The aftermath of Covid-19: The rise of pandemic animosity among consumers and its scale development

Abstract

Consumer animosity captures negative attitudes to foreign products and impacts willingness to buy them. Existing constructs nevertheless fail to account for an emerging dimension: pandemic animosity. This article heeds recent calls to develop a pandemic animosity measurement scale. Its purpose is to: (i) extend Klein et al.'s (1998) animosity model by adding the pandemic animosity construct, (ii) provide a measurement scale for pandemic animosity, and (iii) explain how pandemic animosity impacts consumers’ willingness to buy. Study 1 analyzes qualitative data from in-depth personal interviews with NVivo to identify themes and codes. An expert panel helped reach consensus of all indicators. Study 2 filters scale items using a pilot sample. Study 3 validates a 12-item scale with a larger representative sample. The results contribute to the consumer animosity literature by confirming the existence of pandemic animosity, providing an actionable measure, and confirming its negative impact on consumers’ willingness to buy.

1. Introduction

With the rapid rise in globalization, economic interdependencies, and migratory flows, businesses have a huge incentive to cross international borders to pursue new marketplaces (Deng and Wang, 2016, Guzmán et al., 2017, Luo, 2021, Surugiu and Surugiu, 2015). When choosing a marketplace, understanding consumers’ behavior to determine the correct marketing strategies is crucial (Farah, 2020); especially understanding what factors influence consumers’ evaluations and intentions to buy foreign products (Farah, 2020, Lee et al., 2017). Beyond business expansion opportunities, globalization poses challenges too. Actions taken in one country by governments, corporations, and media may not be well-perceived in another (Maher and Mady, 2010, Sandıkcı and Ekici, 2009). Similarly, the actions and response to natural disasters in one country (i.e., Covid-19 pandemic) may be scrutinized in another (Nawaz et al., 2020, Rojas-Méndez et al., 2022), potentially creating animosity between nations and negative consequences for businesses. In a worst-case scenario, this animosity may lead to consumers boycotting products/services from the offending country as a form of protest or to express their disapproval (Braunsberger and Buckler, 2011, Ettenson and Klein, 2005, Farah and Newman, 2010, Zdravkovic et al., 2021).

Originally developed by Klein et al. (1998), and later extended by Kalliny et al. (2017), the consumer animosity model is comprised of the following dimensions: war animosity, economic animosity, political animosity, and cultural and religious animosity. Nevertheless, due to the Covid-19 pandemic, a new type of animosity emerged (see Appendix-I for definitions). Pandemic animosity describes the remnants of hostility and antipathy related to a pandemic. It can be understood as antipathy towards another country due to its alleged responsibility for causing and spreading the pandemic (Rojas-Méndez et al., 2022). This paper focuses on the recent pandemic of Covid-19. Because the first case of Covid-19 was detected in China, the pandemic unleashed stigmatization and pandemic hate, especially against the East Asian and Chinese diaspora (Ng, 2020, Wang et al., 2021), which turned into pandemic animosity across nations worldwide (Nawaz et al., 2020). Xenophobia, racism, and prejudice in the form of hostility, suspicion, and fear towards Chinese people and people arriving from China were recorded in multiple countries (CGTN, 2020, Halepas and Ferneini, 2020, Jakovljevic et al., 2020, Nawaz et al., 2020, Ng, 2020, Wang et al., 2021). Similarly, the Omicron variant, first detected in Botswana, triggered animosity towards South Africa after the country announced a surge of cases (Mueller & Walsh, 2021). Nevertheless, this is not a new phenomenon; pandemic animosity towards the outbreak and country-of-origin of previous pandemics was recorded too: SARS-2003 in China (Timber & Chiu, 2020), MERS-2012 in the Middle-East (Hyun, 2020), and Ebola-2014 in West Africa (Zurcher, 2014).

Covid-19 animosity has resulted in a massive geopolitical disturbance that is restructuring the international systems (Carracedo et al., 2021, Sharma et al., 2020, Verma and Gustafsson, 2020), accelerating the process of de-globalization and de-regionalization, and establishing some new trade and political alignments around the world (Dareini, 2020, Paul and Dhir, 2021). Pandemic animosity thus becomes comparable in terms of economic impact to that of political, economic, war, and cultural and religious animosities (Ong et al., 2021, Reinicke, 2021, Padhan and Prabheesh, 2021). As animosity cannot be controlled by managers and can have multiple sources, researchers are called to investigate the different dimensions and impacts of consumer animosity in a globalized world (Farah and Mehdi, 2021, Klein et al., 1998, Shoham et al., 2016).

In line with Klein et al.’s (1998) seminal study, recent research finds that consumer brand engagement in the country of market declines due to consumer animosity for global brands from the country of origin due to the Covid-19 pandemic (Wang et al., 2021). Likewise, it negatively affects purchase intentions and travel intentions (Rojas-Méndez et al., 2022). Given that with the exception of Rojas-Méndez et al.’s (2022), and a short scale on animosity towards the Chinese government by Krüger et al. (2020), no measurement scale is available to evaluate the impact of pandemic animosity on consumer perceptions and attitudes, this paper sets to close this gap: understanding how pandemic animosity affects consumer behavior and adding this emerging dimension to the animosity model from a Covid-19 perspective. First, it examines pandemic animosity at a consumer-level and provides a clear conceptualization of the construct. Second, it identifies the pandemic animosity factor’s relevance based on in-depth interviews. Third, it adds the dimension to the animosity measurement model (Kalliny et al., 2017, Su et al., 2020). Finally, it tests and validates the factor structure of the proposed model and examines the effect of pandemic animosity on the outcome variable, consumers’ willingness to buy. The study contributes to the literature by: (i) enhancing academic knowledge by examining the notion of pandemic animosity, (ii) developing a measurement instrument to spur further research on the topic, and (iii) providing guidance to practitioners and corporations for understanding the effects of pandemic animosity and improving their decision making. The overview of the research process is presented in Table 1 .

Table 1.

Overview of the research process.

Research phases	Main objectives	Results
Phase 1: Literature review	Identify pandemic animosity	Congealing pandemic dimension evidence
Phase 2: Interviews	Identify themes and codes	Recorded 29 interviews worldwide
Phase 3: Expert panel	Identify scale items	Total identified items 20 & eliminated 8 items (inclusive)
Phase 4: Scale development	Develop instrument	Pilot study of 132 respondents
Phase 5: Scale validation	Validate instrument	Used 792 respondents worldwide

2. Theoretical background and conceptual development

2.1. Animosity model and pandemic animosity

Tensions among nations, either involving territorial, economic, race, religious, or ideological conflicts, are evident worldwide, and may generate animosity between nations and increase nationalistic sentiments (De Nisco et al., 2016). Investigating such disputes and their impact on consumers’ behavior has gained significant attention (Abraham and Poria, 2020, Abraham and Reitman, 2018, Kalliny et al., 2017). Klein et al. (1998) first investigated the impact of tensions between nations on consumers’ buying behavior, and defined animosity as the remnants of antipathy related to previous or ongoing military, political, or economic events. Animosity hurts consumers’ willingness to buy from the offending country but does not distort their quality evaluations (Klein et al., 1998). Therefore, consumers harboring animosity acknowledge the quality of products originating from the offending nation but tend to refuse buying the products (Kalliny et al., 2017). This hostility might stem from historic as well as current events and has an important impact on consumer behavior, including choice of products and willingness to pay for them (Riefler & Diamantopoulos, 2007; Shimp et al., 2004).

