The Impact of Source Credibility and Risk Attitude on Individuals’ Risk Perception

The Impact of Source Credibility and Risk Attitude on Individuals’ Risk Perception toward GM Foods: Comparing Young Millennials in the U.S. and China

Abstract

This research investigates the effects of source credibility and risk attitude on young millennials’ risk and benefit perceptions and purchase intentions toward GM foods. Results from two samples (young millennials in the U.S. and China) confirmed individuals’ risk attitude significantly influences their purchase intentions toward GM foods. Results also revealed a significant interaction effect of source credibility and risk attitude on risk perception of GM foods among Chinese respondents. Practical and research implications are discussed.

Introduction

Food choices can be personal or social-influence driven. Considering that food choices have important implications for human wellbeing and health, the public’s perceptions of food benefits and risks can lead to simple behavioral changes for broad food categories (Phillips & Hallman, 2013). Recently, the rapid development of genetic engineering technology has made genetically modified foods (GM foods) a prominent public concern. The scientific community and the food industry endorse genetic engineering technology because of its ability to solve the problems of food shortage and production (Clarke, 1997; Shapiro, 1999), improve the nutrient content of foods (Burke, 1997) and help in removing known food-borne allergens (Jones, 1996).

Although GM foods carry notable benefits, the public still remains skeptical about GM foods, as can be readily seen via the anti-GMO opinions expressed about the issue on social media, such as Twitter (Munro, Hartt, & Pohlkamp, 2015). The mass communication literature has convincingly established that various forms of media-delivered messages, including social media posts, can influence people’s opinions and behaviors on different issues (Kareklas, Muehling, & Weber, 2015).

Not surprisingly, the social media environment has created opportunities for factually inaccurate and inconsistent information about GM foods to proliferate, possibly contributing to the uncertainty skepticism that is often linked to GM foods. Previous research has confirmed that source credibility is one of the most important factors exerting influence on consumers’ evaluation on the quality of health information (Kunst, Groot, Latthe, Latthe, & Khan, 2002). To reduce the scientific uncertainty surrounding GM foods felt by consumers, educating consumers to identify the quality as well as the credibility of health information on social media is imperative due to the fact that there’s a significant amount of consumers using social media to obtain health information as relevant to themselves. Such need is particularly urgent when communicating with young millennials, those between the ages of 18 and 25 defined by the Pew Research Center (Fry, 2016), as they are the heaviest users of social media as sources of information and news (Greenwood, Perrin, & Duggan, 2016). 

Although existing research has studied general public’s risk perceptions toward GM foods (e.g., Phillips & Hallman, 2013) and found that source credibility did influence general consumers’ acceptance of GM foods (e.g., Zhang, Chen, Hu, Chen, & Zhan, 2016), no specific research has been designed to find out the effects of source credibility via social media on young millennials, nor has any research focused specifically on social media as an outlet for such health information and news. Therefore, to address this research gap in the literature, we designed a study to investigate the effects of source credibility via social media on young millennials’ risk perception toward GM foods. More significantly, such investigation is designed to compare perceptions from two different groups of samples, young millennials in the U.S. and those in China. The central research question that we asked in this study is:

What kind of influence via social media do source credibility and risk attitude have on young millennials’ risk perceptions, benefit perceptions, and purchase intentions toward GM foods?

By fulfilling the overall research goal, the present study contributes to the literature on health communication in at least three important ways: First, researchers have ascertained that the inquiry of health information via social media is an unusually complex one (e.g., Kunst et al., 2002). Our study contributes to the important role of sources in shaping young millennials’ knowledge and perceptions toward GM foods as an important antecedent of their purchase intention, which has yet to be adequately examined. Second, it is necessary to educate young millennials to effectively evaluate the credibility of various sources in online health communication, which may eventually link to any health-related decision-making.  More concretely, taking into consideration source credibility in identifying the quality of social media posts may lead to a more rational decision-making process. Third, our study extends earlier research on general public risk perceptions toward GM foods by integrating the influence of social media. Furthermore, results provide important information for health communicators because identifying sources that amplify credibility and reduce negative perceptions are relevant for designing health communication messages in reaching target consumers, especially when such risk perceptions may have a global presence.

Background: GM foods consumption in the U.S. and China

Genetically modified (GM) foods are produced from genetically modified organisms (GMOs). GMOs are organisms in which “the genetic material (DNA) has been altered in a way that does not occur naturally by mating and/or natural recombination. The technology is often called ‘modern biotechnology’ or ‘gene technology,’ sometimes also ‘recombinant DNA technology’ or ‘genetic engineering’ (World Health Organization, 2014, “Frequently asked questions,” para.2). GM foods were released into the market in the 1990s. Since that time, modern biotechnology has developed rapidly.

