Newspaper coverage on solidarity and personal responsibility in the COVID-19 pandemic: A content analysis from Germany and German-speaking Switzerland

Solidarity and personal responsibility have been repeatedly called upon during the COVID-19 pandemic. This study quantifies and contextualizes the use of these terms in newspaper coverage in Germany and German-speaking Switzerland based on n = 640 articles from six functionally equivalent newspapers. The term solidarity in the context of the COVID-19 pandemic was mentioned in 541/640 articles (84.5%) and was primarily used during phases with high death rates and comparatively stringent policies in place, supporting the idea that solidarity was used to explain restrictive measures to the population and motivate people to comply with these measures. German newspapers published more articles on solidarity than Swiss-German newspapers, consistent with more stringent COVID-19 policies in Germany. Personal responsibility was mentioned in 133/640 articles (20.8%), meaning that the term was less frequently discussed than solidarity. Articles covering personal responsibility included more negative evaluations during phases of high infection rates as compared to phases of low infection rates. Findings indicate that the two terms were, at least to some extent, used in newspaper reporting to contextualize and justify COVID-19 policy during phases of high infection rates. Moreover, the term solidarity was used in a high variety of different contexts and the inherent limits of solidarity were rarely mentioned. Policymakers and journalists need to take this into account for future crises to not jeopardize the positive effects of solidarity.


Introduction
During the COVID-19 pandemic, public health restrictions have varied depending on the specific pandemic situation in a country and the political environment. While many welcomed strict policy interventions at the beginning of the crisis, the voices of critics and skeptics were getting louder as the pandemic progressed (Georgiou, Delfabbro, & Balzan, 2020;Islam et al., 2020). Policymakers and stakeholders often evoked the concepts of solidarity and personal responsibility of citizens to justify the established measures in pandemic control. In Germany and Switzerland, for instance, solidarity was a frequently used term in politicians' speeches to attune and motivate people to adhere to restrictions enforced to contain the spread of SARS-CoV-2 (Merkel, 2020;Steinmeier, 2020;Swiss Federal Council, 2020). By contrast, the notion of personal responsibility seemed to be more prominently used when legal restrictions were eased to remind people to protect themselves against infection (Federal Office of Public Health, 2021;ZDFheute, 2022). Moreover, personal responsibility was mentioned by economic stakeholders advocating for the easing of restrictions to stimulate the economy (dpa-AFX, 2020; economiesuisse, 2021). These uses of the two terms suggest a normative connotation from policymakers to justify the introduction, waiver or absence of policies aiming at limiting viral spread, thus aiming to increase acceptance of them in the general population. However, overusing these expressions could lead to adverse reactions among people: The individual costs to comply with restrictions bas solidarity are perceived as too high over time (Prainsack & Buyx, 2011). Likewise, personal responsibility assumes rational, knowledgeable agents, but the pandemic crisis, paired with the parallel infodemic, made it difficult for some people to take on this responsibility (Balakrishnan, Ng, Soo, Han, & Lee, 2022;De Landsheer & Walburg, 2022). A closer examination of these two concepts in public debates allows for an understanding of their normative role and relevance in political communication.
Thus, this study aims to quantitatively examine how the terms "solidarity" and "personal responsibility" were used in Swiss-German and German newspaper coverage and whether they were used to contextualize or even justify COVID-19 policies. Investigating newspaper coverage allows for a retrospective reflection on how public debates were shaped, how they changed, and who had a say in them. This reflection can inform future efforts to control major crises by analyzing the underlying rationales of policy strategies applied in the COVID-19 pandemic.
The next section will introduce the concepts of solidarity and personal responsibility as the theoretical background of the study as well as existing literature investigating their implications in the COVID-19 pandemic. Hypotheses derived from theory will be presented at the end of this section. Section 3 will outline the methods used for the empirical part, a quantitative content analysis comparing newspaper coverage Germany and German-speaking Switzerland. Section 4 will present the results along the hypotheses outlined in section 2. The final section 5 will discuss these results with a particular focus on the limits of solidarity and the different tiers of solidarity practices, discuss the limitations of this study, and present our conclusions.

