Reference Schilling, Tolksdorf, Marquis, Faber, Pfoch, Buda, Haas, Schuler, Altmann, Grote and Diercke2021).
Gerontological research on media representations of older adults (before and during the pandemic)
The empirical literature on media representations in magazines, newspapers, television series and advertising (produced before the COVID-19 pandemic) has consistently shown that older people, especially women and the oldest-old, are heavily underrepresented (Zhang et al., Reference Zhang, Harwood, Williams, Ylänne-McEwean, Wadleigh and Thimm2006; Rozanova, Reference Rozanova2010; Bai, Reference Bai2014; Ylänne, Reference Ylänne, Twigg and Martin2015). Demographic ageing has typically been portrayed as a major societal threat (‘grey tsunami’) (O'Neil and Haydon, Reference O'Neil and Haydon2015), along with frailty, dependency and loneliness (Loos and Ivan, Reference Loos, Ivan, Ayalon and Tesch-Römer2018). In the case of older individuals in television commercials and advertisements, most studies have consistently characterised their portrayal as active, happy, strong and well-integrated (Swayne and Greco, Reference Swayne and Greco1987; Featherstone and Hepworth, Reference Featherstone, Hepworth, Featherstone and Wernick1995; Roy and Harwood, Reference Roy and Harwood1997; Kessler et al., Reference Kessler, Schwender and Bowen2010; Williams et al., Reference Williams, Wadleigh and Ylänne2010), and cognitively alert (Kessler et al., Reference Kessler, Rakoczy and Staudinger2004). There are also gendered aspects of ageing in the media wherein positive attitudes are more often ascribed to individuals read as male than female (Hurd et al., Reference Hurd, Bennett and Liu2014; Kovács et al., Reference Kovács, Aczél and Bokor2021), who are, in turn, depicted as vulnerable (Kessler et al., Reference Kessler, Rakoczy and Staudinger2004; Wells, Reference Wells2007; Kessler and Schwender, Reference Kessler and Schwender2012; Pritchard and Whiting, Reference Pritchard and Whiting2015). While raising awareness of the capabilities and agency of older people, an overly positive representation tends to disguise both age-related challenges and the heterogeneity of the age group as a whole (Ylänne, Reference Ylänne, Twigg and Martin2015; Markov and Yoon, Reference Markov and Yoon2021).
Internationally, many studies have investigated the portrayal of older adults in the media during the COVID-19 pandemic with a primary focus on verbal materials. Analyses have found evidence of their portrayal as frail, vulnerable and dependent (Bravo-Segal and Villar, Reference Bravo-Segal and Villar2020; Allen and Ayalon, Reference Allen and Ayalon2021; Reference Yang, Au, Wong and LouYang et al., in press), being depressed, alone, living in care homes, as passive non-contributors to society (Bravo-Segal and Villar, Reference Bravo-Segal and Villar2020; Allen and Ayalon, Reference Allen and Ayalon2021; Allen et al., Reference Allen, Odziemczyk, Perek-Białas and Ayalon2021; Lagacé et al., Reference Lagacé, Doucet, Dangoisse and Bergeron2021; Morgan et al., Reference Morgan, Wiles, Williams and Gott2021; Zhang and Liu, Reference Zhang and Liu2021), being isolated in their homes (Allen et al., Reference Allen, Odziemczyk, Perek-Białas and Ayalon2021) and cognitively impaired (Köttl et al., Reference Köttl, Tatzer and Ayalon2022). Next to these overall homogenously loss-oriented representations, one study also found a number of nuanced portrayals of older adults, as well as overly positive portrayals of young-old healthy contributors to society (Köttl et al., Reference Köttl, Tatzer and Ayalon2022). To the best of our knowledge, the only study of visual content pertinent to the media representation of older adults during the COVID-19 pandemic is a study on the visual presentation of social groups in Finnish newspapers during its early stages, from January to August 2020 (Martikainen and Sakki, Reference Martikainen and Sakki2021). Portrayals differed between age groups, with older adults presented exclusively as faceless, lonely and isolated behind windows or on balconies.
Visual imagery of older adults from the perspective of communication studies
Photographs are not a reproduction of reality but composed by the photographer, who presents a particular point-of-view to the viewer (Burgin, Reference Burgin1982). Given that images are mostly processed unconsciously, the photographer's choices in terms of composition are rendered almost invisible to the viewer. Nonetheless, variations in camera distance or focus are unconsciously unpacked and assigned similar meanings by different people (Coleman, Reference Coleman, D'Angelo and Kuypers2010). Variations in camera distance affect how close the person depicted is portrayed socially as well as spatially (Dan, Reference Dan2018: 20). Long shots (i.e. whole body) indicate social distance as viewers are unable to discern the individual features of the person photographed; medium shots (i.e. upper half of the body) are neutral in terms of social distance, while closer distances (head and shoulders) and close-ups convey a greater level of intimacy to the viewer (Grabe and Bucy, Reference Grabe and Bucy2009). Faces, in particular, are perceived most positively by audiences, with greater agency and competence (Archer et al., Reference Archer, Iritani, Kimes and Barrios1983; Konrath and Schwarz, Reference Konrath and Schwarz2007), as well as a willingness to act (Slovic, Reference Slovic2007) ascribed to the individuals in the images. Further, different points of view convey different levels of social engagement to the person depicted. While a frontal perspective engages the viewer, the back view of a body can lead to greater detachment between viewer and subject (Kress and van Leeuwen, Reference Kress and van Leeuwen1996).