The U.S. now leads the world in producing GM crops (James, 2016). According to the U.S. Department of Agriculture, Economic Research Service (2017), currently 89% of U.S. corn acres are planted with genetically engineered, herbicide-tolerant (HT) seeds and 85% of domestic cotton acres are produced with genetically engineered, insect-resistant seeds. Similarly, China also leads the world in the research and development of biotechnology (Zheng, Gao, Zhang, & Henneberry, 2017). GM crops is the core product of this technology and has been widely breeding in China (Zhang et al; 2016). According to a report, China is the eighth largest grower of GM crops in 2016 (James, 2016). Although biotechnology has been advanced rapidly, there has been increasing concern and discussion about the regulation and safety of GM foods in China.

To date, no substantially adverse health effects brought by GM foods have been documented. According to a comprehensive report released by the National Academies of Sciences, Engineering, and Medicine (2016), the GM foods on the market are safe to eat and do not injure the environment. Regulatory authorities such as China’s Ministry of Agriculture and the Food and Drug Administration (FDA) in the U.S. have also insisted that GM foods sold on the markets have met the safety requirements (e.g., FDA, 2018). Furthermore, scientific research has indicated biotechnology can help to solve the problems of food shortage and production (Clarke, 1997; Shapiro, 1999), improve the nutrient contents of foods and reduce the use of toxic pesticides (Noussair, Robin, & Ruffieux, 2004).

However, many Americans and Chinses still maintain a negative attitude toward GM foods and assume that risk is involved in eating GM foods. They persisted that the potential side effects have not yet discovered, and GM crops can cause environmental pollution (Zheng et al., 2017). These controversies have led to increasing criticism of GM foods in the U.S. and China, which is especially obvious on Twitter and Weibo, a Chinese version of Twitter (Munro et al., 2015). This particular phenomenon reveals a significant knowledge gap between what is considered acceptable in science and what is socially accepted.

Given that GM foods is a hot topic discussed by both Chinese and American social media users, and both China and the U.S. are the world leading countries that produce GM crops (James, 2016), by examining the similarities and differences between young U.S. and Chinese millennials’ perceptions and purchase intentions toward GM foods, the results of this study can contribute to behavioral science literature in the field of health and risk communication, meanwhile, provide practical suggestions for heath communicators in both the U.S. and China.

Literature Review and Theoretical Framework

The role of source credibility in health communication

Source credibility has been an important area of research in persuasion communication for quite some time. Hovland and Weiss (1951) conducted the first study considering source credibility as a theoretical construct. This construct indicated that individuals are more likely to be persuaded when the information source appears to be credible (Hovland, Janis, & Kelly, 1953). That is, the credibility of the communicator influences the response to the communication. As stated by Hovland, Janis and Kelly (1953): “the very same presentation tends to be judged more favorably when made by a communicator of high credibility than by one of low credibility” (p. 35).

Credibility is defined by Callison (2001) as “the judgments made by a message recipient concerning the believability of a communicator” (p.220). Existing literature demonstrates that expertise and trustworthiness are the two major factors of the credibility of a communicator (e.g., Grewal, Gotlieb, & Marmorstein, 1994; Hovland et al., 1953). Expertise refers to “the extent to which a communicator is perceived to be a source of valid assertions” (Hovland et al., 1953, p. 21). Research has confirmed the following dimensions can be used to explain the concept of expertise: authoritativeness (McCroskey, 1966), competence (Whitehead, 1968), knowledge and experience (Ohanian, 1991). Trustworthiness refers to “the degree of confidence in the communicator’s intent to communicate the assertions he considers most valid” (Hovland et al., 1953, p.21). According to Ohanian (1991), dependable, reliable, sincere, trustworthy, and honest are the dimensions of trustworthiness. Source trustworthiness also deals with the self-interest of communicators (Kareklas, Muehling, & Weber, 2015). For example, individuals tend to judge a communicator as less trustworthy when they find that the communicator gains benefits from persuading them, which leads to a less persuasive effect (Kelman & Hovland, 1953).

Source credibility is a key construct in public relations research since it is a positive characteristic of public relations’ message source and can act as buffer in crisis communication (DiStaso, Vafeiadis, & Amaral, 2015). Previous research has suggested that the selection of a credible spokesperson can enhance the acceptance of the message and may result in desirable attitude changes (Chebat, Filiatrault, & Perrien, 2001; Coombs, 2014). In addition, Coombs (2014) indicated that organizations perceived as having high credibility by stakeholders can be more effectively in dealing with rumors attacking the organizations. 

