Analysis of Product-Country Image from Consumer’s Perspective: The Impact of Subjective Knowledge, Perceived Risk and Media Influence

Limited research has empirically explored the antecedents that explain product-country image in the context of international marketing. This topic is particularly important as consumers’ concerns about sustainability and the country of origin are relevant factors influencing purchase decisions relating to foreign products. This study addresses the gap by developing a causal relationship model to explain which are the main determinants of how consumers perceive product-country image on the basis of key aspects that define sustainability (environmental, social, quality/safety, and economic factors). This research is focused on the Spanish horticultural sector, Europe’s leading fresh produce supplier, which has historically experienced major crises that have affected its image abroad. The model is tested by using PLS-SEM on a sample of 303 consumers in Germany, one of the main destination markets of Spanish horticultural products. The results indicate that consumers’ subjective knowledge about image crises of the country and its products/services and consumers’ perceived risk of these products and information sources (organic and induced image) influence the formation of product-country image. The results have implications in the design of marketing campaigns and in the improvement of sustainable strategies for international companies.


Introduction
Sustainability is a complex aspect, which has become a relevant issue among countries, industries, and citizens. Thus, the problems derived from lack of sustainability have worsened in the field of business due to the phenomenon of globalization. Nowadays, many consumers are increasingly inclined to make informed purchasing decisions regarding goods and services that promote environmental, social, and economic sustainability [1]. For this reason, organizations should incorporate sustainability criteria in the design of their products and services in order to reduce or even eliminate the negative outcomes arising out of economic, social, and environmental issues [2]. This is associated with the concept of 'critical consumption', that is, the purchase or boycott of goods and services for political, ethical, or environmental reasons [3].
Despite the intentions of a sector and/or organization to develop its goods and services in a sustainable manner, certain incidents can have a detrimental environmental and social impact, and result in damage to future economic sustainability [4]. In the food industry, negative publicity incidents involving unsustainable agricultural practices have been predominant in the market. These incidents or events can result in crises of food origin and also related to environmental, social, and economic problems. Food crises, which are a main source of public concern, have led to the rapid presence of health alerts and image crises in the country of origin, that can result in the reluctance to buy specific foreign goods and services from suppliers that consumers consider to follow unsustainable foreign products and services, which is a relevant topic in the international marketing literature (e.g., [20]). Second, there is scant research that analyzes the relationship between subjective knowledge and perceived risk, with contradictory results (e.g., [21][22][23]). Third, empirical evidence of the influence of perceived risk on PCI research is scarce (e.g., [24,25]). Therefore, it is necessary to delve into the role of this variable in the formation of PCI. Fourth, the related research mostly looks at investigating consumers' perceptions and behaviors associated with tangible products, while research analyzing the country image effect [26] and PCI [27] in the services sector is scarce. In order to contribute to filling this gap, this study investigates consumers' perception of PCI when it comes from both a tangible product category and a service offering. Finally, media influence through organic and induced image has been widely analyzed in the context of tourist destinations (e.g., [28]), but as far as we are aware, no previous studies have analyzed this effect in the context of international marketing and the consumption of foreign products.
Against this background, and in an attempt to shed light on some of the main determinants of how consumers perceive product-country image on the basis of key factors that define sustainability, the purpose of the current study is two-fold: • First, to examine the role of consumers' subjective knowledge (about image crises that have affected a particular country and about its products and services) and perceived risk as factors affecting PCI. • Second, to analyze how two fundamental information sources (organic and induced image) moderate the relationship between perceived risk and PCI.
The remainder of the paper is structured as follows. First, we review the literature on the country-of-origin effect and the related concept of PCI. Second, we examine the factors that influence PCI, that is, subjective knowledge, perceived risk, and media influence, and we propose and justify the conceptual model and its hypotheses. Subsequently, we outline the research methodology and the empirical analysis used in the study to test the hypotheses. Next, the research findings are described and discussed, then the main conclusions, and the theoretical and managerial implications are examined. Finally, the paper recognizes the limitations of the study and identifies directions for further research.