Moreover, certain bodily characteristics are often interpreted in stereotypical ways (Hall, Reference Hall1976). Specifically, the more stereotypes present in a person's individual features, the more that person is read as a specific type (Jewitt and Oyama, Reference Jewitt, Oyama and Hughes2012). When detailed shots focus on specific body parts, they can be considered a pars pro toto for the person themselves. Thus, the symbolism ascribed to a respective body part is then ascribed to the entire person. In a recent study, older adults were found to be primarily portrayed visually through their hands or legs and feet (Martikainen and Sakki, Reference Martikainen and Sakki2021), which prominently featured the cues of old age – wrinkles, age spots, or restrictions in autonomy and mobility. Nonetheless, a detail shot of an upper arm receiving a vaccine might be drawing attention to a major symbol of the pandemic – vaccinations – rather than being an old age clue.
Current study
Drawing on the most popular German online news sites (i.e. bild.de, t-online.de, n-tv.de and spiegel.de), this study is the first to examine how older adults were portrayed visually in the media during the first phase of the COVID-19 pandemic in Germany (from the beginning of global media coverage of the pandemic in January 2020 to the end of the country's second pandemic wave one year later). By applying visual quantitative content analysis, we assessed and integrated indicators from both gerontology and communication studies to generate a thorough description of the visual representation of older people during this exceptional historical moment. To do so, we examined the frequency of a diverse set of characteristics of older individuals in the N = 604 cover photos of articles identified through the keywords ‘older people’ and ‘COVID-19 pandemic’ (as well as synonyms) published in the four major German online news sites between 1 January 2020 and 31 January 2021.
We expected the visual representation of older adults in our sample to match the results of analyses of verbal materials during the COVID-19 pandemic. Specifically, integrating the literature described above from gerontology and communication science, we derived the following hypotheses:
• Hypothesis 1: Given the generally one-sided, negative stereotypical depictions in verbal materials, we expected to find a high frequency of portrayals of high social distance and detachment between older individuals and the viewer, i.e. visual images that included the back of the body or the whole body; further, we expected a low frequency of visible faces and a high frequency of individual body parts of older people.
• Hypothesis 2: Additionally, we expected to find a high frequency of portrayals of vulnerable, sad, cognitively inactive, passive, frail and care-dependent older individuals in a care setting or at home.
• Hypothesis 3: In alignment with this hypothesis and contrary to earlier findings of the drastic underrepresentation of older women, we further expected to see a high number of older individuals passing as female or androgynous, emphasising the vulnerability of older people.
• Hypothesis 4: Lastly, given the visual segregation of older adults in the COVID-19 pandemic, we expected a high number of images that emphasised the physical distance between the viewer and any older individuals depicted to denote their isolation and thereby strengthen the focus on the vulnerable status of older adults.
Methods
Sample procedure
Data were collected by web scraping all articles published on Germany's most clicked online news sites, namely bild.de (427 million clicks), t-online.de (392 million clicks), n-tv.de (215 million clicks) and spiegel.de (245 million clicks) (Informationsgemeinschaft zur Feststellung der Verbreitung von Werbeträgern, 2020) in the period between 1 January 2020 and 31 January 2021. Each article's cover photo was identified either by being the only one, or in the case of multiple photographs, by having the highest resolution, as identified by the web scrape (i.e. the most central photograph). The sample also included images of older individuals in articles behind the paywall of those outlets that do not provide all articles for free to readers: Spiegel.de and bild.de. Although primarily chosen for their reach, the websites in question also offer some variety. While bild.de is the online outlet of the tabloid Bild, which has the highest readership in Germany, spiegel.de is the online outlet of the left-centre news magazine Der Spiegel. The two other outlets provide a mix of news and entertainment: t-online.de is a news outlet published by the multi-channel media agency Ströer and n-tv.de is owned by the media conglomerate Bertelsmann.
Generating the sample of articles
The articles were filtered with keywords specific to older adults AND the COVID-19 pandemic: (a) ‘ältere’ [older], ‘ältere Menschen’ [older people], ‘hochbetagt’ [aged], ‘hochaltrige’ [very old], ‘Pflegeheimbewohner’ [nursing home resident(s)], ‘Senioren’ [seniors], ‘Seniorenheim’ [old age home], ‘Altenheimbewohner’ [retirement home residents], ‘Pflegeheim’ [nursing home], ‘Altenheim’ [retirement home], ‘Großeltern’ [grandparents], ‘Großmutter’ [grandmother], ‘Großvater’ [grandfather], ‘Rentner’ [retirees], AND (b) ‘Corona’ [corona], ‘COVID-19’ [COVID-19], ‘Pandemie’ [pandemic] in central parts of the article headline OR subline OR skyline OR image sub OR intro, resulting in N = 3,560.
Selecting the sample of cover photos from the sample of articles
From the total number of N = 3,560 cover photos from this sample, two independent judges (female, 22 years; male, 65 years) selected the cover photos that depicted at least one older individual, 65 years or older. In a pilot study, differentiating within this age group into ‘young olds’ (65–80 years) and ‘old olds’ (80+ years), as done in some previous studies (e.g. Kessler et al., Reference Kessler, Schwender and Bowen2010), proved to be too difficult to judge.