Animosity is different from consumer ethnocentrism, which represents unwillingness to purchase any foreign product without negative country-specific attitudes (Klein et al., 1998). In contrast to country-of-origin effects (e.g., Gürhan-Canli & Maheswaran, 2000), animosity does not affect the perception of quality of products but has a direct impact on purchasing decisions (Klein et al., 1998). Recent studies confirm that consumer animosity and ethnocentrism are two distinct concepts (Lee et al. 2017; Riefler & Diamantopoulos, 2007; Shoham et al. 2016). In fact, animosity can stem from (regional) stereotypes and social identity, by contraposing oneself to the disliked society, region, or country (Shimp et al., 2004). Apart from disliking, animosity implies open hostility, ill will, displeasure, and even enmity. Factors such as traditionalism, patriotism, and nationalism (in contrast to internationalism) predict consumer animosity (Tian & Pasadeos, 2012). Multiple social attributes, such as dogmatism, exclusionism, and other cultural orientations also contribute to consumer animosity, while cosmopolitanism is negatively related to it (cf. Leonidou et al., 2019). Personal implications (e.g., losing a loved one to a war or to a pandemic) (Tian & Pasadeos, 2012) and attributing blame on the country-of-origin (Zdravkovic et al., 2021) further magnify animosity feelings.

Fundamental differences exist among the constructs of war, economic, and political animosity developed by Klein et al. (1998), and the constructs of religious and cultural animosity established by Kalliny et al. (2017) and supported by Alvarez et al., 2020, Souiden et al., 2018. Researchers have thus attempted to extend the animosity model to include other situations and dimensions, such as regional animosity (Shimp et al., 2004), implicit animosity (Cai et al., 2012), and people-related and physical environments (Perviz et al., 2014) (see Appendix-II). Furthermore, animosity can be classified as situational or stable (De Nisco et al., 2016). Situational animosity impacts a particular moment, whereas stable animosity evolves into a deeper and long-lasting aggression that accumulates. Cultural and religious animosities are stable and difficult to eradicate and manage. War, economic, and political animosities are usually situational and can normally be resolved over time. Nevertheless, although animosities are not supposed to be stable, as stability predicts more consumer ethnocentrism (Lee et al., 2017), stable animosities are easy to ignite (Jung et al., 2002).

The analysis of the existing literature allows to posit the following: (1) consumer animosity literature is scarce and requires more attention (Farah and Mehdi, 2021, Shoham et al., 2016); (2) consumer animosity has a great impact on consumer behavior (Perviz et al., 2014, Shoham et al., 2016); (3) consumer animosity is a distinct concept, related but not substitutable by consumer ethnocentrism (Lee et al. 2017; Riefler & Diamantopoulos, 2007; Shoham et al. 2016) (cf. Fig. 1 ); (4) consumer animosity is temporary rather than stable and needs to account for current events that might cause the tensions (Campo and Alvarez, 2021, Hoang et al., 2022, Lee et al., 2017); (5) new studies should extend measurement models to account for new roots of animosity (Farah and Mehdi, 2021, Riefler and Diamantopoulos, 2007). To respond to these theoretical drawbacks, this study focuses on one of the most prominent causes of animosity in recent times: the Covid-19 pandemic (Campo & Alvarez, 2021).

Fig. 1.

The placement and relevance of pandemic animosity.

This study focuses on pandemic animosity that can be considered situational, but that despite its temporal effect still requires attention since animosity is always somewhat temporary (Ettenson and Klein, 2005, Lee et al., 2017). As the Covid-19 pandemic prolongs, it continues to have considerable impact on the whole planet and could become an underpinned implicit association with its country of origin and as a result become more stable (Cai et al., 2012).

2.2. Evidence of pandemic animosity

The Covid-19 pandemic is highly complex in origin, spread, effects, and consequences at medical, social, cultural, political, economic, and religious levels (Jakovljevic et al., 2020). It has affected companies’ ability to export their products to many countries (Verschuur et al., 2021, Walmsley et al., 2021). For example, the tougher rules imposed by the U.S. administration following the Covid-19 pandemic limited Chinese exports worldwide, especially medical exports (Oxford Analytica, 2020). Furthermore, Covid-19 triggered pandemic-hate and dehumanization within the country and against immigrant communities of Chinese and East-Asians around the world (CGTN, 2020, Nawaz et al., 2020, Xu et al., 2021). The distrust and hatred toward Chinese and East-Asians also contributed to a well-documented number of foul incidents such as violence, verbal attacks, and boycotts of Chinese businesses (Dareini, 2020, Tessler et al., 2020). Russell Jeung, chair of the Asian-American studies department at San Francisco State University, said: “when it becomes normalized… it dehumanizes Chinese and Asians… following xenophobic policies or statements, the next week we would see boycotts of Asian businesses and then the following week more interpersonal attacks on Asian individuals” (Timber & Chiu, 2020). Selected mainstream media and related social media threads propagated hatred against China and Chinese population (Han & Marwecki, 2020). Pandemic animosity is thus exhibiting parallel consequences to war and economic animosity, and it is important to understand its impact on foreign product buying intentions.

3. Study 1- interviews and expert panel

3.1. Methodology

The lack of theoretical guidelines and literature about pandemic animosity motivated the qualitative study design to develop the scale items. In-depth personal interviews were used to authenticate the proposed dimension and identify potential additional properties (Kim et al., 2018, Malhotra et al., 2006). The study followed standard interview guidelines (Kvale and Brinkmann, 2009, Qu and Dumay, 2011). By adopting a localist approach, semi-structured interviews were designed (Alvesson, 2003). This approach provides the benefits of flexibility, accessibility, intelligibility, effectiveness, convenience, and disclosure of important/hidden facets of behavior (Kvale and Brinkmann, 2009, Qu and Dumay, 2011). Interviews were scheduled during the peak of Covid-19 restrictions, between May and June 2020. Specifically, 29 in-depth personal interviews were conducted (17 via video-call, 7 via audio-call, and 5 via email) using stratified purposive sampling (Kalliny et al., 2017). A mixed contact approach was used due to Covid-19 lockdown measures. On average, each interview lasted 54 min. The study aimed for representation from all continents and included people affected by Covid-19. Informants include international scholars, professional employees, business people, and one person that stays at home. Both genders and different age groups were included.

Concerning the demographic characteristics, the inclusion of all continents was ensured to minimize any continent or country specific bias and get true generalized feelings about pandemic animosity. The diversity of cultural backgrounds (i.e., people from China, neighboring countries, and occidental countries) and of various countries of residence with Covid-19 related restrictions, further allowed to minimize any country-specific bias and elevate the risk of other type of animosities (e.g., war animosity with Japanese, economic and political animosities with the US). The distribution between various age, gender, and occupation groups equally contributed to a more general understanding of pandemic animosity. Nevertheless, the sample was relatively young (32 years old average) and mostly employed (28 out of 29), therefore risking lower historically related biases yet more occupation related concerns (see Table 2 for participants’ characteristics).

Table 2.

Interview participants characteristics.