Source credibility has also been widely studied in health communication. Research finds that consumers evaluate the quality of health information by looking into the accuracy, believability, trustworthiness, truthfulness, readability, and completeness of the information (Bates, Romina, Ahmed, & Hopson, 2006; Hu & Sundar, 2010). It is interesting to see that previous research has generated inconsistent findings about the effects of source credibility on people’s evaluation of the quality of health information. Some studies found people took source credibility into account when judging the quality of the health information. One study used semi-structured interviews and found some people use scientific evidence as a cue for trustworthiness to identify the quality of health information (Stavir, Freeman, & Burroughs, 2003). Another study found accuracy, credibility, currency, clarity, and ease of understanding the health content are the crucial criteria for college students to assess the quality of health websites (Escoffery et al., 2005).

However, other work has indicated people often make little use of source credibility in evaluating the quality of the health information on the Internet. For example, Bates and colleagues (2006) used the web pages of the National Cancer Institute (NCI), the American Lung Association (ALA) and the American Cancer Society (ACS) as credible information sources and a generic web page without specific information source to test source credibility. The results indicated that participants who received lung cancer prevention information from the web page of the NCI, ALA and ALS did not perceive the information to be more trustworthy, truthful or complete than participants who received the same information from a generic web page. In addition, Hu and Sundar (2010) found people did not perceive health information concerning “the use of sunscreen” and “raw milk consumption” to be more credible when the information is from a professional source (e.g., a doctor) than from a layperson. If people pay little attention to source credibility when evaluating the quality of health information, the persuasive effects of health information from a credible source would be jeopardized.

Previous research has also suggested that health risk messages provided by a credible source can lead to a greater message compliance (e.g., Schouten, 2008; Umeh, 2012). For example, Karelkas et al. (2015) found that source credibility significantly influences consumers’ attitudes and behavioral intentions for vaccination.

Although previous studies have examined the effects of perceived source credibility on public reactions to the health risk information related to GM foods (Frewer, Scholderer, & Bredahl, 2003; Zhang et al., 2016), mixed results were generated. Based on the persuasion model developed by Hovland, Zhang et al (2016) investigated the influence of source credibility on consumers’ acceptance of GM foods in China. The study found biotechnology research institutes, government offices devoted to the management of GMOs, and GMO technological experts are three professional and credible sources, which can effectively persuade consumers in China to accept GM foods. However, sources such as non-GMO experts, foods companies, and anonymous information found on the Internet were considered as less believable sources, which are unable to lead to favorable attitudes toward GM foods. In contrast, Frewer, Scholderer and Bredahl (2003) found that the perceptions of the information source characteristics (expertise and trustworthiness) didn’t have a significant influence on changing participants’ risk attitude towards GM foods. Rather, they argued that pre-existing attitudes toward GM foods influenced participants’ perceptions of source credibility.

All told, there is a lack of consistency about the effects of source credibility on consumers’ attitudes and behavioral intentions toward health issues. Further, there is a dearth of findings on the effects of source credibility on young U.S. and Chinese millennials’ risk perceptions – whether concerning GM foods or other health issues. Thus, this study focused on the effects of source credibility via social media on young American and Chinese millennials’ perceptions toward GM foods.

The link of risk attitude to source credibility

People’s risk attitude is another antecedent to the formation of risk perceptions (Phillips & Hallman, 2013). Risk attitude is “a person’s standing on the continuum from risk aversion to risk seeking” and can be seen as a personality trait (Weber, Blais, & Betz, 2002, p.264). Considering risk attitude as a trait characteristic, there are generally existing two subgroups of individuals (Goldstein, Johnson, & Sharpe, 2008): “those who have a tolerance and even a preference for risk and those who are more cautious and would prefer to avoid risk” (Phillips & Hallman, 2013, pp.739-740).

Comparing the personality traits between managers and entrepreneurs, Stewart and Roth (2001) have found entrepreneurs are more willing to take risks, prefer to think flexibly, and bear more responsibility than managers. Furthermore, research has found that risk seeking individuals are more likely to achieve personal success (MacCrimmon & Wehrung, 1990). In addition, Weber, Blais, and Betz (2002) found people’s risk attitude was associated with gender, specifically “women appeared to be more risk averse in all domains (financial, health/safety, recreational, ethical, and social) except social risk” than men (p. 263).

However, there have not been sufficient studies to date that have identified the role of risk attitude in consuming the information related to GM foods. Given the fact that the GM foods market is rapidly growing and continues influencing people’s daily food consumption, it is necessary to understand how people’s personality traits such as risk attitude may influence their perceptions toward GM foods.

Therefore, drawing from previous research on source credibility and risk attitude, we proposed the following set research questions and hypotheses to guide our study:

RQ1: Will different sources of information demonstrate different levels of source credibility?

RQ2: If different sources of information demonstrate different levels of source credibility, how would such differences influence participants’ risk perceptions, benefit perceptions, and purchase intentions for GM foods?

RQ3: Will participants with different levels of risk attitudes have different risk perceptions, benefit perceptions, and purchase intentions for GM foods?

H1a: Risk-seeking participants will have lower risk perceptions toward GM foods than risk-averse participants

H1b: Risk-seeking participants will perceive more benefits of GM foods than risk-averse participants.