Country-of-Origin: Products and Services
The pressure on organizations to implement more sustainable and responsible management practices by safeguarding against environmental, social, and economic issues does not only come from governments, since "consumers are also increasingly looking for products and services that reflect their own values and provide a 'feel good' emotion by indirect support of the environment and society" [29] (p. 111). This management approach is getting to be a differentiating aspect as well as a competitive advantage in the marketplace. Furthermore, the consumer's judgment on a product and/or service depends on its association with a particular country, that is, the effect of Country-of-Origin (COO) [30]. Thus, the origin of a product or service has a strong effect on the consumer's evaluation, and it is the reason that countries can mainly benefit from this when the category of product or service offering associated with a country has a positive and prominent meaning for consumers [31].
The country's image at macro and micro level and the related term of COO remains important areas of research in consumer behavior, international marketing, and international business (e.g., [32,33]). At the macro level, country-of-origin image (COI) has been defined as "the total of all descriptive, inferential and informational beliefs one has about a country" [34] (p. 193). At the micro level, the term PCI is used, and it has been conceptualized as "the overall perception consumers form of products from a particular country, based on their prior perceptions of the country's production and marketing strengths and weaknesses" ( [35], p. 480). Most origin-effect research has identified PCI as the source construct for origin effect [36]. Accordingly, PCI can be measured through the image of the goods and services that are considered to be typically of high quality, from a specific country that is known for exporting better goods and services than other countries and for its popular exports around the world [30]. In the literature on country image, there are two principle streams: (1) The general image of the country concerning its economic, technological, and political development (COI) [37], and (2) the image related to the product evaluation that can be attribute-specific (PCI) [38]. It is highlighted in PCI research that "although the size of the effect may vary across products, consumers and situations, the impact of PCI is real, pervasive and measurable" [39] (p. 1178). The concept is multidimensional in its nature, including products, services, brands, and countries, which may be involved in its production and marketing [40]. In addition, PCI literature has established that consumers differentiate across the image that they have of domestic and foreign products [41]. There is an image associated with the home country of consumers and another with foreign countries (see, for example, [42]). This research will focus on the second type of image-i.e., foreign PCI-as it is consistent with the main purpose of this research: To explore the extent to which PCI is influenced by some key antecedents in times of image crises. For this purpose, we focus on a particular product category (that is, fruits and vegetables) that has been severely damaged in recent years by the influence of the media (projecting a negative image in terms of environmental degradation, the excessive use of fertilizers and pesticides, and the bad social conditions for immigrant workers, among others). In general terms, it is expected that the subjective knowledge of consumers about those foreign products and the image crises suffered by the country of origin of them affect their perception of risk in the consumption of those products and their product-country image. Moreover, this study will analyze the role of organic and induce image in the generation of product-country image by consumers in the destination markets. Foreign PCI is shown in Figure 1 along with the other five components of the conceptual model of this research. and institutions are vital to counteract the negative effect of the media and provide a positive PCI [93]. Few researchers have evaluated the effects of media coverage on perceptions and attitudes related to PCI. With regard to attitudes towards a particular foreign country and its products/services, [25] have shown that when the product/service and the country aspect are associated with a risk, the influence of the information sources moderate the attitudes toward that PCI. In this sense, a better understanding of media framing and strategic communicators may help alter negative perception or cultivate favorable opinion toward the products made in a specific country through mass communication. Moreover, previous authors [94] suggest that it is important to measure a conflictive sector where information sources have more influence on the image of a specific place according to the consumers' perceived risk. Following this reasoning, in the context of image crises, it is expected that an organic image provided by fake news strengthens the relationship between perceived risk and PCI, whereas induced image provided by the sources of information from organizations and institutions will help to reduce this negative effect, undermining the relationship between perceived risk and PCI. In response to these arguments, the following hypothesis is formulated:

Hypothesis 5 (H5). The information sources formed by (a) organic image moderate (strengthen) the relationship between perceived risk and product-country image, and (b) induced image moderate (undermine) the relationship between perceived risk and product-country image.
The conceptual model of this research and the hypotheses formulated are summarized in the following Figure 1.

Participants and Procedure
We specifically chose the Spanish fruit and vegetable industry because Spain is prominent in the export of these types of food products and, historically, has suffered from several image crises in multiple markets (e.g., Germany, France, United Kingdom, etc.). We selected Germany as our empirical setting as it is the main destination market for Spanish horticultural products. In addition, the main image crises that have affected this sector in the last years have come from the German mass media, resulting in a devastating Generally, PCI literature models a representative product structure in terms of tangible goods or services related to the activities of a country [43]. Following the Cue Utilization Theory, consumers make an assessment about the product/service in terms of intrinsic and extrinsic cues. The former cues are recognized as part of the product and/or service itself (e.g., attributes or key components), which cannot be manipulated without altering the nature of the product or service. The latter cues are attributes associated with the product and service that are not inherent (i.e., packaging, COO, and prices). Previous research has demonstrated that extrinsic cues are usually more easily observable by the consumer than intrinsic cues [44]. In fact, COO is conceptualized as a relevant extrinsic and cognitive cue, influencing consumers' decision making and contributing to the formation of their preferences [45]. Despite the assumptions made by COO literature that consumers are affected by the origin of a product based on the type of product category they are dealing with [46], the literature on the effect of the COO associated with services offered is scarce and sometimes contradictory. Some academics argue that the effect of COO on services is similar to what happens when it comes to tangible products [47], whereas other researchers contemplate that the special characteristics possessed by services (i.e., inseparability, intangibility, heterogeneity, and perishability) improve the effect of COO on consumers [48]. In addition, the COO effect may vary between service categories (e.g., [49]). However, studies that analyze the way in which information about the COO of a service offering affects consumer purchase behavior are mainly limited to airlines, banking, insurance, and food and beverage services [27]. Thus, specific attention to both product and services might help to shed light on the topic.
With this background in mind, the literature on PCI has mainly focused on the influence variables such as consumer ethnocentrism, cosmopolitanism, materialism, nationalism, or consumer animosity (e.g., [15][16][17]). However, the effect of consumers' subjective knowledge, perceived risk, and information sources in the process of PCI formation for both foreign products and services has received surprisingly little scholarly attention. In an attempt to advance theoretical development in the area of PCI, this study sheds light on the role of relevant PCI influencing factors in a context of image crises associated with unsustainable practices in the COO.

Theoretical Model and Research Hypotheses
In the proposed model, several factors influence the image of a foreign productcountry: (i) Country-specific aspects (i.e., consumers' subjective knowledge about image crises in terms of environmental, social, quality/food safety and economic issues, and subjective knowledge about foreign products/services); (ii) consumer perception (i.e., perceived risk based on environmental, social, quality, and economic risks); and (iii) information sources (i.e., organic and induced image).