#.	Gender	Belonging continent	Nationality country	Stay during pandemic	Respondent background	Age	Covid-19 experience	Type of contact	Time duration
1	Female	N. America	USA	USA	Postgraduate	21	No	Video Call	54 Minutes
2	Male	N. America	USA	USA	Businessman	45	Yes	E-mail	N/A
3	Male	S. America	Venezuela	China	PhD Scholar	32	No	Video Call	35 Minutes
4	Female	S. America	Granada	China	PhD Scholar	33	No	Video Call	50 Minutes
5	Female	N. America	Canada	Canada	Graduate	22	No	E-mail	N/A
6	Female	N. America	Canada	Canada	Graduate	20	No	E-mail	N/A
7	Female	Europe	England	3rd Country	Teacher	37	Yes	Video Call	67 Minutes
8	Male	Europe	England	England	Professor	51	No	Video Call	45 Minutes
9	Male	Europe	France	China	PhD Scholar	35	No	Audio Call	72 Minutes
10	Female	Europe	Hungary	Hungary	Employed	26	No	Video Call	56 Minutes
11	Female	Asia	Russia	Russia	Postgraduate	28	No	Video Call	65 Minutes
12	Male	Asia	China	China	Businessman	58	No	Audio Call	42 Minutes
13	Female	Asia	China	China	Lawyer	23	No	Audio Call	52 Minutes
14	Male	Asia	Iran	Iran	Professor	52	Yes	Video Call	57 Minutes
15	Female	Asia	UAE	UAE	House wife	30	No	E-mail	N/A
16	Male	Asia	Pakistan	Pakistan	Self-employed	35	No	Audio Call	63 Minutes
17	Male	Asia	Pakistan	Pakistan	Professor	45	Yes	Audio Call	45 Minutes
18	Male	Asia	India	China	PhD Scholar	26	No	Video Call	53 Minutes
19	Male	Asia	India	3rd Country	Medical Doctor	36	Yes	E-mail	N/A
20	Male	Asia	Nepal	Nepal	PhD Scholar	34	Yes	Audio Call	67 Minutes
21	Female	Asia	Japan	China	Graduate	23	No	Video Call	55 Minutes
22	Female	Asia	South Korea	South Korea	Graduate	22	No	Video Call	71 Minutes
23	Female	Asia	Indonesia	China	Employed	32	Yes	Audio Call	51 Minutes
24	Male	Australia	New Zealand	China	Graduate	23	No	Video Call	43 Minutes
25	Male	Australia	Australia	China	Postgraduate	26	No	Video Call	35 Minutes
26	Female	Africa	Algeria	China	Postgraduate	29	No	Video Call	56 Minutes
27	Female	Africa	Morocco	China	PhD Scholar	27	No	Video Call	64 Minutes
28	Female	Africa	Cameron	China	PhD Scholar	27	No	Video Call	45 Minutes
29	Male	Africa	Ethiopia	China	PhD Scholar	33	No	Video Call	48 Minutes

The interviews were analyzed using NVivo and a pool of expert panel (details provided in the next section) (Saldaña, 2021). A research design using a mix of traditional tools and advanced software like NVivo was used to interpret the interview data, to maximize data interaction while ensuring a rigorous and productive analysis process (Maher et al., 2018)—leveraging researchers’ imaginative insights to make sense of the data combined past literature knowledge and software tools (i.e., NVivo) to fully scaffold the analysis process. Following standard procedures, interview data and pandemic literature guided the generation of the codes and themes; cross-coding, generated through the NVivo and expert panel, was applied to reach agreements (Glaser and Strauss, 2017, Kalliny et al., 2017, Su et al., 2020).

3.2. Interview analysis

Interviewees were probed about different aspects of the pandemic, and responses were recorded, transcribed, and analyzed. The analysis confirms that the pandemic has divided the world into three narratives. The foremost and larger group favors the U.S. stance and blames China. The second and medium group represents those in favor of China, and blames the U.S. The third and smaller group, represents neutral arguments, believes that the virus is natural, and blames no one. Moreover, some respondents (from all groups) believe that Covid-19 is a work of non-state actors, such as pharmaceutical companies. Nevertheless, the larger group of informants, including those from the smaller group, feel animosity against China given the pandemic’s consequences: social distancing, lockdowns, and safety measures. The majority of participants (from all groups) expressed frustration, sadness, hopelessness, and worry about the pandemic effects on humans and the economy. All recognized that the information they had was based on their personal experiences and media sources they consult.

The majority of participants (from all groups) criticized the mistreatment of the Chinese, albeit every informant believed the virus did originate in China. Officially or unofficially calling the virus a “Chinese virus” or “Kungflu” was not deemed justifiable and considered an act of racism. The majority of informants (from all groups) also criticized Chinese habits of eating wild animals, which could have been the source of Covid-19. Whereas some respondents (from all groups) appreciated China’s political system in implementing the control measures that led to a quick recovery from the pandemic, some did not. The medium and smaller group believed that China had efficiently managed the outbreak situation within China and that the spread of Covid-19 was not due to Chinese mismanagement, but because of the mismanagement of their own governments and policies, and the delay in implementation of preventive measures. While the larger group placed all responsibility on China, and believed that China failed to manage the situation and rather tried to hide the outbreak information. All respondents, nevertheless, considered it a “deadly global disaster” in the history of mankind, and believe the pandemic controlling measures did more damage than the virus itself.

3.3. Expert panel and NVivo analysis

To make the process easy and understandable, it was divided into five different stages and illustrated through an example from the data (e.g., PA5: I could never forgive China for the human and economic loss in my country during Covid-19 pandemic) (see Fig. 2 for process details). In stage-1, 29 semi-structured qualitative interviews were conducted and raw data was recorded and transcribed. In stage-2, an expert panel was convened to perform a content analysis of the interview transcripts. The respondents’ statements were systematically analyzed to identify codes, categories, and themes, and consequently scale items, from the data. Using investigator triangulation (Natow, 2020), the expert panel consisting of 3 marketing professors and 5 Ph.D. scholars (3 Chinese and 5 from other nationalities) read the transcripts multiple times to gain familiarity with the content (Tesch, 1990). Each response was read word-by-word to identify text that captured the vital concepts. Peer-to-peer discussion and reflection of the analysis process allows researchers to constantly compare, trial, view different perspectives, develop meaningful insights from micro-to-macro level (Maher et al., 2018). The analysis of the qualitative data and mutual discussion between researchers resulted in the initial coding. The focus of the panel was to understand each response under the light of Covid-19 and pandemic animosity feelings.

Fig. 2.

Qualitative data analysis process.

The initial coding resulted in 436 codes after coding or labeling each response (codes related to the example—i.e., the deadly pandemic, deaths everywhere, economic issues, unemployment, economic crisis, economic destruction, human and economic loss, economic war, economic agenda behind virus, US-China economic war, future uncertainty, sad situation, people are dying, miserable situation, need economic support from government, poor economic conditions, lost my father, people lost jobs, etc.). Then the related codes were used to form emergent categories—categories related to the example (i.e., human deaths, economic issues, unemployment, future uncertainty). Thus, the categories generated were grounded on the interviewees’ unique perceptions. To ensure reliability, a classification sample of 18 responses was initially evaluated. A test–retest was conducted where each judge independently created exhaustive and mutually exclusive response categories.

In stage-3 the data obtained through interviews were analyzed using NVivo, which complements working parallel to other coding methods while allowing data management facilitation and answering complex questions (Maher et al., 2018). Data keywords, codes and themes were generated by the software and shared with the expert panel to compare and refine their previous work. In stage-4 the judges were asked to cross-match the codes and themes of NVivo with their initial work and to classify the responses into the developed scale items in three weeks independently. This slowed down the process but allowed for a meaningful interaction with data (Maher et al., 2018). This process refined the repetition of codes generated by the panel (codes refined for the example, i.e., human loss, economic loss, sad situation). In stage-5 each judge scrutinized the initial data with NVivo data and created themes (theme for the example—i.e., human deaths and economic losses). This independent creation of themes helped the judges to phrase the scale items. After each judge exceeded a 0.800 cut-off (Keaveney, 1995), they compared their independently created categorization schemas to create a single consistent scale item system (scale item finalized for the example—i.e., “I could never forgive ‘China’ for human and economic loss in my country during Covid-19 pandemic”) (Su et al., 2020).

3.4. Results

Participants defined pandemic animosity using the following keywords: anger, hatred, dislike, and ill feelings. Respondents stated that the pandemic is spreading animosity as it is generating confrontation rather than cooperation driven by the human losses, economic crisis, and lack of global collaboration. The interviewees’ qualitative data derived into 20 different descriptive items, but further review by the expert panel identified 8 repetitive items that were excluded. The remaining 12 scale items are presented in Appendix-III. The items generated are consistent with the presented theoretical framework.

4. Study 2- scale refinement

4.1. Methodology

Study 2 collected survey data for pilot testing and validating the scale items of pandemic animosity, along with the other dimensions of consumer animosity. Simple definitions of consumer and pandemic animosity were provided. All items were adapted to the context (except pandemic animosity) and measured using a seven-point Likert scale. A screening question to identify Chinese respondents was added to exclude them. Additionally, basic demographic data was collected (i.e., gender, age, education, professional status, nationality, and stay during the pandemic). The questionnaire was developed on “Microsoft Forms”. The online link was shared on WeChat, WhatsApp, and Facebook to foreigners using stratified snowball sampling techniques. The researchers requested international academics, Ph.D. scholars of a northeast Chinese university, to help collect responses from their country residents from different social strata especially those who were living in their home countries (see Table 3 for details). Further, the anonymity of the responses helped reduce response bias (Nunnally & Bernstein, 1994). 132 responses were collected from foreigners residing inside or outside China during the month of August 2020.