Selecting the sample of older individuals from the sample of cover photos
As photographs can show more than one person, we instructed judges to base their assessment on one index individual as indicated in the photographs (defined as being closest to camera if identifiable, otherwise, the person taking up the most space) that the second author had identified in advance. Photographs of older celebrities (N = 334) were excluded. In order to be categorised as a celebrity, the judges needed to identify the character's name and whether this person had a Wikipedia entry, since they might be primarily perceived by other characteristics than their chronological age. The interrater reliability on both coding categories (i.e. index individual and celebrity) was almost perfect: Cohen's κ = 0.84. The raters discussed differences in their coding until they reached a consensus. This procedure resulted in N = 604 photographs, including ones that appeared multiple times, in order to cover the quantity of depictions of older adults.
A total of 53 photographs (8.8%) appeared multiple times, accounting for 44.7 per cent (N = 270) of the total sample. The three photographs used most often were used 46 times (Figure 1), 27 times (Figure 2) and 24 times (Figure 3). Most photographs came from a newswire database (79.0%; N = 477), far fewer from individual photographers (13.1%; N = 79) or other sources including the news site's photo archive (N = 38). (It should be noted that no information was available on whether the image was taken before or during the COVID-19 pandemic.)
The most used photograph in the sample.
Source: Picture Alliance/dpa/Oliver Berg.
The second most used photograph in the sample.
Source: Picture Alliance/dpa/Christoph Schmidt.
The third most used photograph in the sample.
Source: Picture Alliance/dpa/Tom Weller.
Indicator coding procedure (visual content analysis)
Six trained and independent judges (all female psychology students, aged between 24 and 52; median = 25.5 years) rated the final sample of N = 604 individuals in the cover photos with indicators from gerontology and communication studies (each pair of judges coded a separate set of indicators). Each judge obtained all photographs in high resolution on a separate device, and a coding sheet. We could not rule out that the age and gender of the judges influenced their assessment of the characteristics of the older individuals and therefore decided to employ judges from the same age and gender group with the exception of the group with the 52-year-old judge. However, studies on age stereotypes (e.g. Hummert et al., Reference Hummert, Garstka, Shaner and Strahm1994) have shown that young, middle-aged and older adults, as well as women and men, basically share similar representations of old age. We thus believe that ratings from older and male judges will not fundamentally differ from that of our judges. Each pair of judges separately rated a set of indicators and discussed the differences in ratings until they reached a consensus. To avoid stimulating the raters' subjective age biases, they were unaware of the hypotheses of the study.
For the coding of the indicators, the eight judges were trained to apply stipulated definitions for each indicator and were given examples of idealised types for each indicator. The training was based on a manual and lasted six hours. Judges practised rating cover photos not contained in the study until interrater consistency was established. Reliabilities ranged from Cohen's κ = 0.65 to 0.99. Table 1 displays the reliabilities for all measures.
Reliabilities
Measures
Table 2 gives an overview of the dimensions and categories.
Frequency distribution across coding categories
Notes: 1. N = 604. 2. N = 447 (without individuals depicted ‘alone’).
Presentations of the older character's body
The judges coded whether the older individual's body was presented as total body (head to toe from the front), primarily hands, primarily vaccination arm, primarily back of body (some part of the body or total body from the back view), primarily head/shoulders, primarily upper body half (from hip upwards) or primarily legs/feet.
Visibility of the older individual's face
The judges coded the full visibility of face, partial visibility of face or no visibility of face. Full visibility of face encompassed the face including nose, eyes, mouth and forehead fully visible. Partial visibility of face included visible faces that were partly covered or partly visible so that not all elements (nose, mouth, eyes and forehead) were visible. No information was coded if the face was not visible (view from the back or no portions of the head visible).
Setting
The judges coded whether an older individual was depicted in a home setting, care setting, medical setting, public setting, testing/vaccination setting or other setting. Home setting was coded when private living spaces in the home or in a private garden were visible (highly individualised and/or decorated). Care setting included scenes in or outside nursing homes (e.g. rooms or hallways, little decoration, standardised furniture, possible care-givers visible). Medical setting was dominated by medical equipment (e.g. clean rooms, hospital beds, medical professionals), yet also outside the hospital (hospital building and/or park clearly identifiable due to signage). Public setting included open-air settings (parks, lakes, etc.) as well as spaces accessible to the public (supermarket, restaurant, public transportation) or public buildings (police, courthouse). Testing/vaccination setting included all testing or vaccination scenes, possibly showing an injection in the upper arm, or a swab. Other settings did not fit one of the settings described above and included occupational (offices) or recreational indoor settings (musical lessons, painting, etc.). No information was coded if a setting was not clearly identifiable (blurry background, extreme close-up, etc.).
Physical state
The judges coded whether the older individual's physical state was weak, medium or strong. Weak physical state included restricted mobility, frailty (e.g. bent forward, bedridden, also detail shots of hands without any activity), being in need of care (i.e. supported by another person or a walker, pushed in a wheelchair). Medium included states that were neither powerless nor energetic, while strong included being fit, active and energetic (e.g. exercising, dynamic walking). No information was coded if the physical state was not clearly identifiable (extreme close-up or detail shot).