Subjective Knowledge and Perceived Risk
In the literature, consumer knowledge has been considered as an important variable that affects consumer behavior [50,51]. Consumers constantly gather and store information in the memory as knowledge, and they use this knowledge in the process of evaluation and choice of products/services [52]. Consumer knowledge is formed by three components: Subjective knowledge (i.e., what individuals think they know about the company, country, product, brand and/or service; also known as perceived knowledge), objective knowledge (i.e., what consumers actually know), and experience-based knowledge [53]. Divergence between subjective and objective knowledge has been observed, with subjective knowledge having more influence on actual behavior in the specific context of sustainable practices (e.g., [54]). In addition, subjective knowledge is not consistently correlated with objective knowledge [50].
For the purpose of this research, we will focus on consumers' subjective knowledge due to its special relevance in sustainable practices in consumption [54]. It can be defined as the consumer's perception of their level of knowledge regarding a foreign country or COO. In particular, this concept will be referred to: (1) Consumers' perceived knowledge about the different image crises associated with unsustainable practices that a foreign country has suffered in terms of production and consumption; and (2) their perceived knowledge about the products and services that come from that foreign country.
First, we will analyze from a consumption perspective the situation of a COO that has been the object of multiple negative news stories (real or fake news) circulated in other countries and markets, generating an image crisis for this country. An image crisis can be defined as a "specific and unexpected event (or series of events) that causes a high level of uncertainty and threatens or is perceived as a threat to the achievement of key organizational goals" [55] (p. 116). Since consumers often use information from the COO to evaluate products in terms of quality, risk, or value [56][57][58], image crises of a foreign country can influence consumer behavior [59]. In the context of the food industry, image crises have been associated with several food safety, environmental, and social issues, and economic viability problems [60]. This type of information has been disseminated through mass media and represents the fundamental pillars of sustainability. Previous works have argued that some measures to improve food safety result in lower economic efficiency and lower sustainability [61]. Consequently, consumers' anger or discomfort towards a particular country they perceive as hostile may affect their perception of the country image and the export performance of products associated with it [62].
The problems associated with sustainability most often apply to food production and consumption than other human activities, as the provision of food is in fact the human activity with the highest individual environmental impact [61]. In this sense, consumers may use the term "sustainability" as a type of shorthand for the "green and good" to describe production and consumption systems related to a broader range of attributes, such as community-based efforts to build healthy, just, and local food systems [63]. In the case of the Spanish horticultural sector, it has been the subject of several negative news stories generated mainly in other countries through the various information sources blaming quality/food safety (e.g., integrity of producers, suppliers, and sellers concerning the quality of products, fulfilment of requirements of its products, the quality of the products is inconsistent with the descriptions given on product packaging or ads, etc.; see for example, [64]); environmental issues (e.g., water consumption or the type of production in climate change; see for example, [65]); social risks (e.g., labor conditions and illegal immigration, among others; see for example, [66]); and economic risk (e.g., economic inequality; see for example, [66]). All these confusing and disturbing news stories have been the origin of several image crises for this sector [67], generating subjective knowledge in consumers from the main destination markets (i.e., Germany, United Kingdom, and France) about the country and its products and services.
Few studies have empirically demonstrated that consumers' subjective knowledge influences perceived risk, and some results have been contradictory or inconclusive. Subjective knowledge is closely related to confidence in decisions and it is a key factor in decision making. Therefore, it can reduce consumers' risk perception, which is the degree to which one person perceives the likelihood of experiencing an unfavorable outcome of a decision or an action [68]. Thus, [23] analyzed tourists' decision making and showed that subjective knowledge appears to have the strongest influence on tourist risk perceptions compared to objective knowledge and other types of prior knowledge. Previous study [21] found that consumers' assessment of the level of their knowledge influence their risk perception and actions. Moreover, [22] proved that the effect from the level of knowledge about genetically modified food is inversely proportional to perceived risk, and [69] only found empirical evidence for the influence of subjective knowledge on physical perceived risk also in this context.
According to this background, consumers' subjective knowledge about the image crises of a foreign country and its products is expected to influence consumers' perceived risk towards the products from that country. In particular, considering that one of the main reasons for the origin of the image crises suffered by countries is related to unsustainable practices, consumers' subjective knowledge can be measured in terms of environmental, social, quality/safety, and economic factors that affect their perception of sustainability in foreign countries. Moreover, perceived risk has been conceptualized as a multidimensional variable that includes components that are in line with the main dimensions of sustainability, that is, environmental, social, quality/safety, and economic risk [70]. On the other hand, following this line of reasoning, it may be assumed that consumers' subjective knowledge about products and services that come from a particular foreign country could influence their perception of the risks (i.e., environmental, social, quality/safety, and economic risks) that are derived from those products and services. Consequently, we propose the following hypotheses: Hypothesis 1 (H1). Subjective knowledge about image crises influences consumers' perceived risk.

Hypothesis 2 (H2).
Subjective knowledge about foreign products and services influences consumers' perceived risk.

Subjective Knowledge and Product-Country Image
According to the COO image cognitive-processing link, consumers with specific product knowledge, and consequently high skills to evaluate a particular product, tend to rate country-products either more or less positively than consumers with less knowledge [71]. However, consumers who have little knowledge about a product from a particular country will become more dependent on the stereotype that they hold about that country [51]. Therefore, research on COO effects should take into account factors affecting consumers' prior knowledge of imported goods and/or product experience. The relationship between consumers' knowledge and image has been specially studied in the tourism industry. Thus, some studies have concluded that the image of a place is significantly influenced by an individual's knowledge about the country or the destination [72,73]. Therefore, individuals will be able to build a mental representation of the place based on their knowledge, which in turn produces their own perceived personal images [74]. However, in the food industry, consumers have limited information about the conditions in which foods from other countries are made [51]. It is the reason that most consumers make inferences based on the perception of general conditions in the specific country [75], forming, for example, their perceived PCI from the subjective knowledge that they have about foreign products and services. In fact, previous research has addressed the importance of inferences in the formation of a company's overall image [76]. For example, [77] and [78] argued that consumers' subjective knowledge about organic foods greatly affected perceptions, attitudes, and purchase intentions. A previous study [79] found that subjective knowledge about organic vegetables has a positive effect on perception and consumption behavior for Belgian consumers. Thus, since consumers use their subjective knowledge about products and services to determine their perceived image, information about sustainable practices in the food industry can positively contribute to consumers' perceived image about foreign products and their COO. Along these lines, the National Research Council highlighted the benefits of sustainability for the food industry. In its 2010 report, "Toward Sustainable Agricultural Systems in the 21st Century", it states that sustainability meets four goals [80]: (i) Satisfy human food, feed, and fiber needs, and contribute to biofuel needs; (ii) improve environmental quality and the resource base; (iii) sustain the economic viability of agriculture; and (iv) enhance the quality of life for farmers, farm workers, and society as a whole.
Based on these arguments, it is reasonable to think that consumers' knowledge about the Spanish horticultural products and services influence their PCI. Hence, the following hypothesis is formulated:

Hypothesis 3 (H3).
Consumers' subjective knowledge about foreign products and services influences their product-country image.