Table 3.

Samples’ demographic characteristics.

Characteristics	Criterion	Study–1 (n29)	Study–2 (n132)	Study–3 (n792)
Gender	Male	14	79	456
	Female	15	53	336
Age	Below 20	00	18	122
	21 to 30	15	49	335
	31 to 40	09	37	159
	41 to 50	02	18	123
	Above 50	03	10	053
Education	Graduate/under	09	44	377
	Postgraduate	12	62	342
	PhD scholars	08	26	073
Professional status	Student	17	78	437
	Employed	09	35	217
	Businessman	02	15	126
	Housewife	01	04	012
Stay during pandemic	In China	14	56	335
	In home country	13	69	421
	In third country	02	07	036
Belonging	Asia	13	35	205
continent	(Japan, South Korea, Russia,
(Responding countries)	India, UAE)
	North America	4	21	173
	(USA, Canada)
	South America	2	18	108
	(Brazil, Venezuela, Granada)
	Europe	4	25	97
	(UK, France, Hungry)
	Australia	2	14	75
	(Australia, New Zealand)
	Africa	4	19	134
	(Eritrea, Cameron, Algeria,
	Ethiopia, Egypt, Morocco)

The pilot survey included people of different gender, age, education, and occupation groups to better understand pandemic animosity feelings and opinions. Inclusion of all continents allowed to minimize continent or country specific bias. The country of stay during the pandemic (i.e., in China, home country, or 3rd country) helped capture mixed feelings about Covid-19 related restrictions (see Table 3 for participants’ characteristics).

4.2. Scale assessment

Descriptive statistics were inspected to remove items displaying unsatisfactory psychometric properties. Three items (PA1, PA4, and PA6) had a mean below the midpoint of 4 and thus excluded. Item-to-total correlations were calculated for all remaining items and remained above 0.4 (Kim et al., 2012). Multivariate normality was assessed to determine whether structured equation modeling (SEM) assumptions were satisfied. The findings indicate absolute values of univariate skewness below 2, and absolute values of univariate kurtosis below 3, suggesting that the data does not deviate from a normal distribution (Kline, 2015, Su et al., 2020). Two items (PA11, PA12) were deleted given their factor loadings below 0.650. Cronbach alpha’s values ranged from 0.870 to 0.913, and composite reliability values ranged from 0.841 to 0.925, suggesting satisfactory scale reliabilities (Fornell and Larcker, 1981, Hair et al., 2019). Average Variance Extracted (AVE) values ranged from 0.586 to 0.882, proving convergent validity (Hair et al., 2019). The square roots of the AVE ranged from 0.765 to 0.931, which relative to the correlation coefficients of the constructs (0.560 to 0.834) were all greater than the correlation coefficients, proving discriminant validity. Variance inflation factor (VIF) values remained below the threshold value of 5 or more strict 3, indicating no common method bias or multicollinearity (Chin, 1998, Hair et al., 2011). The resulting pandemic animosity scale was validated in study 3 (see Table 4 ).

Table 4.

Descriptive statistics for items of pandemic animosity.

Item code	Description	Mean	Standard Deviation	Skewness	Kurtosis
PA1.	‘Chinese’ people scare me due to ‘Covid-19 pandemic’.	3.545	1.924	−1.160	0.278
PA2.	I don’t like ‘China’ due to ‘Covid-19 pandemic’.	4.689	1.661	-0.779	-0.227
PA3.	I will not forgive ‘China’ for damages during ‘Covid-19 pandemic’ a deadly global disaster.	4.144	1.711	-0.901	-0.006
PA4.	I am very angry at ‘China’ due to ‘Covid-19 pandemic’.	3.818	1.910	−1.055	0.191
PA5.	I could never forgive ‘China’ for human and economic loss in my country during ‘Covid-19 pandemic’.	5.174	1.579	0.176	-0.829
PA6.	I hate people from ‘China’ due to ‘Covid-19 pandemic’.	3.561	1.943	−1.077	0.305
PA7.	I would not like to meet people from ‘China’ which is origin of ‘Covid-19 pandemic’.	4.326	1.734	-0.609	-0.346
PA8.	I would never forgive ‘China’ because of the suffering during ‘Covid-19 pandemic’.	4.811	1.286	0.666	-0.656
PA9.	I believe ‘Covid-19′ is ‘Chinese’ virus.	4.659	1.481	-0.031	-0.612
PA10.	I believe ‘China’ is responsible for ‘Covid-19 pandemic’ spread in world.	4.879	1.255	0.976	-0.628
PA11.	I think ‘China’ is responsible for hiding the information regarding ‘Covid-19 pandemic’ outbreak.	4.705	1.313	0.158	-0.393
PA12.	The eating practices of ‘Chinese’ are crazy because they eat wild animals (Which is the possible source of coronavirus).	4.902	1.375	0.631	-0.669

Note. ‘China’/ ‘Chinese’/ ‘Covid-19 pandemic’ in italics represents the context specification, which could be replaced with other countries for their context specification or other health contexts.

5. Study 3- validation study

The objectives of study 3 were: (1) to validate the animosity model via a replication model (Su et al., 2020), especially the pandemic animosity dimension, (2) to examine the generalizability, validity and reliability of the tool, and (3) to test the factor structure fit in a nomological arrangement. Past literature reveals that willingness to buy (Abraham & Reitman, 2018) is a valuable construct to assess the nomological validity of the emerging pandemic animosity scale.

Past animosity literature focuses on its multiple origins (De Nisco et al., 2016, Park and Yoon, 2017, Tabassi et al., 2012), such as wars (Kalliny et al., 2017, Klein et al., 1998), economic and political arguments (Ettenson and Klein, 2005, Klein, 2002), and cultural and religious differences (Kalliny et al., 2017). As previously mentioned, pandemic animosity has recently played a critical role in political and trade relations (Jakovljevic et al., 2020), and in creating adverse consequences for the country of origin effect on Chinese products (Jakovljevic et al., 2020, Timber and Chiu, 2020).

Past research establishes that consumer animosity towards a particular country exerts negative impacts on buying and consumption behavior (Ettenson and Klein, 2005, Kalliny et al., 2017). Animosity directly and negatively affects consumers’ willingness to buy products of the offending country in the short and long run (Abraham and Reitman, 2018, Shimp et al., 2004, Shoham and Gavish, 2016). All dimensions of consumer animosity, except pandemic animosity, are well researched. However, there is empirical evidence that Covid-19 has negatively affected consumers’ willingness to purchase Chinese and China-related products and services (Rosenbaum and Russell-Bennett, 2020, Wei et al., 2021). Even inside the country, Wuhan products became a victim of stigmatization (Hao & Wang, 2021). While these studies suggested that the lower intent to purchase Chinese and China-related products (e.g., going out to a Chinese restaurant in the US) can come from stigmatization over China’s role in Covid-19 pandemic, they have not investigated pandemic animosity as such. This research thus adds the pandemic animosity dimension and posits that, within a Covid-19 pandemic context, consumer animosity negatively affects consumer’s willingness to buy Chinese products. Formally stated:

Hypothesis: Consumer animosity (economic, political, war, cultural, and pandemic) negatively influences consumers’ willingness to buy Chinese products within a Covid-19 context.