Social context
The judges coded whether the older individual was alone, primarily in a private social context, primarily in a professional care context or in a private social and professional care context. Alone included only the older individual in the photograph. Primarily private social context included scenes with the older individual among family member(s) and/or friend(s), or another person who was not care or medical personnel. Primarily professional care context included scenes in which the older individual depicted was being taken care of by care-givers or medical personnel. Private social and professional care context included scenes of the older individual in a care or medical setting with family members and/or friend(s) or other residents in a nursing home. No information was coded if the social context was not clearly identifiable (extreme close-up or detail shot).
Social support
In those cases where the older individual was not coded as alone, the judges coded whether the person received support, gave support, was engaged in reciprocal support, companionship or in a neutral social interaction. Received support entailed material, informational, instrumental or emotional support given to the older individual (i.e. being comforted or supported/led by another person; also if the hand is being supported). Gave support included the older individual giving material, informational, instrumental or emotional support (i.e. explaining something to others, doing volunteer work or comforting another person). Reciprocal support entailed instrumental or material reciprocal support (i.e. pushing a wheelbarrow, carrying boxes together). Companionship included the older individual and another person spending free time together, doing something together (i.e. walking together, spending time including being separated by a window). Neutral included the older individual depicted with other person(s), but without visible support, or engaged in a neutral undertaking (i.e. waiting at the bus stop without any interaction depicted). No information was coded if it was unclear who was giving support to whom (insufficient details).
Emotional state
The judges coded whether a person was in a negative, neutral or positive emotional state. Negative emotional states included sadness, worry/anxiety, disgust, despair or anger (e.g. tears in eyes, hands in front of the face looking down, frowning, furrowed brows, drooping eyes, lines across forehead). Positive emotional state included happiness, love or joy (e.g. laugh lines, eyebrows pulled up, head tilted upwards, smile). Neutral included states that were neither negative nor positive. No information was coded if the emotional state was not clearly identifiable (no face visible or face primarily covered by a mask).
Pronounced cognitive activity
The judges coded whether the person showed pronounced cognitive activity or no pronounced cognitive activity. Pronounced cognitive activity included scenes in which the person performed an activity that required a high level of mental acuity, including attention, reasoning and planning (e.g. playing an instrument, sitting at a laptop, crafting, dialling a number on a mobile phone), while no pronounced cognitive activity included no such activity. No information included images without (visible) faces or signs of activity.
Gender
The judges coded whether the person passed as male, female or androgynous. No information included detail shots without any signifier of gender identity.
Physical distance
In photographs with older individuals and at least another person, the judges coded whether the older individual maintained physical distance to the other person, maintained no physical distance or maintained physical distance through windows or on a balcony. Physical distance included scenes with physical distance (i.e. visible distance of approximately two metres between the older individual and at least one other person when shopping, going for a walk, sitting apart, greeting by touching feet), while no physical distance included scenes in which the older individual and at least one other person were physically close to each other (e.g. hugging, sitting next to each other, less than approximately two metres). Physical distance through windows or on a balcony included scenes of balcony/window visits (older individual is in home/on a balcony, friend(s)/family member(s) outside the window/balcony), or interactions with a fence between the older individual and the other person. No information was coded if the background was too blurry.
Analytical strategy
We used the sample of cover photos (N = 604), including photographs that appeared multiple times, to analyse the older individual's presentations of the body, visibility of the face, setting, physical state, social involvement, gender, emotional state and cognitive activity. For two indicators, namely social support and physical distance, a reduced sample size (N = 447), excluding photographs with older individuals coded as alone, was used given that the rating of these indicators required the presence of more than one person. Table 2 gives an overview of the assessed dimensions, the coding categories and the frequency distributions across the categories.
In order to test each hypothesis, chi-square goodness-of-fit tests were undertaken to investigate the frequency distribution of categories across the observable indicators. Specifically, we tested the difference between the observed and the expected frequency for each dimension to identify categories with a higher frequency. We used SPSS (version 25.0, IBM Corp., Armonk, NY, USA) to perform all statistical analyses.
Results
The presentations of older bodies
There were statistically significant differences in the presentation of older bodies: χ2 (6, N = 604) = 161.4, p ≤ 0.001. The upper body of older individuals was portrayed most frequently (24.5%), followed by hands (19.9%), back of the body (19.7%) and the total body (15.2%). Legs and feet (9.9%) and head and shoulders (9.8%) were coded at a similar frequency. In a few photographs, the vaccinated upper arm was coded (1.0%). The detail shot of a wrinkled hand holding on to a walker symbolising an older person is illustrated by Figure 2.
The visibility of older faces
There were statistically significant differences in the frequency of the visibility of the older faces: χ2 (2, N = 604) = 204.5, p ≤ 0.001. In most cover photos, the face of the older individual was not visible (i.e. absent information; 60.8%) (as examples, see Figures 1–3). At a similar frequency, the face was partly visible (19.0%) or completely visible (20.2%). Follow-up analyses showed that the differences between no visibility and partial visibility taken together, and completely visible faces, were statistically significant: χ2 (1, N = 604) = 214.5, p ≤ 0.001.
Setting
Judges coded no information in a fifth (20.4%) of cases. There were statistically significant differences in the frequency of settings: χ2 (5, N = 485) = 334.5, p ≤ 0.001. Older individuals were primarily coded within a care setting (35.9%) (Figure 3), followed by a public setting (16.6%), home setting (10.3%) or a testing or vaccination setting (10.3%), rarely a medical setting (4.3%) or other setting (2.8%). Follow-up analyses showed that the differences between care, medical and home settings taken together against all other settings were statistically significant: χ2 (1, N = 481) = 34.5, p ≤ 0.001.