Perceived Risk and Product-Country Image
For more than half a century, the concept of perceived risk has been extensively studied as a fundamental facet of consumer behavior. Previous research suggests that perceived risk represents the subjective "evaluations of the probability as well as consequences of a (negative) outcome" [81] (p. 8). According to [25], news framed with risks elicits negative perceptions of the PCI in terms of both the product/service aspect and the country aspect, understanding image as "the expression of all knowledge, impressions, prejudices, and emotional thoughts a person or group of persons has/have of a particular object or place" [28] (p. 4). Thus, when negative outcomes are likely or when uncertainty is high, perceptions of risk increase [25]. More specifically, negative events and/or incidents associated with social, economic, political, and food security and safety scandals, among others, that increase the perception of negative consequences have a damaging effect on the country' image [28]. For example, Thailand has been subject to political and security risks (stemming from social scandals), which have consequently resulted in external negative images of this country. In the particular case of the Spanish horticultural sector, it has been subject to incidents about bad social conditions for immigrant workers, environmental degradation, and the excessive use of fertilizers and pesticides, most of them resulting in an increase of perceived risk by consumers and, consequently, damaging its projected image.
Consumers form assessments about PCI based on risk associated with foreign products, which are particularly salient when product categories are prone to crises or risk [24]. Because of the multiple crises that have occurred in the food industry, consumers are becoming increasingly skeptical about the quality of foreign-made products/services from the country in question. In addition, there is evidence that perceived risk influences consumers' decision-making process (regardless of whether that risk is real or perceived), and negatively affect consumers' image of products and services (e.g., [82]). Similarly, [83] argued that a favorable image could be created by reducing the perceived risk associated with particular risk factors (such as environmental, social, quality, and economic risk associated with unsustainable practices). Since the majority of studies have analyzed risk perception in the context of destination attributes [25,28], little is known regarding the influence of perceived risk on PCI in the field of image crises. Therefore, it is reasonable to expect that, when consumers' perceived risk of foreign product and service increases after image crises, the PCI decreases. In response to these arguments, the following hypothesis is formulated: Hypothesis 4 (H4). Consumers' perceived risk influences their product-country image.

Media Influence
According to the Theory of Image formulated by [84], the world is a psychological and distorted representation of objective reality residing and existing in the mind of individuals. Based on this theory, the mental construction of the image of a product or service is made up of the organic image and the induced image. Following [85], organic image is formed though passively acquired information from an individual's interaction with non-commercial information sources (e.g., blogs, mass media, information received from friends and relatives, print media, and word-of-mouth), whereas induced image is formed through the active search for information (i.e., promotional and non-promotional sources, for example, by organizations, suppliers, cooperatives, public institutions and/or governments). Therefore, consumer perceptions derived from secondary information sources have a direct influence on the perceived image formation process [73].
Media influence (i.e., by general information sources and/or commercial information sources provided by organizations and institutions) is the source of a wide variety of information (real and/or fake), providing knowledge and creating consumers' perceived image, which consequently form the basis of consumer choices [86]. When sudden news (e.g., generates an image crisis) breaks, the information sources have an impact on the perceived risk [87], demonstrating an unfavorable effect on consumers' decisions [88,89]. Prior works found that the effect of organic image is greater in relation to the induced image [12]. Consumers that do not have direct experience with a particular foreign country and its products/services rely more on the mass media sources to get to know that country and its products/services [90]. In this sense, researchers suggest that mass media play an important role in cultivating people's perceived image of foreign countries and that people's evaluations of other countries are largely based on media coverage [75]. Furthermore, a PCI created by the media can be improved or impaired depending on whether positive outweighs negative media coverage, or vice versa. The influence of media coverage on the formation of an overall national image is pertinent to the question of how media shape a PCI. Media coverage on foreign products may likewise influence people's perceptions of and attitudes toward a given PCI. The image of a particular country that is considered hostile might not be easily changed [62], as the pre-existing stereotyped images of the country are likely to have unfavorable effects on the organic stage of the images. Despite this, companies and associations can induce a more positive image by investing more in promotion. Moreover, on many occasions, credibility of information sources is a key issue in 'risk communication' that can be defined as "any purposeful exchange of information about health or environmental risks between interested parties" [91] (p. 172). Particularly, risk communication is the act of conveying or transmitting information between companies and associations (and other interested parties, e.g., media, government, etc.) in regard to: (a) Health or environmental risk levels; (b) the significance or meaning of health or environmental risks; or (c) decisions, actions, or policies aimed at managing or controlling health or environmental risks [92]. It is the reason that the efforts by organizations and institutions are vital to counteract the negative effect of the media and provide a positive PCI [93]. Few researchers have evaluated the effects of media coverage on perceptions and attitudes related to PCI.
With regard to attitudes towards a particular foreign country and its products/services, [25] have shown that when the product/service and the country aspect are associated with a risk, the influence of the information sources moderate the attitudes toward that PCI. In this sense, a better understanding of media framing and strategic communicators may help alter negative perception or cultivate favorable opinion toward the products made in a specific country through mass communication. Moreover, previous authors [94] suggest that it is important to measure a conflictive sector where information sources have more influence on the image of a specific place according to the consumers' perceived risk. Following this reasoning, in the context of image crises, it is expected that an organic image provided by fake news strengthens the relationship between perceived risk and PCI, whereas induced image provided by the sources of information from organizations and institutions will help to reduce this negative effect, undermining the relationship between perceived risk and PCI. In response to these arguments, the following hypothesis is formulated: Hypothesis 5 (H5). The information sources formed by (a) organic image moderate (strengthen) the relationship between perceived risk and product-country image, and (b) induced image moderate (undermine) the relationship between perceived risk and product-country image.
The conceptual model of this research and the hypotheses formulated are summarized in the following Figure 1.