H1c: Risk-seeking participant will have higher purchase intentions toward GM foods than risk-averse participants.

In addition, we also want to know whether source credibility and risk attitude will have an interaction effect on participants’ risk perceptions, benefit perceptions, and purchase intentions for GM foods. Thus, the following research question is proposed:

RQ4: For participants who have different risk attitudes, what are the similarities and differences between their risk perceptions, benefit perceptions, and purchase intentions for GM foods in response to different levels of source credibility via social media?

Methods

Research design

This study used a 2 (risk attitude: risk averse vs risk seeking) x 4 (source credibility: government vs food company vs social media influencer vs scientist) between-subjects design. The first factor, risk attitudes, was measured. The other factor, source credibility, was manipulated in the experiment. Each individual participant was randomly assigned to one of the four experimental condition which presented one of the four different information sources. Different sources (government, food company, social media influencer and scientist) were incorporated into the identical fictional social media posts. The fictional social media posts were designed for Twitter and Weibo. The tweet and Weibo message outline the benefits of GM foods. To better approximate real-world exposure to a Tweet or a Weibo message, all the messages were designed to look like they appear on each information sources’ Twitter/Weibo desktop version homepage.

The questionnaire and the fictional social media post were first created in English and were then translated into Chinese and verified by two bilingual researchers. Two versions of the questionnaire (Chinese and English) were pretested in each country (n = 40 per country) to allow final adjustments before carrying out the main study.

For the experiment conducted in the U.S., FDA was used as the government source since it regulates the safety of foods and supervises the production of GM foods. Louisa Stark was used as the scientist source since she is the director of Genetic Science Learning Center in the University of Utah and she is an expert in the field of biology. Monsanto Company was used as the food company source because Monsanto is a leading producer of genetically modified seeds. Rolf Degen was used as the social media influencer source. He was a science writer and book author in psychology, neuroscience and evolution and he had over 3K followers on Twitter. The reason for choosing Rolf Degen as the source of social media influencer is because he has tweeted information about GM foods in real life. In order to differentiate social media influencer and scientist, specific modifications were made on the social media influencer stimuli. Rolf Degen was framed as “bestselling author, interested in science, evolution and history” with 125K followers. The scientist was framed as “Ph.D. Director, Genetic Science Learning Center, University of Utah. Research Professor, Human Genetics, University of Utah.”

For the experiment conducted in China, China’s Ministry of Agriculture was used as the government source whose function is like FDA including regulating GM foods and issuing GMO safety certificate. Ning Yan, a professor of Tsinghua University whose research interest is structural biology, was used as the scientist source. Monsanto Company (China) was used as the food company source. Sijin Chen was selected as the social media influencer source since he has over 500K followers on Weibo and has posted information related to GM foods on Weibo previously. Due to the concentrated Chinese social media users on Weibo, a popular account on Weibo usually has followers over 500K.

Participants, procedure and measures

The population studied in this research are young millennials in the U.S. and China. Participants in the U.S. were recruited from two large universities in the Southeastern region of the United States. Participants in China were recruited from three different universities in mainland China. Only young millennials students, those between the ages of 18 and 25 are selected. A total of 517 millennials participated in this study with 279 from China and 238 from the U.S. Among them, 242 in China and 207 in the U.S. completed the study. Thus, the final sample consists of 449 completed responses from both countries.

Upon beginning the study, participants were asked about their frequency of social media use in daily life. The second section measured participants’ level of awareness for several controversial health and risk issues, such as GM foods.

The third section was used to assess the level of perceived credibility of the four different information sources. The measure of perceived source credibility was adapted from Ohanian’s source credibility- trustworthiness subscale (Ohanian, 1990). A brief description of the information source adapted from dictionary definitions or the organization’s mission statement appeared before the participants answered the related questions. For example, participants would read the definition of FDA: “The Food and Drug Administration is a federal agency of the United States Department of Health and Human Services, whose mission is to protect the public health and advance the public health” before they answered the FDA related questions. Participants were asked to assess the perceived credibility of each information source by rating seven 7-point item pairs.

In the fourth section, participants were asked about their attitudes toward risk in order to measure their willingness to take risk. The measure of risk attitude was adapted from the health/safety subscale of Weber, Blais and Betz’s (2002) risk taking scales. Eight items were included, for example, “buying an illegal drug for your own use,” which were measured by the seven-point Likert scale from “Very unlikely” to “Very likely.” Reponses to these items were averaged together. A median split was used to divide participants into two different groups: risk seeking and risk averse.

After the fourth section, participants were randomly assigned to one of the four stimuluses. Participants were invited to read the designed tweet/Weibo message sent from one of the four information sources. After viewing the stimuli for a minimum of 20 seconds, participants were asked to answer the manipulation check items and several survey questions about their perceptions of the risks and benefits of GM foods and future purchase intentions.