Physical state of the older individuals
In terms of physical state, judges coded no information in a tenth (11.4%) of cases. There were statistically significant differences in frequency in social context: χ2 (2, N = 535) = 271.67, p ≤ 0.001. We found a high frequency of photographs portraying older individuals as physically weak (54.6%), followed by medium physical state (30.8%). A strong physical state was rarely depicted (3.1%). Figure 2 exemplifies the weak physical state of the older individuals as slightly hunched over and leaning on a walker.
Social context
Regarding social support, judges coded no information in a fifth (21.4%) of cases. There were statistically significant differences in frequency regarding social context: χ2 (3, N = 475) = 101.6, p ≤ 0.001. Older individuals were depicted most frequently in a professional care context (30.1%) or alone (26.0%), followed by a primarily private context (15.7%). The combination of private and professional contexts was rarely depicted (6.8%).
Social support
Regarding social support, judges coded no information in a fifth (22.2%) of cases. There were statistically significant differences in frequency in social context: χ2 (4, N = 447) = 295.49, p ≤ 0.001. Most frequently, older individuals were depicted as receiving support (46.5%), less frequently in companionship (25.3%), rarely in neutral situations of support (5.1%) or as giving support (0.9%), and never in situations of reciprocal support (0.00%).
Emotional state
Regarding emotional state, judges coded no information in almost three-quarters (71.9%) of cases (among them, 16.9% of older individuals were coded wearing a mask). There were statistically significant differences regarding emotional state: χ2 (2, N = 170) = 48.3, p ≤ 0.001. Older individuals were coded similarly often as being in a positive emotional (13.2%) and neutral state (12.6%). Rarely, cover photos showed a negative emotional state (2.3%).
Pronounced cognitive activity
In the case of cognitive activity, judges coded no information in more than half (55.0%) of cases (faces needed to be sufficiently visible). Older individuals were more frequently coded without pronounced cognitive activity (41.9%) than with pronounced cognitive activity (3.1%).
Gender
Regarding gender, judges coded no information in almost a quarter (23.5%) of cases. The gender of the individuals was not distributed equally across the three categories: χ2 (2, N = 462) = 226.33, p ≤ 0.001. Older individuals were most frequently coded as passing as female (47.5%), less frequently as passing as male (25.2%) and rarely as androgynous (3.8%). The difference in frequency between passing as female and androgynous taken together and passing as male was statistically significant: χ2 (1, N = 462) = 226.33, p ≤ 0.001.
Physical distance
Regarding physical distance, judges coded no information in less than a tenth (8.3%) of cases. The indicator physical distance was not distributed equally across the four categories: χ2 (3, N = 447) = 631.39, p ≤ 0.001. Most frequently, older individuals were not portrayed as keeping distance (76.3%), while fewer images portrayed them keeping distance (11.2%), or keeping distance through windows or on balconies (4.2%).
Discussion
At the beginning of the COVID-19 pandemic, scholars from the field of gerontology warned in numerous opinion papers that the pandemic might increase the marginalisation of older adults and lead to a resurgence of ageism in its various guises (for an overview of these opinion papers, see Silva et al., Reference Silva, Silva, Bacurau, Francisco, Assumpção, Neri and Borim2021). Researching the media portrayals of older adults in verbal contexts, various studies found they were indeed negatively stereotyped (e.g. Allen et al., Reference Allen, Odziemczyk, Perek-Białas and Ayalon2021; Köttl et al., Reference Köttl, Tatzer and Ayalon2022; Lagacé et al., Reference Lagacé, Doucet, Dangoisse and Bergeron2021; Zhang and Liu, Reference Zhang and Liu2021) in contrast to the increasingly diverse media representations of older adults in pre-pandemic times (see Ylänne, Reference Ylänne, Twigg and Martin2015). Our study of older individuals in cover photos (N = 604) on online news sites in Germany during the early period of the COVID-19 pandemic (1 January 2020 to 31 January 2021) echoes and enriches these findings. As described below, our study identified a variety of negative stereotypes and a visual othering of older adults in the, for the most part, generic and symbol-heavy photographs. Our analysis reveals that the unique components of the photographs intertwine and reinforce each other into an overall composition that visually frames older adults as an other. Specifically, the composition of the photographs visually segregates older adults, who literally rarely face the viewer, conveying little competency and agency, often stereotyped through a selective view of their wrinkled hands and immobile legs (deindividualisation and dehumanisation). In addition, the photographs uniformly show older adults as vulnerable, dependent and alone (homogenisation and objectification), thus reinforcing their othering by emphasising features negatively associated with the fourth age. This visual othering amplifies the overall misrepresentation of older adults in a stereotypical and prejudiced manner – a phenomenon that should be understood as visual ageism.