Theory
Prainsack and Buyx define solidarity as "practices that express the willingness to support others with whom we recognise similarity in a relevant respect" (Prainsack & Buyx, 2011). They suggest solidarity to work on three tiers: the individual, group, and institutional level (Prainsack & Buyx, 2017). While solidarity practices pose costs on an individual level, they are to the benefit of the broader community. Personal responsibility has been discussed in different contexts (e.g., Buyx, 2008;Steinbrook, 2006) and refers to the underlying ideal of the autonomous, independent individual that is held responsible for his or her actions. It is linked to both free choices and an understanding of the implications and consequences of actions (McKeon, 1957). Thus, while the two concepts of solidarity and personal responsibility both relate agents to actions that take place in a socio-cultural context, solidarity asks for supporting others in a community despite individual costs while personal responsibility stresses individual freedom.
Both concepts were linked to motivating people to protective behavior during the COVID-19 pandemic . Findings from a 9-country qualitative interview study on solidarity in the COVID-19 pandemic indicate that policies overly focusing on personal responsibility might counteract people's motivation for solidarity (manuscript under review). Collective responsibility, however, has been found to motivate people to comply with COVID-19-related policies , including using contact tracing technology (Lucivero et al., 2022), wearing masks (Zimmermann, Eichinger, Schönweitz, & Buyx, 2021), and getting vaccinated (Liu, Ao, Zhao, & Zhang, 2022). In a white paper written through the COVID-19 Rapid Response Impact Initiative, Cammett and Lieberman stressed the importance of solidarity-promoting public messaging already in March 2020, emphasizing the central role of mass media and societal leaders in that context (Cammett & Lieberman, 2020). Studies conducted in the early phases of the pandemic, as well as during and after lockdown measures, particularly referred to solidarity as a crucial concept to overcome the pandemic (e.g., Marzana et al., 2022;McCormick, 2020;Moss & Sandbakken, 2021;Peng & Berry, 2021;van Breen et al., 2021). Relatedly, there were repeated calls for global solidarity to overcome regional resource bottlenecks and for pandemic containment (Jecker & Atuire, 2021;Zagefka, 2022). Many scholars linked global solidarity with equity and justice principles, stating that global solidarity is needed to achieve equity in the context of COVID-19 (Ebrahim et al., 2020;Killeen & Kiware, 2020).
As the pandemic progressed, evidence and notions of the limits or failure of solidarity in the COVID-19 pandemic became apparent (Previtali, Allen, & Varlamova, 2020;West-Oram, 2021). Many authors lamented the failure of global solidarity in the context of the global distribution of scarce resources, such as COVID-19 tests or vaccines (e.g., Figueroa et al., 2021;Obinna, 2022;Sekalala et al., 2021;Syntia Munung, Ujewe, & Afolabi, 2022). On a more national or regional level, it was suggested that calls for solidaristic practices could sometimes also lead to polarization instead of social cohesion (Chadwick, 2020;Stjernswärd & Glasdam, 2021). For instance, intergenerational solidarity might also turn into ageism once solidarity towards the elderly is exhausted (Ayalon et al., 2020;Barrett, Michael, & Padavic, 2021;Barth et al., 2021;Ellerich-Groppe, Pfaller, & Schweda, 2021;Fraser et al., 2020). This polarization was already evident in April 2020, when some residents from Germany and German-speaking Switzerland strongly and spontaneously distinguished between those complying vs those not complying with restrictive measures .
On an individual level, a person can act both responsibly and in solidarity by taking responsibility for themselves while supporting others (Deutscher Ethikrat, 2020). Moreover, the concepts can be used to theoretically explain the trajectories of COVID-19 policies in different countries: Personal responsibility, stressing the freedom of the individual, may be used to justify the absence or easing of policies restricting individual freedom. Solidarity, on the other hand, stresses communal aspects and can, therefore, be seen as a justification for restrictive measures limiting individual freedom. As indicated by the Oxford Stringency Index (Ritchie et al., 2020), COVID-19 policy was generally less stringent in Switzerland than in Germany ( Figure A. 2). Therefore, mirroring the meaning of these terms to justify policymaking during the pandemic, we developed the following hypotheses: German newspapers published more articles about solidarity than Swiss-German newspapers (H1a) and Swiss-German newspapers published more articles about personal responsibility than German newspapers (H1b). If the terms were mainly used to justify COVID-19 policies, then we expect the terms solidarity and personal responsibility to be most often used by political actors (H2).
Moreover, if solidarity was used as a justification for stringent COVID-19 policies, the following additional hypotheses apply: -Coverage about solidarity was higher in phases with high stringency of policy measures than in phases with lower stringency (H3a) and more positively/less negatively evaluated (H4a) -Coverage about personal responsibility was higher in phases with lower stringency of policy measures than in phases with high stringency (H3b) and more positively/less negatively evaluated (H4b)