Participants and Procedure
We specifically chose the Spanish fruit and vegetable industry because Spain is prominent in the export of these types of food products and, historically, has suffered from several image crises in multiple markets (e.g., Germany, France, United Kingdom, etc.). We selected Germany as our empirical setting as it is the main destination market for Spanish horticultural products. In addition, the main image crises that have affected this sector in the last years have come from the German mass media, resulting in a devastating organic image (The most significant examples are the discovery in 2006 of residues of the insecticide isofenphos-methyl in sweet peppers that come from Spain, and the outbreak of the lethal bacterium E. coli in 2011 that was alleged to come from Spanish cucumbers.). Therefore, it is reasonable to suggest that image crises may play an important role in German consumers' subjective knowledge, perceived risk, and PCI.
In a first stage, we used 7 complementary pre-tests with academics and 30 pre-tests with consumers for item screening purposes. Subsequently, we selected many judges with mixed demographic characteristics (i.e., gender, age, and region) and individuals who consume fruit and vegetables to ensure the "relevance and representativeness of measurement items" [95] (p. 75) and that all items were comprehensible to consumers. Next, we drew the sample using an online survey software (LimeSurvey) that allowed customization according to participant responses, and a professional market research agency (panel) specializing in online research executed the survey. The sample procedure was stratified random sampling. Before taking the survey, a pre-study was conducted (n. 50) on German consumers. Appendix A shows the set hard limits used (i.e., a maximum number of respondents for the age, gender, and region). Thus, the sample was demographically and geographically representative of the German territory, with a balanced representation of consumers from the different regions of that country (see Appendix A). Participants were awarded points for contributing to our survey, which is a common practice [33,96]. Data were collected in January 2020. Of 490 invited respondents, 303 answered the questionnaire over a three-week survey period, yielding a highly satisfactory response rate of 61.83%; an adequate sample size in comparison to those employed in similar studies (e.g., [33,97]).
To analyze the proposed relationships of the conceptual model, we selected the Partial Least Squares (PLS) technique, which is widely acknowledged in the international business area [98]. Following the guidelines established by reference [99], the measurement model and the structural model were evaluated. For this purpose, the software SmartPLS 3.3.2. was utilized. Additionally, a two-step approach proposed by [100] was then adopted for the analysis. Thus, firstly, the measurement model was evaluated by analyzing the reliability and validity of each of the variables included in the proposed model. This is then followed by the structural model's evaluation, which involves estimation of the path coefficients between the constructs. A hierarchical component model (HCM) or higherorder model encompassing second-order constructs was selected for this study. To assess the model, we used a mixture of the repeated indicator approach and the use of latent variable scores in a two-stage approach in PLS-SEM. The two-stage approach can address the issue that any additional latent variable as a predecessor is always approximately zero and non-significant because almost all of the high-order construct variance is explained by its first-order constructs [101].

Constructs Measurement
The questionnaire design was based on scales developed in previous studies. To ensure content validity, most measurement items were adapted from validated scales (see Appendix B for items and sources). We asked respondents to answer questions that captured consumers' subjective knowledge about image crises and about foreign products/services, perceived risk, media influence and PCI.
Consumers' subjective knowledge about image crises was measured using a four-item scale adapted from the studies by [102] and [103], indicating the level of understanding and the ease of obtaining information about image crises in relation to the Spanish horticultural sector. Consumers' subjective knowledge about foreign products and services was measured using the scales by [102,104,105]. The questionnaire included 10 items used to operationalize four perceived risk dimensions (i.e., environmental, social, quality, and economic risk). Environmental risk was measured by three items, social risk was measured by four items, economic risk was measured by one item, and quality was measured by three items [106]. We adopted from [35] a four-item scale to operationalize PCI assessments (of Spanish horticultural products/services). We measured media influence using [107] five-item organic image scale and four-item induced image scale. Thus, respondents were asked about their level of agreement or disagreement with the influence exerted by general and noncommercial information sources (i.e., blogs and social media, internet, electronic media, print media, and information from relatives and friends) and commercial information sources provided by organizations and institutions from the Spanish horticultural sector (advertising, sales promotion, sponsorships, fairs and exhibitions, etc.) on the image that they have in regard to Spanish horticultural products and services.
All the items were measured using a seven-point Likert scale from 1 (strongly disagree) to 7 (strongly agree), indicative of the respondent's level of agreement or disagreement with the different statements included in the questionnaire. Since the survey was conducted in Germany, translation and back-translation were performed to ensure cross-language equivalence between the English and original German version of all questions. The items and associated code of the multi-item scales used in the analysis are shown in Appendix B.