Distanced, disembodied and faceless
As expected in Hypothesis 1, the stereotypical visual portrayal of older adults in articles concerning the COVID-19 pandemic was constructed by a high degree of social distance, selected body parts and the absence of faces. Given that the COVID-19 pandemic was framed as a problem of older people, often segregating the older population from the young (Lichtenstein, Reference Lichtenstein2021), it is not surprising that we found older adults to be visually distanced in the cover photos. The frequent depictions of the back of the body (15.2%) and the whole body (15.2%) visually introduces a social distance to the viewer, consequently separating them from young or supposedly ageless, midlife viewers (van Dyk, Reference van Dyk2016). While the most intimate portrayals showing head and shoulders were rare (9.8%), we also found a less-distant portrayal of older individuals (24.5% upper body) that could point to a benevolent ageism (Vervaecke and Meisner, Reference Vervaecke and Meisner2021), which defines older adults as a homogenous risk group for which one has sympathy and which one feels the need to protect. In addition to visually segregating older adults, older bodies are disembodied (Parry, Reference Parry2010) with the focus on selected body parts (hands: 19.9%; legs and feet: 9.9%), which turns them into stereotypes (Jewitt and Oyama, Reference Jewitt, Oyama and Hughes2012). With the focus on body parts that highlight physical markers of decline and decay associated with old age (Chivers, Reference Chivers2011), including wrinkles and age spots or restrictions to autonomy depicting legs and feet, older adults are essentialised through features that are negatively associated with the fourth age (Rozanova, Reference Rozanova2010; Uotila et al., Reference Uotila, Lumme-Sandt and Saarenheimo2010).
Further, the dramatic absence (60.8%) or partial depiction of faces (19.9%) de-individualises older adults and in turn segregates them further as individual narratives are important for viewers to engage (Bleiker et al., Reference Bleiker, Campbell, Hutchison and Nicholson2013). In contrast, no other age group was portrayed visually in news stories during the COVID-19 pandemic without showing their faces (Martikainen and Sakki, Reference Martikainen and Sakki2021). Given that faces show emotions and personality, they not only create an intimate connection to the viewer but also reflect agency and competence (Archer et al., Reference Archer, Iritani, Kimes and Barrios1983; Konrath and Schwarz, Reference Konrath and Schwarz2007). In addition to a high frequency of showing only the back view of older individuals' bodies, they are thus characterised as objects rather than as subjects with agency. Overall, the faceless stereotyping creates a dehumanising portrayal of older adults (Crimp, Reference Crimp, Grossberg, Nelson and Treichler1992; Dan and Ren, Reference Dan and Ren2021).
The stereotype of the vulnerable and dependent old person
Globally, news articles framed older adults as vulnerable, dependent, lonely and in need of care during the COVID-19 pandemic. Our findings indicate – in line with Hypothesis 2 – that news photographs often reflected the representation visually. Specifically, we found portrayals of often physically weak (54.6%) older individuals in need of professional care (30.1%), with limited social interaction (26.0%), receiving support (46.5%), rarely characterised as cognitively active (3.1%), and often portrayed in care settings (35.9%) or at home (10.3%). Older individuals depicted with a care provider, or alone and at home, emphasises dependence and symbolically segregates them from the rest of society (Martikainen and Sakki, Reference Martikainen and Sakki2021), feeding into the ageist narrative of the dependent and lonely older person (Lagacé et al., Reference Lagacé, Doucet, Dangoisse and Bergeron2021; Morgan et al., Reference Morgan, Wiles, Williams and Gott2021; Zhang and Liu, Reference Zhang and Liu2021). While being at home itself can be interpreted in a positive light, as known from studies on advertisements (Swayne and Greco, Reference Swayne and Greco1987; Williams et al., Reference Williams, Wadleigh and Ylänne2010), findings on distanced, de-individualised and dehumanising portrayals as well as the overly negative depictions suggest a rather one-sided, negative interpretation. Adding to the stereotypical portrayal of bodies and the absence of faces, their portrayal as weak and in need of care conveys the idea that they are passive objects and targets of others' efforts (Kress and van Leeuwen, Reference Kress and van Leeuwen1996; Bell and Milic, Reference Bell and Milić2002), implying a lack of self-determination and social power instead of agency and strength. In congruence with arguments that women are figuratively ascribed to ‘decline and demise’ (Meyers, Reference Meyers2002: 163), and in line with Hypothesis 3, older individuals were portrayed most frequently as female passing (47.5%).
This representation of older adults does not fully reflect their representation in the German population: only about one-quarter of individuals aged 65+ (26.4%) are in need of care (Federal Statistical Office of Germany, Reference Femers2021). While the percentage is much higher for people aged 90+ (76.3%), only a minority of older adults in Germany live in long-term care facilities (4% of people aged 65+ and 18% of those aged 80+ years). This number of 80–90-year-old individuals in need of care in a facility is in stark contrast to the large number of photographs in our sample. While feelings of loneliness increased during the COVID-19 pandemic, 76–90-year-old adults did not feel lonelier than 46–55-year-old adults (Huxhold and Tesch-Römer, Reference Huxhold and Tesch-Römer2021). Even though older people faced the effects of preventive measures such as social distancing, quarantine and general self-isolation in different settings, including nursing homes and in the community, their lives during the pandemic were as heterogeneous and complex as those of younger age groups (Simonson and Tesch-Römer, Reference Simonson and Tesch-Römer2022). Yet, the visual representations identified in our study ignore the heterogeny of older adults' lived experiences.