Material and methods
Newspaper coverage was selected as the analytic body for this research, excluding other media content (i.e, social media or online news platforms). First, newspaper articles are stable, published entities that can be systematically searched for in online databases. By contrast, content on online news platforms or social media, for instance, is constantly changing, edited, and deleted, and the algorithms of search engines are biased and intransparent (McMillan, 2009). Second, newspapers simultaneously inform the public about ongoing issues and reflect the public debate (Peters, 1994), thereby shaping the public's view on things while being influenced by issues that move public debates themselves. During the COVID-19 pandemic crisis, newspapers and other traditional journalistic media regained importance as reliable, trustworthy and accessible information sources (Amann, Sleigh, & Vayena, 2021;Zimmermann, Fiske, et al., 2021). As such, they played an important role in the COVID-19 pandemic by motivating individuals to comply with policies and enact protective behavior, while reflecting societal happenings and situational changes (Cammett & Lieberman, 2020).

Comparative approach
Switzerland and Germany are neighboring countries in Western Middle Europe. In 2021, Germany had 83.24 million inhabitants (Statistisches , and Switzerland had 8.74 million (Bundesamt für Statistik, 2022). Switzerland has three major language areas (German-speaking, French-speaking, and Italian-speaking) that differ in terms of culture, which was shown to be relevant for the spread and containment of SARS-CoV-2 (Deopa & Fortunato, 2021;Mazzonna, 2020). Particularly the German-speaking part of Switzerland is culturally and economically closely connected to Germany. As such, we apply a quasi-experimental most similar system design by comparing newspaper coverage from Germany and German-speaking Switzerland, which are embedded in similar media systems but differing policy responses in the COVID-19 pandemic (Berg-Schlosser, 2003;Teune & Przeworski, 1970). While both countries reacted similarly when SARS-CoV-2 started to spread in Europe, COVID-19 policy was less stringent in Switzerland than in Germany from the summer of 2020 on, as measured by the Oxford Stringency Index (Ritchie et al., 2020).

Selection of articles
The time frame of analysis was 1 January 2020 -30 April 2022, from the start of the COVID-19 pandemic until most COVID-19-related restrictions were lifted in Switzerland and Germany. From each country, two newspapers with different political orientations and comparatively high readerships as well as the most-read tabloid were selected for the analysis ( ). 1 All included newspapers belong to private publishing groups and are not government-owned. Relevant articles were retrieved through a keyword search in Factiva (Dow Jones) using the search algorithm: (COVID* OR Corona* OR SARS-CoV-2) AND (Solidarität OR solidarisch* OR eigenverantwort*) (searching titles and lead texts only). Full articles were downloaded and manually screened whether they met the inclusion criteria. Journalistic newspaper articles that covered solidarity and/or personal responsibility in the context of the COVID-19 pandemic were included. Articles without editorial content (e.g. glossaries, content summaries, TV programs) were excluded from the analysis. When front-page articles introduced a longer article on the same relevant topic, the full text was retrieved from Factiva in an additional, targeted search (n = 26) (see Figure A. 1 for the detailed search and selection strategy).