Two items were used to check the manipulation of information source: (a) “As best you can recall, the tweet/Weibo message you just saw was sent by?” by selecting one of the following options: the government, a scientist, a food company, a social media influencer or not sure, and(b) “As best you can recall, which of these categories best describes the profile photo of the Twitter/Weibo account that you just saw?” by selecting one of the following options: personal picture, the name of a government agency, corporate logo, other (please specify) or not sure. In the Chinese version, the options for item (b) were: personal picture, the name of a government agency, corporate logo, animal, other (please specify) and not sure.

Risk perception of GM foods was measured by adapting the perceived risk scales developed by Thelen, Yoo, and Magnini (2011), for example, “Eating a genetically modified food is risky.” Benefit perception of GM foods was measured by asking participants to rate five benefit-related items created by researchers of this study to fit the scenario of GM foods. One example is “Genetically modified foods can help to solve the problems of food shortage and production.” Purchase intention was measured by adapting the item from Loebnitz, Schuitema and Grunert (2015), which asked “How likely are you to buy genetically modified foods in the future?” All items were measured using seven-point Likert scales ranging from “1 = strongly disagree” to “7 = strongly agree” or from “1 = very unlikely” to “7 = very likely.” Finally, participants were asked to answer basic demographic questions such as gender, age, education, ethnicity, major, religion and political party affiliation and political ideology.

Manipulation check

This study used crosstab to check the manipulations. For the study conducted in the U.S., the results of manipulation check for the first manipulation item, “as best you can recall, the tweet you just saw was sent by,” indicated among the 53 participants who viewed the government stimuli, 62% (N = 33) of them chose the right answer which means they answered the tweet was sent by government. Forty-eight participants saw the food company stimuli and 73% (N = 35) of them chose the right answer. Among the 53 participants who saw the scientist stimuli, 55% (N = 29) chose the right option. Fifty-three participants were exposed to the social media influencer stimuli and 70% (N = 37) answered the first manipulation item correctly. For the second manipulation item “as best you can recall, which of these categories best describes the profile photo of the Twitter account that you just saw,” results showed that among the 53 participants who viewed the government stimuli, 60% (N = 32) of them chose the right answer. Among 48 participants who saw the food company stimuli, 81% (N = 39) of them chose the right answer. Fifty-three participants viewed the scientist stimuli and 76% (N = 40) answered this manipulation item correctly. Among the 53 participants who saw the social media influencer stimuli, 74% (N = 39) of them answered the second manipulation item correctly.

For the study conducted in China., the results of manipulation check for the first manipulation item indicated among the 59 participants who viewed the government stimuli, 44% (N = 26) of them chose the right answer. Sixty-four participants saw the food company stimuli and 47% (N = 30) of them chose the right answer. Among the 58 participants who saw the scientist stimuli, 24% (N = 14) of them chose the right option. Sixty-one participants were exposed to the social media influencer stimuli and 71% (N = 41) answered this manipulation item correctly. For the second manipulation item, results showed that among the 59 participants who viewed the government stimuli, 44% (N = 26) of them chose the right answer. Among 64 participants who saw the food company stimuli, 53% (N = 34) of them chose the right option. Fifty-eight participants were invited to view the scientist stimuli and 38% (N = 22) answered this manipulation item correctly. Among the 61 participants who viewed the social media influencer stimuli, 85% (N = 52) of them answered the second manipulation item correctly.

Results

Reliability analyses

A series of reliability analyses were run to test the reliability coefficients of the source credibility scale, the risk attitude scale, the risk perception scale, and the benefit perception scale (see Table 1). Results indicated that the internal consistency of all above scales was acceptable.

RQ1: Will different sources of information demonstrate different levels of source credibility?

Results indicated that each information source had different levels of source credibility. For the study conducted in the U.S., participants perceived scientist (M = 5.39, SD = 1.10) and government (M = 5.29, SD = 1.36) had a higher level of source credibility. On the contrary, participants believed company (M = 3.78, SD = 1.43) and social media influencer (M = 3.56, SD = 1.22) had a lower level of source credibility. For the study conducted in China, participants perceived scientist (M = 4.53, SD = 1.24) and government (M = 4.17, SD = 1.19) were the information sources with a higher level of credibility. However, company (M = 3.44, SD = 1.15) and social media influencer (M = 3.23, SD = 1.15) were perceived by participants as the information sources with a lower level credibility.