The news stories in our sample centred on the entire older population, yet they primarily focused on negative features attributed to the fourth age. While older women and the oldest-old have been underrepresented in the media (Rozanova, Reference Rozanova2010; Ylänne, Reference Ylänne, Twigg and Martin2015; Myrczik et al., Reference Myrczik, Bowen, Franke, Taeuber and Kessler2022), our findings show that these groups do appear in the media, albeit in the context of vulnerability. We are aware that there is a ‘universal ontology of human vulnerability’ (Gilleard and Higgs, Reference Gilleard and Higgs2011: 135), wherein people of all ages are vulnerable and thus interdependent, and we, thus, need to be careful about potentially interpreting the dependency of older individuals as inherently negative. However, it is precisely the homogenising portrayal of older individuals as weak, in need of care, alone, rarely providing social support (0.9%) or cognitively active that demands attention. By essentialising older individuals as vulnerable and passive in addition to the lack of intimacy in the visual composition, they are turned into something other, separate from a supposedly ageless midlife (van Dyk, Reference van Dyk2016). Thus, our findings can be interpreted as a visual othering of older adults during the COVID-19 pandemic. From a psychological perspective, othering older adults (Whitfield, Reference Whitfield2001; Jonson, Reference Jonson2013) as more vulnerable to the virus than oneself might be interpreted as a form of self-protection when confronted with one's own mortality (Martens et al., 2006) – a possibility that applies to journalists as much as to the rest of society (Kessler, Reference Kessler2020). However, we would like to note that this selection of faceless photographs might also be pragmatic, in response to the need to protect the privacy of older adults in nursing homes (Xu, Reference Xu2022).
There was one exception to Hypothesis 2: in contrast to our hypothesis, we found few sad portrayals (2.3%) of older individuals, with more images depicting positive (13.2%) or neutral emotional states (12.6%). However, we caution that these results are based on a small percentage of all photographs in which faces were fully visible. Considering the overall negative characteristics attributed to older individuals, this result stood out, even though gerontological studies have previously demonstrated the resilience of older adults during the COVID-19 pandemic (e.g. Sterina et al., Reference Sterina, Hermida, Gerberi and Lapid2022). One could argue that image editors might have shied away from showing faces with negative emotions, concerned that viewers may feel fear, sadness or disgust (Baberini et al., Reference Baberini, Coleman, Slovic and Västfjäll2015) and thus not read the article. However, the positive portrayal of older adults can also be interpreted as hostile and benevolent ageism co-existing (Reference Yang, Au, Wong and LouYang et al., in press), insofar as older adults are vulnerable, alone and dependent, but are framed as rather content with their situation.
Taken from the archives – stereotypical depictions with little reference to the pandemic
Our hypothesis of a high number of photographs depicting physical distance as a visible indicator of the COVID-19 pandemic (Hypothesis 4) was not supported. Most frequently, older individuals were not portrayed as maintaining physical distance to another person (76.3%), few were portrayed as maintaining physical distance (11.2%), or through windows or on balconies (4.2%). Apart from older individuals depicted alone, we found few symbolic representations of the isolation of older adults, especially when compared to our findings on companionship (25.3%) and a private social context (15.7%), or being out in public (16.6%). The latter findings might point to a small segment of nuanced reporting on older adults that was found in German print media as well (Köttl et al., Reference Köttl, Tatzer and Ayalon2022), yet the segregating and othering presentations of older individuals in the cover photos were more frequent and thus dominant overall.
Another reason for a lack of visual reference to the pandemic might be that it was difficult to take photographs of people, particularly at the beginning, when protective measures were very strict – restricting journalists in contacting people in general and older people in particular. Without an explicit visual reference (e.g. the low number of older individuals coded wearing a mask – 16.9%) to the COVID-19 pandemic, older adults were stigmatised as inherently rather than situationally vulnerable due to political measures or increased susceptibility. Abstracted from the present COVID-19 pandemic, the photographs become symbolic and generic and might be (re)used in any other context (Vobic and Tomanic Trivundza, Reference Vobic and Tomanic Trivundza2015). The sample's high number of photographs from large newswires (79%) – typical for Western media (Fahmy, Reference Fahmy2005) – might explain the stereotypical portrayals given that newswires distribute stock photography, i.e. photographs that broadly symbolise certain issues (Kraemer and Lobinger, Reference Kraemer, Lobinger and Lobinger2019). Concomitantly, and in line with other studies (Wanta and Roark, Reference Wanta and Roark1994), the selection of news photographs plays a major role (Fahmy and Neumann, Reference Fahmy and Neumann2011) insofar as a small number of photographs (8.8%) accounted for almost half of the photographs in our sample, highlighting the importance of the selection pattern. The selection might be explained by the everyday practices of online newsrooms increasingly shaped by time constraints and limited human resources (Le Cam and Domingo, Reference Le Cam, Domingo, Vos and Heinderyckx2015), leaving little room for reflection on the selection of stereotypical photographs. Further, in response to the increasing pressure to find a visual for every online news story despite perhaps not having access to any images directly related to the piece, online journalists resort to symbolic photographs to fill ‘the empty frame’ despite their own discontent with the practice (Vobic and Tomanic Trivundza, Reference Vobic and Tomanic Trivundza2015). Further, the historically exceptional situation of the COVID-19 pandemic might have stimulated the use of stereotypes as they help to reduce complexity in overwhelming situations (Fineman, Reference Fineman2011).