Operationalization and data collection
Article metadata (title, length, publication date, newspaper) were retrieved from Factiva. To assess what was covered by newspapers, who had a say in coverage, and how it was evaluated, we developed a codebook based on previous newspaper content analyses analyzing similar issues (Gerhards & Schäfer, 2009;Schäfer, 2009;Zimmermann, Aebi, Kolb, Shaw, & Elger, 2019). We measured how extensive the terms "solidarity" and "personal responsibility" were discussed in articles, in what contexts the terms were mentioned, how they were evaluated (measured as the presence or absence of positive and negative evaluative statements about the implementation of solidarity and personal responsibility practices) and what actors were cited using those terms (details about these variables are presented in the supplementary codebook). A preliminary list of context categories was derived through a qualitative content analysis procedure including a step-wise abstraction into categories (Mayring, 1994), applied to a random selection of 57 articles. The actor and evaluation categories were drawn from a codebook used in a previous study on a different health topic (Zimmermann et al., 2019).
Coder training and codebook refinement took place in four coding sessions and was performed by the main author and a student assistant. An inter-coder reliability test was performed on 57 articles that were sorted by publication date and newspaper; then, every 10th article was selected. Because articles published between November 2021 and April 2022 were coded later, the reliability test only relied on articles published before November 2021. We use Scott's Pi (Scott, 1955(Scott, , 2009) to calculate the reliability score because our variables are all nominal with highly variable degrees of freedom, which are taken into account for this measurement. For the categories feature and evaluation of solidarity, we reached good reliability with scores higher than 0.7, for the other categories the reliability scores ranged between 0.4 and 0.7, which is considered sufficient reliability (Krippendorff, 2009) (see Table A. 1 for detailed reliability scores). After inter-coder reliability testing, results were screened for repetitive differences. Coder errors and codebook ambiguities were discussed and coder instructions in the codebook were refined and specified accordingly. No changes were made concerning categories or variables. To further improve reliability, a second coder double-checked the coding for articles where the first coder experienced uncertainties during coding.

Data analysis
Data were analyzed using SPSS 28. For the longitudinal analysis, phases were defined based on pandemic development as mirrored by death rates and the Oxford Stringency Index (Table A. 2; Figure A. 2). Since these phases differed in length, the average number of articles per month was calculated for data analysis to ensure comparability between the different time frames. A Chi-Square Goodness of Fit Test was performed to test whether the proportion of the number of articles about solidarity and personal responsibility was equal between countries (H1). A Z-Test for Independent Proportions was applied to test for significant differences in proportions, applying, if necessary, the Benjamini-Hochberg correction for multiple testing (H2, H3). The significance threshold was set to 0.05.
Because the media coverage was dominated by one newspaper outlet (Süddeutsche Zeitung), we examined whether the findings were consistent when excluding this outlet from the analysis, adding to the robustness of the findings. Only when country comparisons were performed (H1, section 4.1), the functionally equivalent Swiss-German newspaper outlet (TagesAnzeiger) was excluded from the analysis, too.
Some 34/640 articles (5.3%) covered both concepts but only half of them (n = 17) discussed both concepts in more than one sentence. Except for one article, these 17 articles were all published in broadsheet newspapers. The most-mentioned context was compliance for both solidarity and personal responsibility (Table A. 3). Most articles were published throughout the year 2020 (n = 10, 58.8%). In 2021, 5 articles (29.4%) were published; 4 of them in November and December 2021. The remaining 2 articles were published between January and April 2022.
Regarding personal responsibility, the amount of coverage remained relatively stable over time (Fig. 1). When comparing coverage among countries, the number of articles was higher in German-speaking Switzerland than in Germany in phase III (summer 2020) (23/31; 74.2%, p = .036) and German newspapers reported more on personal responsibility in phase IX (12/15; 80.0%, p = .002, Figure A. 3).
To test H2, regional, national, and international politics were combined into one category ("politics"). All other actors were summarized in a second category ("other", Table A. 3). Because "media" were coded as actors in all journalistic commentaries and "citizens/individuals" in all reader letters, these sorts of articles were excluded for this analysis. A Paired-Samples Proportions Z-Test indicate a significantly lower proportion of political actors as compared to other actors in articles about solidarity (132/408, 32.4% vs 229/408, 56.1%, z = − 5.397, p < .001). No significant differences were found in articles about personal responsibility (politics: 40/90, 44.4%, others: 41/90, 45.6%, z = − 0.120, p = .905). These findings were reproducible when excluding Süddeutsche Zeitung from the analysis (see Appendix B.2).
When comparing the proportions of article evaluations between phases of relaxed versus stringent COVID-19 policies (Table 4), personal responsibility was significantly more negatively evaluated during stringent policy phases (z = − 2.026, p = .043). This finding was not reproducible when excluding Süddeutsche Zeitung (see Appendix B.3). The other differences were not statistically significant when applying the Benjamini-Hochberg correction for multiple testing.