Measurement Model
First, a confirmatory factorial analysis (CFA) using the statistical software SmartPLS v.3.3.2 was conducted to evaluate the psychometric properties of the operationalization of our focal constructs (i.e., scales and subscales). As expected, CFA provided evidence for the multi-dimensional (second-order) structure of a variable in the research model (i.e., perceived risk). Perceived risk results comprise four separate risk subscales (i.e., environmental, social, quality/safety, and economic risk). The four separate subscales made significant and strong contributions to overall perceived risk (βs > 0.70; ps < 0.01), with social risk showing the strongest effect, followed by the environmental and the quality and economic subscales. Consistent with the original work by [108] and the subsequent study by [107], media influence (i.e., information sources) was separated into two common categories: Organic image, comprising the sources of general information and, therefore, noncommercial information sources (i.e., does not come from organizations affiliated with the Spanish horticultural sector), and induced image, encompassing commercial information sources such as advertising or marketing-controlled information, that is, information sources from organizations and institutions related to the Spanish horticultural sector (e.g., suppliers, cooperatives, public institutions, and governments, etc.). According to prior research [35], PCI comprises innovativeness, design, prestige, and workmanship. Consistent with the study of [104], consumers' subjective knowledge about image crises was based on negative news stories about environmental, social, quality/food safety, and economic crises. Finally, evidence strongly indicated the one-dimensionality of the consumers' subjective knowledge about Spanish horticultural products and services.
CFA was used for purification of first-order and second-order scales, with the objective of eliminating items/dimensions that did not load meaningfully on the underlying constructs. As a result, one indicator (item ORG-3) was dropped from the media influence (organic image) first-order scale due to insufficient loading. In addition, three indicators (items SK-2, SK-4, and SK-5) were dropped from the concept of subjective knowledge about foreign products/services due to insufficient loading. After CFA purification, all factor loadings were significant at p < 0.01; standardized loadings fell within the ranges 0.82-0.95 and 0.86-0.95, respectively, for first-and second-order constructs. The internal consistency reliability, and the convergent and discriminant validity of the first-and second-order constructs were assessed following the procedure suggested by [109] and [110]. Internal consistency reliability was checked by computing Cronbach's Alpha and Raykov's Rho coefficients (see Table 1). The results were very satisfactory on both first-and second-order levels, with alpha and rho coefficients ranging from 0.82 to 0.95. Composite reliability (CR) and average variance extracted (AVE) were computed to further assess convergent validity [109,110]. On the first-and second-order level, composite reliability values were above the suggested cut-off point of 0.70 (see Table 1). As shown in Table 1, average variance extracted (AVE) ranged between 0.63 and 0.88 and was always higher than the relevant squared inter-construct correlations thus demonstrating discriminant validity [110]. The statistics brief in Tables 1 and 2 offers supportive evidence of reliability and validity of the focal constructs.  As reported in Table 3, we assessed discriminant validity by using the criterion that the AVE of each latent variable exceeded its shared variance (squared correlation) with other constructs [110]. As shown in Table 4, this condition was satisfied for all variables. In addition, [111] recommended examining the Heterotrait-Monotrait ratio of correlations (HTMT) to assess discriminant validity. This latest approach reveals the estimation of the true correlation between the two constructs. Thus, 0.90 is the threshold value suggested for HTMT [111]. Any values that are higher than 0.90 indicate a lack of discriminant validity. Moreover, the confidence interval's HTMT value should not be close to 1. Table 4 reveals that the HTMT criterion has been established for our PLS model. To sum up, the measurement model demonstrates adequate internal consistency, convergent validity, and discriminant validity.

Structural Model
Once the measurement model has been verified, the structural model is contrasted. The Standardized Root Mean Square Residual (SRMR) is an index of the average of standardized residuals between the observed and the hypothesized covariance matrices [112]. The SRMR is a measure of estimated model fit. When SRMR = <0.80, then the study model has a good fit [113], with a lower SRMR being a better fit. As Table 5 shows, the model's SRMR was 0.057, which demonstrated that this study model had a good fit, whereas the Chi-Square was equal to 326.507 and NFI equal to 0.837 was also measured. The model was able to explain as much as 63.5% of the variance in the main outcome construct of interest (i.e., PCI).  Figure 2). Consistent with expectations, consumers' subjective knowledge about image crises was also found to be a significant determinant of perceived risk. This means that consumers' subjective knowledge about image crises positively influence perceived risk (β = 0.631, t = 12.417, ρ < 0.01), which supports H1. Therefore, an increase in consumers' subjective knowledge about image crises decreases the favorable perception of Spanish horticultural products and services in terms of risk. However, the findings do not support the hypothesized (H2) direct effect of consumers' subjective knowledge on perceived risk. Therefore, H2 was rejected. Consumers' subjective knowledge about Spanish horticultural products and services positively affects PCI (β = 0.300, t = 4.56, ρ < 0.01), providing support for H3. Findings indicate that perceived risk has the stronger effect on PCI (β = −0.355, t = 5.14, ρ < 0.01), supporting H4.
The coefficient of determination (R 2 ) value of 0.39 indicated that consumers' subjective knowledge about image crises is responsible for explaining 39.8% of perceived risk of the Spanish horticultural products and services. This shows that German consumers do not feel much affection as they considered Spanish horticultural products as risky because of high environmental, social, quality, and economic risk. not feel much affection as they considered Spanish horticultural products as risky because of high environmental, social, quality, and economic risk. Figure 2. Structural model. Notes: All coefficients are standardized. The dotted path represents a relationship that is not the focus of this study but was studied before analyzing the effect of moderators. ** < 0.001

Moderating Effects
Following the recommendation of [101] to test the hypothetical moderating effects (see H5), we used the two-step technique, that is: Step 1, evaluate the direct effects and, step 2, add the interaction effect. Latent variable interaction terms were computed with the residual centering approach [114]. Firstly, product terms were created by multiplying each indicator of the moderating variable (i.e., organic and induced image) by each indicator of the relevant predictor variable (i.e., perceived risk and PCI). The resulting product terms were then regressed on their respective first-order components and the obtained residuals were finally used as indicators of (orthogonalized) latent interaction terms in the structural model.
We estimated three structural models with latent variable interaction terms using the two-step approach outlined above: M1 focuses only on the moderating effect of organic image (model fit, step 1: χ 2 = 238.50, SRMR = 0.064; step 2: χ 2 = 237.73, SRMR = 0.063); M2 analyzes the corresponding moderating effect of induced image (model fit, step 1: χ 2 = 236.08, SRMR = 0.061; step 2: χ 2 = 233.77, SRMR = 0.060); and M3 examines the simultaneous effect of organic and induced image (Figure 3) on the relationship between perceived risk and PCI (model fit, step 1: χ 2 = 260.16, SRMR = 0.065; step 2: χ 2 = 257.84, SRMR = 0.064). By considering only the organic image as a conditioning influence (M1), results show that organic image has a significant moderating effect on the relationship between perceived risk and PCI (β = 0.131, p < 0.05); this is in line with H5a. When considering only induced image as a conditioning influence (M2), the interaction effect is significant, indicating that induced image favors the relationship between perceived risk and PCI (in line with the H5b). Notably, there was evidence of differential effects of each media influence factor (i.e., organic and induced image). In this sense, for high levels of organic image ( = Figure 2. Structural model. Notes: All coefficients are standardized. The dotted path represents a relationship that is not the focus of this study but was studied before analyzing the effect of moderators. ** ρ < 0.001.