The media's power to shape societal representations of older adults
While viewers' perceptions and interpretations of cover photos might differ from the interpretation of single elements of the photographs, an image can unlock existing individual and collective narratives about older adults (Busselle and Bhandzic, Reference Busselle, Bhandzic, Ross and Lester2011). In addition, it is precisely such symbolic photographs that rely on a broader interpretative frame based on existing societal narratives (Vobic and Tomanic Trivundza, Reference Vobic and Tomanic Trivundza2015). As the literature has shown that ageist narratives have increased over time (Ng et al., Reference Ng, Chow and Yang2021), our results might be interpreted as demonstrating visual ageism (Loos and Ivan, Reference Loos, Ivan, Ayalon and Tesch-Römer2018), reflecting societal biases towards older adults by ignoring their heterogeneity. By selecting primarily negative and stereotypical photographs from archives in which more diverse images that are closer to reality are absent, online news media offered ageist interpretations of older adults and the COVID-19 pandemic. By repeating and normalising these images, they reify these narratives (O'Neill, Reference O'Neill2013). Hence, media practitioners need to be made aware of ageism through workshops and collaborations with scholars and practitioners (Carrigan and Szmigin, Reference Carrigan and Szmigin2000; Milner et al., Reference Milner, Van Norman, Milner, Beard, Biggs, Bloom, Fried, Hogan, Kalache and Olshansky2012; Xu, Reference Xu2022). We know from the literature that stereotypically homogeneous representations of age (Swift and Steeden, Reference Swift and Steeden2020) and ageist interpretations of the pandemic as dangerous only to older-age people (Fraser et al., Reference Fraser, Lagacé, Bongué, Ndeye, Guyot, Bechard, Garcia, Taler, Adam, Beaulieu, Bergeron, Boudjemadi, Desmette, Donizzetti, Éthier, Garon, Gillis, Levasseur, Lortie-Lussier, Marier, Robitaille, Sawchuk, Lafontaine and Tougas2020) negatively impacted them. Findings show that these narratives had severe negative effects on wellbeing and health (Cohn-Schwartz and Ayalon, Reference Cohn-Schwartz and Ayalon2021; Kornadt et al., Reference Kornadt, Albert, Hoffmann, Murdock and Nell2021; Levy et al., Reference Levy, Chang, Lowe, Provolo and Slade2022; Schlomann et al., Reference Schlomann, Bünning, Hipp and Wahl2021) and increased age discrimination (Higgs and Gilleard, Reference Higgs and Gilleard2021).
Limitations and future research
We chose to examine online news sites as the number of online readers has increased significantly in recent years, particularly during the COVID-19 pandemic. Future research should examine whether the visual representations in print newspapers differed from our findings. Further, representations in social media could also yield other, more-differentiated results, as some analyses have shown (Barrett et al., Reference Barrett, Michael and Padavic2021; Xi et al., Reference Xi, Xu, Zhang and Ayalon2021). We acknowledge the overly fragmented analysis through visual content analysis that might seemingly overemphasise the specific content of a photograph while underestimating the overall emotional impact of the entire composition. Thus, future research should focus on how viewers and readers perceived these representations during the COVID-19 pandemic. Our findings need to be interpreted with caution as we considered only the early period of the COVID-19 pandemic, potentially omitting different narratives that appeared after the arrival of the vaccine and the emergence of various virus variants. Another limitation of our study was that we focused solely on older adults, without taking into account other age groups that may have been subject to similar negative visual stereotypes, as the findings of a discourse analysis of three German newspapers suggests (Pelizäus and Heinz, Reference Pelizäus and Heinz2020). We further need to note the lack of positive portrayals, which was in part the result of the omission of celebrities – who are often portrayed positively in news photographs (Martikainen and Sakki, Reference Martikainen and Sakki2021) – from our sample, since we assumed that they are noticed primarily for other characteristics than their chronological age. Nonetheless, older individuals in our study were portrayed with positive emotions when faces were visible. Despite the fact that photographs usually emphasise a news story's message, there might be a congruency between visual and verbal presentation (Dan, Reference Dan2018). Future research might investigate potential discrepancies between the two modes of representation to reveal whether generally negative stereotyping exists or if there is instead a lack of quality images when selecting the visual material for news stories.
Conclusion
This study set out to investigate the visual representation of older adults in the four most popular German online news sites during the early period of the COVID-19 pandemic. Overall, negative stereotypical representations were widespread and seemed to evoke an othering of older adults. They were rarely shown with their faces, but were instead presented as disembodied stereotypes featuring traits negatively associated with the fourth age. Their vulnerability was even more pronounced as the people depicted were primarily women, failing to provide a more realistic representation of older adults in the German population. They were visually segregated and stereotypically presented as vulnerable and dependent. As such, the overall generic cover photos frame older adults as an other, separate from young people or ‘ageless’ midlife. The wider viewership's confrontation with their own mortality in the face of the SARS-CoV-2 virus might have exacerbated an othering of the older population in our societies, although one must simultaneously acknowledge that stereotypes help to reduce the burden of overwhelming situations such as a pandemic by reducing complexity. The overall negative stereotypical framing needs to be understood within the increase in ageism, on the one hand, and routines in online newsrooms with staff opting for stereotypical stock photographs, on the other hand.
To conclude, this study adds to the understanding of societal ageism, focusing on media representations of older adults and journalistic practices that reflect and reproduce ageist narratives. Our analysis highlights that the media need to reflect on these practices as the internalisation of negative age stereotypes has consequences for people of all ages.
Author contributions
JM supervised the data collection, conducted the analysis and wrote the paper. CS contributed to coding the data and revising the paper. AF contributed to revising the paper. E-MK planned the study and revised the paper.
Financial support
This work was supported by the Josef und Luise Kraft-Stiftung, which played no role in conducting and writing the study.
Competing interests
The authors declare no competing interests.
Open access