Discussion
Overall, German and Swiss-German newspaper coverage of solidarity was more extensive than on personal responsibility, reflecting that a pandemic is a global crisis that requires collective efforts (Prainsack & Buyx, 2011;West-Oram & Buyx, 2017). Our findings only support part of the hypotheses (Table 5). Supporting H1a and mirroring more stringent policies in the country, German newspapers published more articles on solidarity than Swiss-German newspapers. However, contrary to H1b, there was no statistically significant difference in coverage among the countries regarding personal responsibility. However, we do report a tendency for more articles published in Swiss-German newspapers. Given the relatively low number of articles published about personal responsibility, a higher N would possibly make the reported tendency statistically significant.
Supporting H3a but not H3b, coverage about solidarity was more common in phases with stringent policies, but there was no difference between the phases in coverage regarding personal responsibility. Political actors were often cited when using those terms, but only for personal responsibility, they were cited significantly more often than other actors. Thus, H2 can only partly be accepted. Moreover, partly supporting H4b, personal responsibility was more negatively evaluated in phases of high stringency. These findings indicate that, to a limited extent, solidarity was used by politicians and the media to justify and explain policies, particularly the implementation of restrictive policies. The term personal responsibility was much less important and did not seem to serve the purpose of refraining from or easing restrictions in times of low stringencies, based on our data. Also, only few articles covered both concepts in more than one sentence, indicating a only limited reflection on the connection or potential opposition of the two terms.
Conceptually, solidarity and personal responsibility are epistemologically different in the sense that solidarity refers to individuals as inherently relational agents embedded in a collective, whereas personal responsibility refers to the individual as a discrete entity (Smiley, 2017). In the literature, however, the concepts are often discussed as being parallel concepts that (should) go hand in hand. Some argue that solidarity comes with obligations and people can (and should) be held responsible for meeting those (Davies & Savulescu, 2019). The German Ethics Council coined the term "solidarische Eigenverantwortung", meaning that people have to take responsibility for their own actions to practice solidarity in the context of the COVID-19 pandemic (Deutscher Ethikrat, 2020, p. 5). As a German research group illustrated based on a journalistic commentary published in Neue Zürcher Zeitung in April 2020, it was sometimes even assumed that solidaristic actions should be based on personal responsibility, which would result in the absence of state intervention (Ellerich-Groppe et al., 2021). Our findings indicate that newspaper reporting did not support this to a larger extent and that the reported linkage of seeing the two concepts as two sides of the same  coin was the exception rather than the standard in newspaper reporting.

The limits of solidarity
While personal responsibility has been evaluated critically in newspaper coverage, solidarity was rather positively assessed, but that changed in the second half of the pandemic. This relates to observations that solidarity cannot be accounted for in the long term because people are not willing to cover the costs of solidaristic actions and expect reciprocity indefinitely (Collins, 2020;Prainsack & Buyx, 2017). Yet, a surprisingly small portion of articles addressed such limits of solidarity explicitly or discussed unsolidaristic behavior. This may point to a lack of awareness in newspaper coverage that solidarity-based obligations and requirements come with such limits (see e.g., Drotbohm & Reichardt, 2020;Ellerich-Groppe, 2023). This might have contributed to overextending the notion of solidarity as a reason for compliance and restrictive policiesparticularly in Germany, as our data indicate. This can be problematic, as solidarity as a normative reason for public policy is disputed, particularly in pluralistic, democratic societies (Lopez Frias & Thompson, 2022).
Moreover, an important ratio of articles mentioned solidarity in the context of a pure declaration of solidarity (e.g. by clapping for healthcare workers). This is not in line with Prainsack and Buyx' definition of solidarity, as these practices did not evoke relevant costs. Instead, it contributed to the use of the word "solidarity" as a political statement, as being on "the right side of things". This became particularly noticeable in the late stages of the pandemic in Germany, where the word "solidarity" was commonly used to frame demonstrations against COVID deniers, thereby creating a divide rather than a sense of general togetherness (Häyry, 2020;Stjernswärd & Glasdam, 2021). This shows that solidarity is a diversified term that is used in various settings, as it lacks a commonly accepted, firm definition (Prainsack & Buyx, 2017).