Paired-samples T Test was used to test the differences among the levels of credibility of the four different information sources. For the study conducted in the U.S., results suggested there was a significant difference between the government’s credibility (M = 5.29, SD = 1.36) and the company’s credibility ((M = 3.78, SD = 1.43); t (206) = 12.31, p < .001). There were also significant differences between the government’s credibility (M = 5.29, SD = 1.36) and the social media influencer’s credibility ((M = 3.56, SD = 1.22); t (206) = 15.30, p < .001), the scientist’s credibility (M = 5.39, SD = 1.10) and the company’s credibility ((M = 3.78, SD = 1.43); t (206) = 13.65, p <. 001), the scientist’s credibility (M = 5.39, SD = 1.10) and the social media influencer’s credibility ((M = 3.56, SD = 1.22); t (206) = 17.89, p < .001), and the company’s credibility (M = 3.78, SD = 1.43) and the social media influencer’s credibility ((M = 3.56, SD = 1.22); t (206) = 2.00, p = .047). However, there was a nonsignificant difference between the government’s credibility (M = 5.29, SD = 1.36) and the scientist’s credibility ((M = 5.39, SD = 1.10); t (206) = -.991, p = .323).

For the study conducted in China, results indicated there was a significant difference between the government’s credibility (M = 4.17, SD = 1.19) and the company’s credibility ((M = 3.44, SD = 1.15); t (241) = 8.97, p < .001). There were also significant differences between the government’s credibility (M = 4.17, SD = 1.19) and scientist’s credibility ((M = 4.53, SD = 1.24); t (241) = – 4.52, p < .001), the government’s credibility (M = 4.17, SD = 1.19) and the social media influencer’s credibility ((M = 3.23, SD = 1.15); t (241) = 10.47, p < .001), the scientist’s credibility (M = 4.53, SD = 1.24) and the company’s credibility ((M = 3.44, SD = 1.15); t (241) = 12.79, p <. 001), the scientist’s credibility (M = 4.53, SD = 1.24) and the social media influencer’s credibility ((M = 3.23, SD = 1.15); t (241) = 13.40, p < .001), and the company’s credibility (M = 3.44, SD = 1.15) and the social media influencer’s credibility ((M = 3.23, SD = 1.15); t (241) = 2.54, p = .012).

To answer RQ 1, we found that in both countries different sources of information demonstrate different levels of source credibility.

RQ2: If different sources of information demonstrate different levels of source credibility, how would such differences influence participants’ risk perceptions, benefit perceptions, and purchase intentions for GM foods?

To test RQ2, we ran three two-way ANOVA tests to measure the main effects of perceived source credibility on risk perceptions, benefit perceptions and purchase intentions. For the study conducted in the U.S. (see Table 2), according to the results of the two-way ANOVA, there were no significant main effects of perceived source credibility on participants’ risk perceptions (F (3, 199) = 2.26, p = .083, ); benefit perceptions (F (3, 199) = 1.39, p = .247, ) and purchase intentions (F (3, 199) = 1.43, p = .235, ) for GM foods.

For the study conducted in China (see Table 3), there were no significant main effects of perceived source credibility on participants’ risk perceptions (F (3, 234) = .65, p = .584,)) and purchase intentions toward GM foods (F (3, 234) = .54, p = .658,)). However, we did find a significant main effect of perceived source credibility on benefit perceptions of GM foods (F (3, 234) = 2.96, p = .033, ). Tukey was chosen as the post hoc tests. The result indicated that participants who viewed the government stimuli (M = 4.04, SD = .97) perceived more benefits of GM foods than the participants who saw the scientist stimuli (M = 3.49, SD = .96), although the result of this study has suggested that scientist was perceived as having a higher level of credibility than government.

Therefore, to answer RQ2, the results indicated that in both countries, different levels of perceived source credibility did not generate different risk and benefit perceptions of GM foods and different purchase intentions for GM foods. However, we found participants in China who viewed the government stimuli perceived more benefits than the participants who saw the scientist stimuli.

RQ3 and Hypothesis Testing

Three two-way ANOVA were used to measure the main effects of risk attitudes on each dependent variable for RQ3 and the H1 set of hypotheses. For the study conducted in the U.S. (see Table 2), results of the two-way ANOVA showed that there were no significant differences between the two risk attitudes on risk perceptions (F (1, 199) = .021, p = .884, ), and benefit perceptions (F (1, 199) = 2.30, p = .131, ) of GM foods. However, there was a significant main effect of risk attitude on purchase intentions toward GM foods (F (1, 199) = 6.98, p = .009, ). Risk-seeking participants (M = 4.54, SD = 1.45) had more purchase intention for GM foods than risk-averse participants (M = 4.03, SD = 1.47). Therefore, H1c was supported while H1a and H1b were not supported in the U.S sample.