5.1. Methodology

To validate the constructs and enhance the external validity of the pandemic animosity scale, as suggested by Churchill Jr (1979) and in line with Kalliny et al., 2017, Su et al., 2020, study 3 recruited a second sample to help reduce the possible single sample bias (Kalliny et al., 2017). Study 3 tests the complete proposed model of consumer animosity, using a common construct from past consumer animosity studies, willingness to buy, via a replication model method. The constructs were adapted from past literature and few additional items were included to capture demographic information. Consumer animosity was measured through a single item animosity scale (to gauge the overall animosity feeling of the respondent) (Rose et al., 2009) and five dimensions; i) economic, political, and war animosity (Klein et al., 1998), cultural animosity (Kalliny et al., 2017), and pandemic animosity (self-generated items). Willingness to buy was adapted from Kalliny et al., 2017, Darling and Arnold, 1988. All scale items were measured using a seven-point Likert scale. Questionnaire online links, generated on “Microsoft forms,” were distributed among foreigners worldwide (using convenience and snowball sampling techniques). Social media applications (WhatsApp, Facebook, Twitter, Instagram, and WeChat) were used to post the questionnaire links every-three days during September and October 2021. The survey instrument was administered to as many possible foreigner participants, representing various ages, ethnicity, educational levels, occupations, and cultural affiliations around the world. A screening question (Chinese/others) was also included. Responses from Chinese respondents or with missing information were removed, resulting in a final sample size of 792. All the necessary measures were taken to ensure anonymity and reduce response bias as recommended by Nunnally and Bernstein (1994).

5.2. Results

The demographic characteristics were taken as control variables (see Table 4). All PLS-SEM measurement and structural model steps were followed as suggested by (Hair et al., 2019, Sarstedt et al., 2022). Because the proposed model includes consumer animosity—a higher-order construct containing five lower-order dimensions—making the model of a reflective-formative type, all the analysis was developed accordingly (Becker et al., 2022). The research applied disjoint two-stage approach, which further allows the use of PLS_predict analysis to assess the out-of-sample predictive power at indicators level (Becker et al., 2022, Shmueli et al., 2019). The disjoint approach involves two steps. In the first step, the measurement model of all lower-order constructs should be validated applying standard evaluation criterion (Hair et al., 2017). When lower-order constructs are validated, the second step includes the evaluation of the higher-order model as a whole as per standard evaluation criterions (Becker et al., 2022).

5.2.1. Measurement model

Factor loadings for all lower-order constructs items exceed the 0.70 threshold level, thus proving their reliability. Similarly, internal consistency reliability was measured using Cronbach’s alpha, rho_A, and composite reliability. Values for all these measures are above the threshold of 0.70 (Sarstedt et al., 2022, Hair et al., 2019, Nunnally and Bernstein, 1994). Convergent validity was assessed using the AVE and the values remained above the 0.50 threshold (Sarstedt et al., 2022, Hair et al., 2019, Bagozzi and Yi, 1988). Recent advances in PLS-SEM consider HTMT criterion a standard way to assess discriminant and give it a priority over the Fornell-Larcker criterion (Franke and Sarstedt, 2019, Henseler et al., 2015). The HTMT values obtained are considered good if significantly lower than the threshold of 0.85 (for conceptually different constructs) and 0.90 (for conceptually similar constructs) (Hair et al., 2021; Sarstedt et al., 2022). In PLS-SEM, the VIF values are used to assess multicollinearity and common method bias (CMB) issues. All values are below the threshold value of 5 or more strict criterion of 3, indicating no multicollinearity or CMB issues (Hair et al., 2011, Sarstedt et al., 2022). Further, Harman’s test was used for CMB, and the unrotated factor structure shows no dominating factor with more than 50 % variance in the sample (Aguirre-Urreta and Hu, 2019, Chakraborty et al., 2021, Dash and Paul, 2021, Podsakoff et al., 2003). (See Table 5, Table 6 ).

Table 5.

Reliabilities, validities & VIF.

Constructs	IC	FL	α	rho_A	CR	AVE	VIF
Consumer animosity
Economic animosity			0.917	0.919	0.922	0.822
China is not a reliable trading partner.	ECA1	0.931					2.025
China wants to gain economic power over other country(s).	ECA2	0.937					1.984
China has too much economic influence in other country(s).	ECA3	0.947					2.563
Chinese are doing unfair business practices in other country(s).	ECA4	0.900					2.001
Political animosity			0.870	0.880	0.920	0.794
Chinese government implements unfair policies against others.	POA1	0.833					1.828
Chinese government distorts history to make China look good and others wrong.	POA2	0.915					2.889
Chinese government should admit its wrong policies against others.	POA3	0.923					3.182
War animosity			0.877	0.882	0.924	0.804
I feel angry toward China because of wars with other country(s).	WAA1	0.834					1.782
I will not forgive China for its attacks on other country(s).	WAA2	0.920					2.534
China should pay for what it did to other country(s).	WAA3	0.932					3.721
Cultural animosity			0.917	0.919	0.922	0.836
Chinese are responsible for all that is wrong with the world.	CUA1	0.946					2.358
I dislike people from Chinese culture.	CUA2	0.951					2.042
I would not like to meet people from Chinese culture.	CUA3	0.942					2.334
People from Chinese culture represent all that is bad.	CUA4	0.926					2.278
Pandemic animosity			0.896	0.903	0.909	0.698
I do not like China due to the Covid-19 pandemic.	PA1	0.648					2.094
I will not forgive China for my mental and physical health damages during the Covid-19 pandemic.	PA2	0.824					2.518
I could never forgive China for human and economic loss in my country during the Covid-19 pandemic.	PA3	0.881					3.657
I would not like to meet people from China, which is the origin of the Covid-19 pandemic.	PA4	0.883					1.170
I would never forgive China because of the suffering during the Covid-19 pandemic.	PA5	0.862					2.100
I believe Covid-19 is a Chinese virus.	PA6	0.855					1.752
I believe China is responsible for the Covid-19 pandemic spread in the world.	PA7	0.868					2.900
Willingness to buy			0.892	0.895	0.913	0.824
I would feel good if I bought Chinese products.	WB1	0.921					2.734
I like to buy Chinese products.	WB2	0.939					2.106
Whenever available, I prefer to buy products made in China.	WB3	0.861					1.992

Note1. The country and Covid-19 pandemic name in italics can be substituted with the animosity target country believed to be the source of any particular pandemic or other health context.

Note2. IC = Item code; FL = Factor loadings; α = Cronbach alpha; CR = Composite reliability; AVE = Average variance extracted; VIF = Variance inflation factor.

Table 6.

Discriminant validity (HTMT criterion).

Constructs	1	2	3	4	5	6
1. Cultural animosity
2. Economic animosity	0.731
3. Willingness to buy	0.680	0.692
4. Pandemic animosity	0.714	0.673	0.614
5. Political animosity	0.546	0.528	0.540	0.582
6. War animosity	0.708	0.652	0.661	0.671	0.575

Note. The shaded diagonal blocks a standard way of HTMT values representation.

The study contained consumer animosity, a higher-order construct comprised of five lower-order constructs (dimensions): war, economic, political, cultural, and pandemic animosities. Given that the Chinese do not follow any religion, the study excluded religious animosity but considered cultural animosity. Nevertheless, there is possibly an overlap between culture and religion (Kalliny et al., 2017). All dimensions are attached to consumer animosity in a formative manner, making it a reflective-formative type model. Using the disjoint approach, the details of the statistical parameters are provided and proved (i.e., redundancy analysis, VIF values, outer weights and outer loadings) (Becker et al., 2022; Hair et al., 2021). For redundancy analysis the reflectively measured single item construct of consumer animosity, which captures the essence of the whole construct, was used (Cheah et al 2018). (See Table 7 ).

Table 7.

Higher-order construct validity parameters.

Higher-Order-construct	Lower-Order-construct	Redundancy analysis (path coefficient)	Inner VIFs	Outer Weights	T stats	P values	Outer loadings
Consumer animosity	ECA	0.858	2.824	0.223	6.756	0.000	0.888
	POA		1.452	0.037	2.371	0.009	0.615
	CUA		2.113	0.079	2.005	0.022	0.875
	WAA		2.831	0.076	2.989	0.001	0.817
	PA		2.595	0.697	14.827	0.000	0.889

Note. ECA = Economic animosity; POA = Political animosity; CUA = Cultural animosity; WAA = War animosity; PA = Pandemic animosity.