The three tiers of solidarity
According to Buyx' theoretical approach (2011, 2017), solidarity can be practised on the individual, the group, and the contractual level. These levels were also represented in newspaper coverage, with a particular focus on individual solidarity. Contractual solidarity was represented to a lesser extent, as reported in the context categories solidaristic political decision-making and international solidarity. Of note, existing national solidarity institutions, such as social security systems, were not captured by this analysis if they were not specifically mentioned in the context of the COVID-19 pandemic. Another potential reason for this is that these aspects might not be discussed under the concept of solidarity, even though they can be understood as institutionalized solidarity (Prainsack & Buyx, 2011).
Equivalent to individual and group-level solidarity, solidaristically motivated support of others, compliance with policies, and vaccination have been studied and discussed in COVID-19-related scholarly literature. First, the support of others has been shown to have increased in the early stages of the pandemic in Germany (Bertogg & Koos, 2021). Yet, after the initial phase with a first wave and lockdown, solidaristic support declined (Ntontis et al., 2022). Second, solidarity has been acknowledged early in the pandemic as a prerequisite to compliance with COVID-19 policies (Cammett & Lieberman, 2020). Indeed, people in German-speaking Switzerland and Germany perceived solidarity as an important motivator for compliance in the early stages of the pandemic Liekefett & Becker, 2021;Zimmermann et al., 2022). Also, mask-wearing was repeatedly associated with solidaristic motivation (Cheng, Lam, & Leung, 2020;Zimmermann, Eichinger, et al., 2021). Third, solidarity is reported as a motivation for COVID-19 vaccination in many studies, in line with the narrative in the newspapers in our study (Kerrigan et al., 2023;Majid, Ahmad, Zain, Akande, & Ikhlaq, 2022). Yet, in the context of vaccination prioritization at the beginning of the COVID-19 vaccination campaign, the narrative of  Table 5 Overview and evaluation of hypotheses.

Hypotheses
Interpretation based on findings

H1a
German newspapers published more articles about solidarity than Swiss-German newspapers.

H1b
Swiss-German newspapers published more articles about personal responsibility than German newspapers.
Rejected; no statistically significant difference

H2
The terms solidarity and personal responsibility were most often used by political actors.
Supported for personal responsibility; rejected for solidarity H3a Coverage about solidarity was higher in phases with high stringency of policy measures than in phases with lower stringency.

H3b
Coverage about personal responsibility was higher in phases with lower stringency of policy measures than in phases with high stringency.
Rejected; no statistically significant difference

H4a
Coverage about solidarity was more positively evaluated in phases with higher stringency of policy measures and more negatively in phases with lower stringency Rejected; no statistically significant difference

H4b
Coverage about personal responsibility was more positively evaluated in phases with lower stringency of policy measures and more negatively in phases with high stringency.
Partly supported (more negatively in phases with high stringency) giving others priority because they were more endangered was not always linked to genuine solidarity, but sometimes also to vaccine hesitancy Knotz, Gandenberger, Fossati, & Bonoli, 2021).