Moderating Effects
Following the recommendation of [101] to test the hypothetical moderating effects (see H5), we used the two-step technique, that is: Step 1, evaluate the direct effects and, step 2, add the interaction effect. Latent variable interaction terms were computed with the residual centering approach [114]. Firstly, product terms were created by multiplying each indicator of the moderating variable (i.e., organic and induced image) by each indicator of the relevant predictor variable (i.e., perceived risk and PCI). The resulting product terms were then regressed on their respective first-order components and the obtained residuals were finally used as indicators of (orthogonalized) latent interaction terms in the structural model.
We estimated three structural models with latent variable interaction terms using the two-step approach outlined above: M1 focuses only on the moderating effect of organic image (model fit, step 1: χ 2 = 238.50, SRMR = 0.064; step 2: χ 2 = 237.73, SRMR = 0.063); M2 analyzes the corresponding moderating effect of induced image (model fit, step 1: χ 2 = 236.08, SRMR = 0.061; step 2: χ 2 = 233.77, SRMR = 0.060); and M3 examines the simultaneous effect of organic and induced image (Figure 3) on the relationship between perceived risk and PCI (model fit, step 1: χ 2 = 260.16, SRMR = 0.065; step 2: χ 2 = 257.84, SRMR = 0.064). By considering only the organic image as a conditioning influence (M1), results show that organic image has a significant moderating effect on the relationship between perceived risk and PCI (β = 0.131, p < 0.05); this is in line with H5a. When considering only induced image as a conditioning influence (M2), the interaction effect is significant, indicating that induced image favors the relationship between perceived risk and PCI (in line with the H5b). Notably, there was evidence of differential effects of each media influence factor (i.e., organic and induced image). In this sense, for high levels of organic image (β = −0.144, t = 2.48, ρ < 0.01) the relationship between perceived risk and PCI increases, supporting H5a. Furthermore, for high levels of induced image (β = 0.280, t = 2.91, ρ < 0.01), the influence of risk on PCI remains stable, which supports H5b. These results provide evidence that the impact of perceived risk on PCI may be mitigated when the consumer is influenced by sources of information from organizations and institutions related to the Spanish horticultural sector (e.g., suppliers, cooperatives, public institutions, and governments, etc.), which is called an induced image. However, the relationship between perceived risk and PCI may be increased when the consumer is influenced by the source of information that comes from various external media (e.g., internet, information received from relatives and friends, blogs, and social media, etc.). , the influence of risk on PCI remains stable, which supports H5b. These results provide evidence that the impact of perceived risk on PCI may be mitigated when the consumer is influenced by sources of information from organizations and institutions related to the Spanish horticultural sector (e.g., suppliers, cooperatives, public institutions, and governments, etc.), which is called an induced image. However, the relationship between perceived risk and PCI may be increased when the consumer is influenced by the source of information that comes from various external media (e.g., internet, information received from relatives and friends, blogs, and social media, etc.).

Theoretical Implications
According to our findings, this paper makes the following research contributions. Overall, this study is one of the first attempts that provides empirical evidence on the role that some key variables in consumer behavior (i.e., subjective knowledge, perceived risk, and media influence) play in explaining consumers' perception of the product-country image. The study describes underlying mechanisms that explain the effect of image crises experienced by a particular foreign country on potential consumers of its products and services. First, this research demonstrates that consumers' subjective knowledge about image crises suffered by a foreign country (defined as reflecting the main sustainability dimensions of environmental, social, quality/safety, and economic factors) contributes to determining the risk perceived by consumers related to the products and services that come from that country. This risk is associated with environmental, social, quality, and economic factors, all of them key elements in the definition of sustainability. Therefore, these results reflect that consumers' subjective knowledge about image crises plays an important role in risk perceptions of products and services from a particular foreign country. In line with [69], our findings revealed that consumers' knowledge about foreign products and services is not a significant variable that explains perceived risk. This confirms the special relevance that negative publicity and even fake news has on consumers'