For the study conducted in China (see Table 3), results of the two-way ANOVA showed that there was no significant difference between the two risk attitudes on risk perception of GM foods (F (1, 234) = 1.55, p = .214, ). However, there were significant main effects of risk attitude on benefit perceptions (F (1, 234) = 8.49, p = .004, ) and on purchase intentions (F (1, 234) = 15.81, p < .001, ) toward GM foods. Risk-seeking participants (M = 3.98, SD = 1.05) perceived more benefits of GM foods than risk aversion participants (M = 3.57, SD = 1.08). In addition, risk-seeking participants (M = 3.60, SD = 1.46) had a higher purchase intention for GM foods than risk-aversion participants (M = 2.90, SD = 1.21). Therefore, the hypothesis H1b and H1c were supported but H1a was not supported.

To answer RQ3, the above results indicate that participants in the U.S. with different levels of risk attitudes had different purchase intentions for GM foods but had similar risk and benefit perceptions. In addition, participants in China with different levels of risk attitudes had different benefit perceptions and purchase intentions for GM foods while had similar risk perception.

RQ4: For participants who have different risk attitudes, what are the similarities and differences between their risk perceptions, benefit perceptions, and purchase intentions for GM foods in response to different levels of source credibility via social media?

In order to investigate the interaction effects of perceived source credibility and risk attitude on the three dependent variables for RQ4, a two-way ANOVA was conducted on each dependent variable. For the U.S. sample (see Table 2), the results indicated that there were no significant differences among the four different information sources and the two different risk attitudes on risk perceptions (F (3, 199) = .78, p = .508, ); benefit perceptions (F (3, 199) = .88, p = .453, ) and purchase intention (F (3, 199) = .01, p = .998, ).

For the sample in China (see Table 3), results found that there was a significant interaction effect of perceived source credibility and risk attitude on risk perceptions of GM foods (F (3, 234) = 2.72, p = .045, = .03). This result suggested that risk-seeking participants and risk-aversion participants were affected differently by perceived source credibility on risk perceptions of GM foods. Pairwise comparisons were used to test the simple effects. Specifically, risk-aversion participants’ risk perceptions of GM foods were similar no matter they saw government stimuli (M = 3.83, SD = 1.37); social media influencer stimuli (M = 4.11, SD = 1.35); scientist stimuli (M = 4.11, SD = 1.42) or company stimuli (M = 3.61, SD = 1.00). However, among the risk-seeking participants, those who viewed the scientist stimuli (M = 3.39 SD = .96) had a significantly lower risk perception of GM foods than whose who saw the company stimuli (M = 4.12, SD = 1.24). However, results found that there were no significant differences among the four different information sources and two different risk attitudes on benefit perceptions (F (3, 234) = .97, p = .408,  = .01) and purchase intentions (F (3, 234) = .76, p = .520,  = .01).

Therefore, to answer RQ4, the results showed that in the U.S. risk-averse and risk-seeking participants were not affected differently by source credibility on any of the three dependent variables. However, the results showed that in China risk-averse and risk-seeking participants were affected differently by perceived source credibility on risk perceptions of GM foods but not on benefit perceptions and purchase intentions for GM foods.

Discussion and Conclusions

The aim of this study is to investigate the influence of perceived source credibility and risk attitude on young Chinese and U.S. millennials’ risk perceptions, benefit perceptions, and purchase intentions toward GM foods. In sum, four research questions and one set of hypotheses were tested. The results of this study suggested that government, scientist, food company and social media influencer were perceived as having different levels of source credibility. For the four information sources adopted in this study, we found scientists and the government were perceived by participants as high credibility information sources in both the U.S. and China. In contrast, food company and social media influencer were perceived as low credibility information sources in both the two countries.

In addition, this study also found risk attitude had a significant influence on participants’ benefit perceptions and purchase intentions for GM foods. Specifically, risk-seeking participants had higher benefit perceptions and purchase intentions for GM foods than risk-averse participants. This study also found risk-averse and risk-seeking participants in China were affected differently by perceived source credibility on risk perceptions of GM foods. Among the risk-seeking participants, those who viewed the scientist stimuli had a significantly lower risk perception of GM foods than whose who saw the company stimuli.

Previous research in persuasion has well established that many forms of media-delivered messages, such as online PSA and social media posts, are able to affect public opinion on various issues (e.g., Kareklas et al., 2015; Mangold & Faulds, 2009). This current study extends this literature in a health risk-related context (i.e., the current debate on the safety of GM foods) through investigating the effects of perceived source credibility via social media and risk attitude on young Chinese and U.S. millennials’ risk perceptions, benefits perceptions and purchase intentions towards GM foods. Recognizing a growing amount of people using social media to obtain health information as relevant to themselves (Witteman & Zikmund-Fisher, 2012), this study carefully designed the stimuluses as they would appear on the respective information sources’ Weibo/Twitter account. Since to our knowledge, this is the first study examining the interaction effects of perceived source credibility and risk attitude on young Chinese and U.S. millennials’ risk perceptions, benefits perceptions and purchase intentions, we believe our study can make important theoretical and practical contributions to the health risk communication literature in the field of public relations research.