5.2.2. Structural model

The explanatory power of the model is assessed using coefficient of determination (R²). The R² measures the variance of each endogenous variable and the in-sample explanatory power of the research model (Shmueli & Koppius, 2011). R² values of 0.75, 0.50, and 0.25 are considered good, average, and weak, respectively (Hair et al., 2011). The obtained R² value is 0.771, indicating the model’s high explanatory power. The effect size (f2) is somehow redundant to the size of the path coefficients (Geisser, 1974, Stone, 1974). The threshold values for f² are 0.02 for small, 0.15 for medium, and 0.35 for large effect sizes (Cohen, 2013). The obtained value of f² is 0.489, indicating a large effect size. Researchers argue that a model with a certain level of explanatory power may produce varying levels of predictive power, thus the research should apply PLS_predict, a process comprised of out-of-sample prediction (Sarstedt and Danks, 2022, Shmueli et al., 2016). The application assesses the model on a training sample and evaluates its predictive performance on a hold-out-sample (Shmueli et al., 2019). All indicator level (i.e., WB1, WB2 and WB3) values of RMSE and MAE obtained from PLS model are smaller than the values of LM model, thus, authenticating the predictive power of the current model (Becker et al., 2022, Shmueli et al., 2019). (See Table 8 ).

Table 8.

R², f² & PLS_predict.

Criterion	CA	WB	PLS Model			LM Model
			RMSE	MAE	Q2predict	RMSE	MAE	Q2predict
R2		0.771
f2	0.489
PLSpredict
		WB1	1.062	0.802	0.639	1.072	0.810	0.632
		WB2	0.871	0.673	0.742	0.876	0.685	0.739
		WB3	1.029	0.766	0.750	1.040	0.795	0.745

Note. CA = Consumer animosity; WB = Willingness to buy; R² = Coefficient of determination; f² = effect size; RMSE = Root-mean-square error; MAE = Mean absolute error.

The current research stream on PLS-SEM advocates the importance of using the SRMR value for model fit (Hubona et al., 2021, Sarstedt et al., 2022), and the obtained value is 0.042 (below the threshold of 0.1 or more strict 0.08), representing a good model fit (Henseler et al., 2014). In addition, a CB-SEM approach, which strongly relies on the concept of model fit, was applied (see appendix-IV). Scholars argue that the fit indices should be taken together to have a better perspective (Dash & Paul, 2021). Based on overall fit indices obtained, PLS-SEM implies an acceptable fit but a poor fit is observed through CB-SEM indices.

The final part of the structural model presents the statistical significance and relevance of the path coefficient results (Hair et al., 2019). Table 9 and Fig. 3 indicate that the relationship of consumer animosity and willingness to buy Chinese products is negatively significant, thus supporting the hypothesis (CA → WB: β = -0.733). The result aligns with previous findings (Jakovljevic et al., 2020, Shoham and Gavish, 2016, Timber and Chiu, 2020). Furthermore, each dimension’s impact on the willingness to buy Chinese products was calculated through disjoint approach and reported. Pandemic animosity has the largest negative impact on willingness to buy, confirming that due to the ongoing Covid-19 crisis, peoples’ pandemic animosity feelings are peaking.

Table 9.

Model results along dimensional impacts.

Hypotheses		Path coefficients	T stats	P values	Result
Consumer animosity	→ WB	-0.733	59.761	0.000	Accepted
Individual dimensions’ impact on willingness to buy
Pandemic animosity	→ WB	-0.654	15.646	0.000	1st
Economic animosity	→ WB	-0.202	7.152	0.000	2nd
War animosity	→ WB	-0.076	3.335	0.001	3rd
Cultural animosity	→ WB	-0.071	1.990	0.047	4th
Political animosity	→ WB	-0.034	2.335	0.020	5th

Note. WB = Willingness to buy.

Fig. 3.

Model Paths.

6. Discussion

This study extends Klein et al.'s (1998) consumer animosity model by successfully adding the new dimension of pandemic animosity and developing its scale. All the studies conducted support the existence of pandemic animosity, and that the pandemic has divided the world in terms of opinions regarding Covid-19, its origin, and who to blame. The results show that pandemic animosity is distinct and measures a different sort of animosity than previous dimensions. The findings show that pandemic animosity negatively affects consumers’ willingness to buy Chinese products. This finding aligns with the impact of other consumer animosity dimensions towards willingness to buy. This finding is also supported by recent findings that indicate that pandemic animosity is driven not only by beliefs but also by emotions (Rojas-Méndez et al., 2022).

6.1. Theoretical implications

The study provides empirical support for Klein et al.'s (1998) theory of consumer animosity. It validates the extension of the animosity model by adding the pandemic animosity dimension and its clear conceptualization to better understand consumer behavior. Together with Rojas-Méndez et al.’s (2022), the pandemic animosity scale presented herein will help scholars measure the phenomenon comprehensively. The systematic mixed method approach used helped prove its existence and develop the measurement instrument, which is very helpful for understanding general animosity or pandemic animosity alone. The study further investigates the consumer animosity outcome variable and contributes to the academic literature on consumers’ willingness to buy. By doing so, it further proves the developed scale's generalizability, reliability, and validity. It also confirms the existing relationship, which previously lacked the pandemic animosity scale to investigate the relationship comprehensively and appropriately within today’s world context.

The pandemic has fundamentally changed the world, forcing society to rethink its healthcare, media, economic, and political systems; its social, cultural, and religious norms; and its scientific research, international relations, and relation with environment and nature (Jakovljevic et al., 2020; Panwar, 2020). It has also challenged several business ideologies: from globalization to nationalization (Cleveland & McCutcheon, 2022), from capitalism to neoliberalism (Anker, 2021), and from over-significance of work to understanding work-life balance (Amankwah-Amoah et al., 2021, Carnevale and Hatak, 2020). Scholars and policymakers must therefore rethink and reorganize social, legal, political, and economic systems to tackle such disasters in the future. Although the adverse effects of the pandemic are hopefully temporary, effective international cooperation to tackle such disasters in the future is imperative (Sharma et al., 2021). The pandemic animosity scale could be used to measure animosity and its effects on different behavioral outcomes within this rapidly evolving business environment.

This study proposes a tool to evaluate the extent and the impact of pandemic animosity on consumer and business decisions. Future research can use it to evaluate how contextual factors and pandemic-related narratives adopted by governments impact animosity among the population. This could be done through the prism of agenda-setting theory (Sheng & Lan, 2019). Indeed, official narratives divided the world into three groups (Dareini, 2020, Jakovljevic et al., 2020). The first—representing the U.S. narrative, including the U.K., India, Australia, some E.U. countries, South Korea, Japan, and others—blames China for the pandemic (e.g., Han & Marwecki, 2020). The second—representing the Chinese, Iranian, and Russian narratives, including some African and Asian countries—believes that Covid-19 was bioengineered by the U.S. to achieve certain economic objectives. The third, believes that Covid-19 is a natural occurrence and that no country should be blamed for it (Nawaz et al., 2020, Rojas-Méndez et al., 2022). Beyond the responsibility of authorities in developing international unrest, the role of media and social media should be equally accounted for (Barkemeyer et al., 2020). According to the conducted interviews, a large number of respondents based their beliefs on China’s role in the Covid-19 pandemic on the information from official media sources and discussions on social media.

6.2. Managerial implications

The research provides several managerial implications. First, the study highlights the critical issue of pandemic animosity in the context of buying foreign products. Although war, economic, political, cultural, and religious animosities have been shown to have an effect, pandemic animosity had not. Businesses around the globe have been victimized due to pandemic animosity, which required attention. Second, although managers were aware of pandemic animosity between the pandemic’s country of origin and a host nation, it was difficult to assess its extent and impact. The developed scale, which can easily be adjusted to various situations and countries, allows managers to assess the extent of pandemic animosity in a specific country or ethnic group. This understanding may aid businesses from a pandemic’s perceived country of origin in forecasting how customers are likely to respond to buying foreign products, allowing them to make informed business decisions about how to manage their existing businesses, launching new products, or establish new businesses.