Theoretical Implications
According to our findings, this paper makes the following research contributions. Overall, this study is one of the first attempts that provides empirical evidence on the role that some key variables in consumer behavior (i.e., subjective knowledge, perceived risk, and media influence) play in explaining consumers' perception of the product-country image. The study describes underlying mechanisms that explain the effect of image crises experienced by a particular foreign country on potential consumers of its products and services. First, this research demonstrates that consumers' subjective knowledge about image crises suffered by a foreign country (defined as reflecting the main sustainability dimensions of environmental, social, quality/safety, and economic factors) contributes to determining the risk perceived by consumers related to the products and services that come from that country. This risk is associated with environmental, social, quality, and economic factors, all of them key elements in the definition of sustainability. Therefore, these results reflect that consumers' subjective knowledge about image crises plays an important role in risk perceptions of products and services from a particular foreign country. In line with [69], our findings revealed that consumers' knowledge about foreign products and services is not a significant variable that explains perceived risk. This confirms the special relevance that negative publicity and even fake news has on consumers' perception of risk in purchase decision making of foreign products. Despite this, the results show that consumers' subjective knowledge about the products and services has a positive and significant influence on PCI. Moreover, the perceived risk stands out as the factor that most influences the consumers' image of a product from a certain foreign countryi.e., in this case, the Spanish horticultural products and services-since they may associate this product-country with social, environmental, food safety/quality, and economic risks. This finding is consistent with previous research that reinforces the relationship between perceived risk and PCI in a conflict zone e.g., [28]. The analyses also validated the complex nature of the concept of perceived risk as proposed in the work of [70], encompassing the dimensions of environmental, social, quality/safety, and economic risk. Thus, high levels of perceived risk negatively affect the formation of PCI.
Second, our findings show that both organic and induced image condition PCI in different ways. Induced image (i.e., generated by organizations and institutions) has a negative moderating effect, since the link between perceived risk towards PCI shows a weaker relationship as a result of the moderating effect of a high induced image. However, organic image (i.e., created through electronic media and information received from relatives and friends) exerts an opposing moderating effect, strengthening the link between perceived risk and PCI. Thus, organic and induced image as information sources have countervailing influences on the effect of perceived risk on PCI. These results reflect the importance of the information sources from organizations and institutions (i.e., suppliers, cooperatives, public institutions, and governments, etc.), since they help to decrease the effect of consumers' perceived risk on PCI in a context of image crises. Furthermore, the importance of the institutions to offer information regarding the practices and measures adopted by the sector is notable, as well as a quick and efficient management against the fake news generated by organic information sources.

Implications for Policy and Practice
At a managerial level, this study shows the importance of image crises in the formation of consumers' perceived risk in terms of sustainability factors such as environmental, social, quality/safety, and economic elements, and, ultimately, in the definition of PCI. Therefore, organizations and institutions of a country should gather information about the level of consumers' subjective knowledge regarding those crises and try to reverse the distorted image that consumers have about that country and its products and services. In the particular case of the Spanish horticultural sector, this country has historically suffered multiple image crises (see, for example, [60,67], mainly resulting from internal problems related to sustainability (i.e., inadequate treatment of immigrant workers, environmental degradation, or the overuse of pesticides and chemical fertilizers, among others) and external fake news (i.e., the dissemination of unverified information in the mass media). Consequently, it is particularly important for this sector to know how to manage and improve its external image in terms of sustainability in the destination markets.
The results of this research show that Spanish horticultural companies and institutions could successfully re-establish the image of their productive and commercial sustainable activities through variables such as consumers' knowledge and perceived risk. In this sense, our findings demonstrate that it is important to reduce the information asymmetry in the destination markets, since consumers must have informed knowledge about the origin and consequences of the image crises in Spain. Therefore, it is central objective to reduce the incorrect and inadequate consumers' knowledge about the origin of the past crises, providing truthful and verifiable information in order to reduce the consumers' perceived risk regarding unsustainable practices. Moreover, the empirical evidence provided by this research supports the investment of the Spanish institutions and managers in marketing strategies to reduce consumers' perceived risk in the destination markets with the ultimate objective of improving PCI.
In addition, managers should be aware about the relevance of the information sources in the consumers' perception of product-country image. In this sense, the development of suitable communication strategies can be an important tool for Spanish horticultural organizations and institutions (i.e., induced image) to reduce the negative effect of risk perception on the perceived image, mainly in an image crisis context. Therefore, these organizations should invest in fast-acting promotional campaigns, trying to control the organic image and mitigating its effect through different media such as, for example, social media. Furthermore, when an image crisis occurs, policy makers should develop marketing strategies to counteract the negative effect generated by organic image (e.g., mass media, TV, and social media) that can influence consumer behavior. In fact, the Spanish Ministry of Agriculture, Fisheries, and Food has designed and made available to the sector a set of tools to promote the integral sustainability of the food industry, with the aim of improving the competitiveness of the industries that constitute the Spanish horticultural sector. Additionally, the information campaigns in the destination markets should accentuate the measures that have been applied in the sector to enhance sustainability in order to reduce the asymmetry of information in retailers and consumers.

Limitations and Further Research
There are several limitations to this study requiring further examination and additional research. First, the study was based on respondents from Germany, since it is the main destination market for Spanish horticultural products, and it is the country where the majority of negative and/or fake news about this sector has originated. Thus, a replication of this study in other destination markets (e.g., United Kingdom and France) may be used to compare the results and to show if there are country-specific differences. Furthermore, potential differences should not only be explored between countries, but also between general product-country or different product categories. In particular, future research should analyze the proposed conceptual model considering different types of products (fruits and vegetables) in order to determine the existence of potential differences in terms of perceptions and consumer behavior considering the product category. Second, the number of works that analyze PCI and its antecedents is relatively scarce. In this sense, it would be interesting to analyze the effect on PCI of variables such as warmth and competence [115], consumer animosity [116], risk communication [117], and place branding [118], as well as their impact on the anti-consumption phenomenon [119]. Finally, regarding the moderating effect, only one variable was considered: The influence of information sources through organic and induced image, representing only two parts of the general picture. Therefore, future research should include other moderating effects such as the usage of COO information [120] and consumer localism [Lehner et al., 2016]. In addition, future research should analyze different media and explore the effect of each of them (e.g., social media, mass media, TV, etc.), to provide useful information to organizations and institutions as to which source of information is the most influential. Moreover, further studies should analyze the moderating effect of the message and contents provided by different information sources on the relationship between perceived risk and PCI.
Author Contributions: All authors have equally contributed to this paper (conceptualization, methodology, formal analysis, and writing). All authors have read and agreed to the published version of the manuscript.   Note: Items were measured using a seven-point Likert scale from 1 (strongly disagree) to 7 (strongly agree).