Limitations
Only a selection of newspapers from each country was analyzed in this study. The amount of coverage on the topics of solidarity and personal responsibility varied considerably between newspapers, indicating important newspaper-specific differences. With the selection of newspapers, we were aiming for functional equivalence but due to these differences, results might likely have differed had we chosen other outlets (for example, the Frankfurter Allgemeine Zeitung instead of Die Welt in Germany). Still, most findings were reproducible when excluding Süddeutsche Zeitung, the outlet publishing almost half of the included articles, andfor country comparisonthe functionally equivalent TagesAnzeiger.
The article selection process also resulted in rather low case numbers, lowering statistical power for comparisons between pandemic phases, for example. Moreover, by using targeted search terms in Factiva, only articles that were directly using the wording Solidarität and Eigenverantwortung were identified as relevant in this study. It is possible that other aspects, for example, institutionalized forms of solidarity, would have been more salient with a more inclusive approach. We instead show how diffuse the words Solidarität and Eigenverantwortung were used in newspaper coverage. Also, for Switzerland, the findings are not generalizable to the French, Italian and Rhaeto-Romanic media landscape because culture and language impact the use of these two terms significantly.
Finally, we chose a quantitative approach comparing the amount of coverage and its aspects in Germany and German-speaking Switzerland. Including a codebook with predefined content categories bears the limitation that an in-depth, qualitative assessment of the article content is beyond this paper's methodological scope.

Conclusion
The terms solidarity and personal responsibility were used to justify and explain COVID-19 policy in newspaper coverage. Solidarity was used to justify the implementation of restrictive policies (particularly in Germany), and personal responsibility to explain the lack thereof. Our findings illustrate the relative importance of the term solidarity in the context of the COVID-19 pandemic. The term "solidarity" was used in various contexts, including symbolic declarations, calls for protective behavior, and to justify COVID-19 policy. The inflationary use of the term while not explicitly acknowledging and discussing the limits of solidarity is problematic, particularly in times of crisis: Acting in solidarity comes with costs that accumulate over time. Calling for the maintenance of solidarity practices without installing institutionalized systems of reciprocity might demotivate people, causing solidarity calls to lose their positive effects. Newspapers and policymakers alike should be aware of this when using the concept of solidarity in future crises and discuss and acknowledge explicitly the limit and exhaustion of solidarity as a practice. Relatedly, the role solidarity may take to provide reasoning in policymaking needs to be more explicitly defined in the future.
Ethical Statement for SSM population health 1) this material has not been published in whole or in part elsewhere; 2) the manuscript is not currently being considered for publication in another journal; 3) all authors have been personally and actively involved in substantive work leading to the manuscript, and will hold themselves jointly and individually responsible for its content.

Declaration of competing interest
The authors declare no conflicts of interest.

Data availability
The data used for analysis is provided as supplemental file

Table B2
Contexts in which solidarity were mentioned in newspaper coverage excluding Süddeutsche Zeitung and TagesAnzeiger (see Table 3 for all newspapers).

B.2Actors cited (H2)
Because no country comparison was involved in H2, only Süddeutsche Zeitung was excluded from this analysis to have higher case numbers and greater statistical power. Similar to the whole analysis, most-cited actors for articles about solidarity were the media (63/236, 21.4%), individual citizens (including reader's letters, 33/236, 14.0%), but regional politics was less well represented (16/236, 6.8%), and there was a higher ratio of celebrities and cultural actors (32/236, 13.6%, Table B. 3). For personal responsibility, the media (27/84, 32.1%) and national politics (20/84, 23.8%) were still the most cited actors, but regional politics (12/84, 14.3%) were again less prominent than in the whole analysis.

Table B3
Actors cited in newspaper coverage excluding Süddeutsche Zeitung (for table including all newspapers see Table 3).

B.3. Pandemic phases (H3, H4)
When testing for the independency of coverage in stages of relaxed versus stringent policies without Süddeutsche Zeitung, findings were confirmed that significantly more articles were published about solidarity during stringent phases (167/266, 62.8%, X2(d1,266) = 17.383, p < .001) and that there was no statistically significant difference for personal responsibility (X2(d1,96) = 0.042, p = .838). Proportions of article evaluations were not statistically significant but case numbers were very low, particularly for personal responsibility.

Table B4
No of articles and evaluations in phases of relaxed versus stringent COVID-19 policies (excluding Süddeutsche Zeitung; for analysis including all newspapers see Table 4).