Third, this construct can be used as country-specific (herein, China), thus marketers could assess the intensity of pandemic animosity towards any country. Marketers could therefore predict how a product and business of a specific origin might be perceived in a target market and devise a marketing plan developing tactics to overcome pandemic animosity and prevent pandemic consequences. Fourth, pandemic animosity between China and the U.S. has effects at a national and individual level. The recent incident in Atlanta in which Asian Americans were killed (Denselow, 2021) is an example of its effect at an individual level. Nevertheless, pandemic animosity is more likely to convert into economic animosity between the U.S. and China at a national level. During such confrontational situations, managers need to be very conscious and sensitive to avoid making business decisions that might increase this animosity. Marketers, for example, could refrain from promoting the country-of-origin identity and rather their brands as global—e.g., Toyota promotes itself as a global brand, especially in Asian countries (Han & Nam, 2019).

Fifth, practitioners in host nations could consider ‘owned-by’ or ‘made-in’ cues while creating marketing messages during pandemic environments. Likewise, international brands could use CSR strategies during a pandemic to help and support the affected communities, to show compassion, sympathy, and solidarity with victims of host nations. Sixth, the proposed model should generate interesting and novel results when applied to different regions and people that view the pandemic from different narratives. Managers must consider such variations in consumer behavior while making decisions. By applying the model and carefully keeping an eye on the current situation, practitioners should be able to more accurately select entry modes, define their global supply chains, design promotional campaigns, build their business identity, and leverage their country of origin.

Finally, the pandemic animosity scale could be used to carry out educative and informative campaigns against hate and racism (Crockett & Grier, 2021). Racism and stigma should be addressed to reduce hate and division, and instead replace them with constructive unity across the board to spread compassion and empathy to strengthen the collective efforts to defeat the pandemic. The pandemic does not discriminate gender, race, cultural or religious beliefs, and thus the battle against viruses should be fought collectively. As the Secretary-General of the U.N. stated: “now is the time for unity” (Ng, 2020). “Today, humanity is facing a severe crisis not only because of Covid-19 but also due to the absence of trust among humans” (Harari, 2020). Destitute global solidarity, trust, unity, empathy, and cooperation; humanity will remain the victim of such catastrophe or even worse (Jakovljevic et al., 2020). Furthermore, online social media platforms must take responsibility for controlling racism and hate and find effective solutions to stop such xenophobic and biased behaviors.

7. Limitations and future research directions

Notwithstanding its contributions, the study has the following limitations. First, the studies’ responses were obtained using stratified, snowball, and convenience sampling techniques. Future studies should adopt other data collecting techniques and explore other contexts and different or future pandemic narratives. Given that the pandemic is ongoing, its evolution (i.e., different variants and waves) will most likely present future research opportunities, such as an extension of this work to capture longitudinal data. Future research can also explore perceived attractiveness and intentions towards China and Chinese products post-pandemic in a business, investment, and tourism context. Given that pandemic animosity is situational and temporary, future research can explore its long-term social and economic impact. In addition, the survey data for study 3 points towards low external validity (i.e., CB-SEM model fit indices show a poor fit). Although several additional tests were used to prove the data’s validity, researchers should further investigate the phenomenon and model by considering longitudinal data at some later point of time, where fewer cases of Covid-19 remain, or where the disease is controlled, to prove the model fit indices accordingly. Despite its limitations, this paper contributes to the business, marketing, and consumer behavior literature by conceptualizing the pandemic animosity construct and developing and validating a scale to measure it and its effects.

CRediT authorship contribution statement

Muhammad Zahid Nawaz: Conceptualization, Formal analysis, Methodology, Software, Validation, Writing – original draft, Writing – review & editing. Shahid Nawaz: Conceptualization, Data curation, Methodology, Resources, Writing - original draft. Francisco Guzmán: Project administration, Writing - original draft, Writing - review & editing. Daria Plotkina: Conceptualization, Supervision, Visualization, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Biographies

Dr. Muhammad Zahid Nawaz is a postdoc research fellow at Dongbei University of Finance and Economics. He has completed his Ph.D. in Marketing from the same university. Previously worked in the corporate sector for more than six years in management positions of sales and marketing. His research work is published in Current Psychology, Public Integrity, Information Technology & Tourism, Sustainability and others. His main research interests include; consumer behavior, consumer brand engagement, consumer animosity, and sharing economy.

Dr. Shahid Nawaz is a serving Assistant Professor of Marketing at The Islamia University. He has more than ten years of teaching and research experience with publications in scholarly journals. His main research interests includes, consumer behavior, consumer brand engagement, consumer animosity, and brand management.

Dr. Francisco Guzmán is a Professor of Marketing. His research focuses on branding and social transformation, particularly on brand co-creation, brand equity, branding and sustainability, corporate social responsibility, public-private collaborations, and political branding. His research has been published in the Journal of International Marketing, Journal of Business Research, European Journal of Marketing, Industrial Marketing Management, Journal of Marketing Management, Journal of Brand Management, and Harvard Business Review América Latina, among others. He currently serves as Co-Editor in Chief of the Journal of Product & Brand Management and Associate Editor of the Journal of Business Research.

Dr. Daria Plotkina is an Associate Professor at EM Strasbourg Business School, University of Strasbourg, obtained her doctorate in Digital Marketing from EM Strasbourg Business School. Her research focuses on online information and interaction, persuasive communication, and financial self-efficacy. Her research was communicated on international conferences and is published in Journal of Business Research, Psychology & Marketing, Journal of Consumer Affairs, and Journal of Retailing and Consumer Services.

Appendix I: Dimensions of consumer animosity with definitions

Dimensions	Study direction	Definition	Authors
Consumer animosity	Scale development	The remnants of antipathy related to previous or ongoing military, political, animosity or economic events.	(Klein et al., 1998)
War	Scale development	The remnants of antipathy related to previous or ongoing military events.	(Klein et al., 1998)
Economic	Scale development	The remnants of antipathy related to previous or ongoing economic events.	(Klein et al., 1998)
Political	Scale development	The remnants of antipathy related to previous or ongoing political events.	(Klein et al., 1998)
Cultural	Scale development	The remnants of antipathy related to cultural differences.	(Kalliny et al., 2017)
Religious	Scale development	The remnants of antipathy related to religious differences.	(Kalliny et al., 2017)
Pandemic	Scale development	The remnants of antipathy related to a pandemic.	(Self-generated)

Appendix II: Literature review

Study	Method	Investigated parties	Dimensions of animosity	Related concepts	Key findings	Relevant research directions
Klein et al. (1998)	Quantitative survey	China against Japan	Sources - war - economic →general	- consumer ethnocentrism - country of origin - hostility	- conceptualization of animosity - significant impact on buying decisions above and beyond the effect of consumer ethnocentrism	- to continue research on animosity - to extent the model and account for other sources of animosity
Shimp et al. (2004)	Quantitative survey	South vs North of USA	Location regional animosity	- social identity - stereotyping	regional animosity correlated with preference and willingness to pay for products/services from one’s own region	to explore the effect of animosity on consumer behaviour at international level
Riefler and Diamantopoulos (2007)	Literature review and quantitative study	Austria	Sources - war - economic - political - personal	behavioural outcomes	- consumers have animosity to different countries for different reasons - not only war and economic tensions are important sources of animosity - animosity ≠ culture dissimilarity	- to consider different causes of animosity - to extend the measurement model - to investigate situational animosity - to focus on context specific animosity
Cai et al. (2012)	Quantitative survey	China against Japan	Type implicit animosity	consumer ethnocentrism	- implicit animosity ≠ consumer ethnocentrism; - significant correlation with war but not with economic animosity - negative impacts on purchase intention	- to investigate how different sources impact on purchase decisions - to pay attention to deep animosity that is not expressed overtly
Tian and Pasadeos (2012)	Quantitative survey	China against Japan	Sources - war - economic	- traditionalism - conservatism - patriotism - internationalism *As antecedents of animosity	- four antecedents of consumer animosity - war and economic animosity decrease purchase intentions - only war animosity reduces product quality judgments	- to investigate situational animosity - to study factors explaining animosity, such as the effect of social norms
Nes et al. (2012)	Exploratory and empirical	US, Norway	Sources - war - economic - political - people	- prejudice and discrimination - boycott	- extending animosity from 2 to 4 dimensions - animosity impacts buying behaviour through affect	- to study different countries - to extend to other contexts and impacts of animosity - to explore animosity in B2B context - to study symbolic and emotional effect of country-of-origin
Perviz et al. (2014)	Qualitative study	Slovenia	Sources - war - economic - political - personal experience - people (mentality, lifestyle, and language) - physical environment	behavioural consequences of animosity	- main sources of animosity: people, politics, and personal experience - consumer animosity influences purchase behaviour in selected product categories	- to carry out triangulating studies - to study effect on different product and services - to investigate the intensity of animosity in a particular moment/ period (situational vs stable)
Harmeling et al. (2015)	Quantitative survey	China vs Japan; Russia vs USA	Affective - agnostic (anger) - retreat (fear) [Source: war]	- attribution theory - national identity	- mediational role of emotions - agonistic emotions are related to NWOM and product avoidance - retreat emotions are related to product avoidance	- to study emotional response based on the source of animosity - to investigate effects of causality attributions on animosity and behaviour
Little and Singh (2015)	Scenario-based experiment	scenario-based animosity, countries not specified	Type - purposeful vs unguided actions - intended/ unintended consequences - event (causal) vs history	locus of control	- should be investigated both how animosity is developed and its consequences	to use less homogeneous sample than (US) students
Shoham et al. (2016)	Meta-analysis	–	Type - general - consumer	- consumer ethnocentrism - cognitive dissonance theory	- animosity ≠ ethnocentrism - negative relationship between consumer animosity and product quality - product quality judgments partially mediate the effect of consumer animosity/ ethnocentrism and willingness to buy	to enrich the research body on consumer animosity
Lee et al. (2017)	An experiment and two quantitative surveys	China against Japan	Type - semantic memory (one’s opinion) - episodic memory (one’s experience)	- consumer ethnocentrism - boycott - memory theory	- consumer ethnocentrism is driven by semantic memory - animosity is underpinned by episodic memory - consumer ethnocentrism is temporally more stable than animosity - animosity influences boycott behaviour during but not after the dispute - ethnocentrism influences boycott behaviour during and after the dispute - consumer ethnocentrism is an antecedent to consumer animosity	to compare countries with differing cultural profiles (individualistic vs collectivist)
Antonetti et al. (2019)	Quantitative survey	China against Japan	Affective - threat emotions (anger and fear) - extreme emotions (contempt and disgust)	social functionality	- extreme negative emotions are key to explain the effects of animosity on behaviour	to investigate additional emotions and other types of animosity
Leonidou et al. (2019)	Quantitative survey	Ukraine against Russia	Source - war - economic - political - social (people)	- realistic group conflict theory - social identity theory - self-categorization theory	- individuals’ personality (OCEAN) influence consumer animosity, except for agreeableness - consumer animosity influences product avoidance, moderated by cultural dimensions	to investigate other possible antecedents of consumer animosity and underlying values and ideologies
Zdravkovic et al. (2021)	Quantitative survey	Croatia vs Israeli-Palestine conflict	Affective anger vs hope	- power - universalism - openness to change	- power value predicts anger toward Palestine - universalism predicts anger toward Israel and hope - openness to change is only related to hope - anger/hope predict consumers’ buying intentions	- to explore other sources of animosity where people are not directly involved - to investigate the attribution of blame
Farah and Mehdi (2021)	Literature Review	–	Source - war - economic - political - people	- consumer ethnocentrism - cognitive/ affective theory	- few studies have investigated animosity - first gap: the inability to specify the trigger - second gap: lack of measurement models	- to investigate animosity - to consider more dimensions of animosity (at least 4), not less - to investigate specific triggers of animosity - to focus on recent events and not historical tensions - to develop more adaptable models to measure animosity - to investigate mechanical causes of animosity development and new causes of animosity
Campo and Alvarez (2021)	Quantitative survey	China against Japan and other countries	Source - historical/military - economic/ social -political/ ideological →general	- perceived risk - cognitive/ affective theory	- target country impact on how animosity is formed differently - the cause of animosity (traditional or other) influences on intention to visit and perceived risk of visiting the country - Tensions > crime and security > environment > government and ideology > history > living conditions > negative feelings > culture, people, and national traits > treatment of women and minorities > war and conflict	- to investigate animosity for every specific cause and target country - to focus on different root causes of animosity (e.g., terrorism, natural disasters) - to study the component of epidemic and especially COVID-19 on animosity
Hoang et al. (2022)	Quantitative survey	Vietnam against China and USA	Source - war - economic	- consumer ethnocentrism - cosmopolitanism - country-of-origin	- animosity findings for developed and developing countries also apply for emerging countries - the relationship between ethnocentrism and product judgement, and the relationship between cosmopolitanism and willingness to buy are not significant in a developing country	- to enrich quantitative methods with qualitative for a better understanding of animosity - to account for current events and influences - to consider the proximity to the target country

Appendix III: Items generated for pandemic animosity

Item code	Theme code	Pandemic animosity
PA1.	Fear of virus	‘Chinese’ people scares me due to the ‘Covid-19 pandemic’.
PA2.	Hate China	I do not like ‘China’ due to the ‘Covid-19 pandemic’.
PA3.	Lockdown situation	I will not forgive ‘China’ for my mental and physical health damages during the ‘Covid-19 pandemic’.
PA4.	Anger/frustration	I am very angry at ‘China’ due to the ‘Covid-19 pandemic’.
PA5.	Deaths/losses	I could never forgive ‘China’ for human and economic loss in my country during ‘Covid-19 pandemic’.
PA6.	Hate/Abuse	I hate people from ‘China’ due to the ‘Covid-19 pandemic’.
PA7.	Contagious virus	I would not like to meet people from ‘China’, which is the origin of the ‘Covid-19 pandemic’.
PA8.	Quarantine/social distance	I would never forgive ‘China’ because of the suffering during the ‘Covid-19 pandemic’.
PA9.	Bio-weapon/ bio-engineered	I believe ‘Covid-19 pandemic’ is a ‘Chinese’ virus.
PA10.	Responsibility	I believe ‘China’ is responsible for the ‘Covid-19 pandemic’ spread in the world.
PA11.	Hiding information	I think ‘China’ is responsible for hiding the information regarding the ‘Covid-19 pandemic’ outbreak.
PA12.	Chinese eat everything	The eating practices of the ‘Chinese’ are crazy because they eat wild animals (which is the possible source of coronavirus).

Note. ‘China’/ ‘Chinese’/ ‘Covid-19 pandemic’ in italics represents the context specification, which could be replaced with other countries for their context specification and/or other health contexts.

Appendix IV: CB-SEM model fit indices

Criterion	Fit Indices	-	Criterion	Fit Indices
CMIN	15.251	–	CFI	0.932
IFI	0.932	–	NFI	0.929
TLI	0.904	–	GFI	0.879
AGFI	0.781	–	PNFI	0.664
RMSEA	0.107	–	SRMR	0.040

Note. CMIN = chi square value (X²/DF); CFI = Comparative fit index; IFI = Incremental fit index; NFI = Normed fit index; TLI = Tucker Lewis index; GFI = Goodness of fit index; AGFI = Adjusted goodness-of-fit; PNFI = Parsimonious normed fit index; RMSEA = Root mean square error of approximation; SRMR = Standard root mean square residual.

Data availability

Data will be made available on request.