The results of this study indicated that there was a significant interaction effect of perceived source credibility and risk attitude on young Chinese millennials’ risk perception of GM foods. Specifically, among the risk-seeking participants, those who viewed the scientist stimuli had a significantly lower risk perception of GM foods than whose who saw the company stimuli. This finding provided evidence that the two antecedents of risk perceptions, perceived source credibility and risk attitude, can exert joint effect on risk perception. In addition, this finding can also provide a practical suggestion for public relations practitioners in China that is public relations practitioners in China should consider using scientists as the information source in order to successfully transmit the benefits of GM foods via social media.

This study also found young Chinese millennials’ risk attitude significantly influenced their benefits perceptions and purchase intentions for GM foods, but did not have an impact on their risk perceptions of GM foods. The result suggested that even though risk-seeking

participants and risk-averse participants had similar risk perceptions of GM foods, risk-seeking participants can perceive more benefits of GM foods and would be more likely to buy GM foods. One possible explanation for this result could be that risk-averse participants and risk-seeking participants are different in their benefit perceptions rather than risk perceptions. Compared to risk-averse people, risk-seeking individuals can see more benefits of a risk, thus they are more willing to take the risk. Therefore, public relations practitioners should spend more time to communicate the benefits of GM foods to risk-averse individuals.

The results of this study indicated perceived source credibility via social media has little or no effect on young Chinese and U.S. millennials’ risk perceptions, benefit perceptions and purchase intentions for GM foods. Here we provide two reasons that may explain the results. Our findings are perhaps understandable if we consider those young millennials’ knowledge and attitude toward GM foods. Results of this study show that young millennials in China and the U.S. felt that they were not very informed about GM foods and they had a general negative attitude towards GM foods. Therefore, it is not particularly surprising that the use of even a highly credible source is insufficient for persuading this audience to change their attitudinal evaluations of GM foods. Rather, a long-term education strategy may be needed to truly shift prevailing opinions. Such sentiments have been shared by Uzogara (2000), who noted that “the public needs to be sufficiently educated on genetic engineering of any product to enhance acceptability of such a food.” (p.202). In addition, according to Frewer et al. (2003), the reason risk communication sometimes failed is because the message was sent from the expert view of what should be known by the public rather from the view of what the public are really concerned about. Therefore, in order to make young millennials be more willing to accept the positive message related to GM foods, researchers and public relations practitioners should understand what are the things that this audience is really concerned about.

Another reason that source credibility via social media has little or no effect on young millennials’ risk perceptions, benefit perceptions and purchase intentions for GM foods is that participants may not take source credibility fully into account when consuming the information from Weibo and Twitter. As suggested by some former studies, Internet users often make little use of source credibility in evaluating the quality of the health information on the Internet (e.g., Bates et al., 2006; Hu & Sundar, 2010). Therefore, the persuasive effects of the health information from a credible source would be jeopardized. Since that researchers and health public relations practitioners should continuously search for ways to improve the health literacy and health information-seeking skills of young millennials, which might influence how young millennials use cues like source credibility when forming opinions.

Limitations and Recommendations for Future Research

The first limitation of this work is the student sample, which is also a limitation of many social science research projects. Future studies should include more diverse participants to investigate the influence of perceived source credibility and risk attitude on perceptions of GM foods. The second limitation is the design of the stimulus as the desktop version. Since participants can also use their mobile phones to access the questionnaire, it would be better if the stimulus was also designed for the mobile version of Weibo. As the desktop version stimulus may not suit the phones’ screen very well, the visual impacts of the stimulus maybe jeopardized. Third, this study only examined government, company, scientist and social media influencer as information sources. The impacts of, for example, family members and friends on the dependent variables remain to be investigated.

Research has found that the persuasive strength of message has a significant influence on persuasive appeal and attitude change (Frewer, Howard, Hedderley, & Sheperd, 1997). Since this study only used one message to communicate the benefits of GM foods, it didn’t manipulate the level of the persuasive strength of the message. Thus, in order to effectively communicate the benefits of GM foods and other health topics, future studies should examine the effects of persuasive strength on public attitudes and behaviors. In addition, the format (video, audio, text) of the message might also influence the persuasive appeal of the message. Research has found that video can generate stronger and more enduring attitude change compared to text since video has a higher level of vividness than text (Coyle & Thorson, 2001). As this study relied solely on text-based stimuli to communicate the benefits of GM foods, future studies could test whether using alternative formats enhances the persuasiveness appeal of the message.

References

Bates, B. R., Romina, S., Ahmed, R., & Hopson, D. (2006). The effect of source credibility on consumers’ perceptions of the quality of health information on the Internet. Medical Informatics and the Internet in Medicine, 31(1), 45-52.

Burke, D. (1997). What biotechnology can do for the food industry. Food Science and Technology Today, 11, 202-209.

Callison, C. (2001). Do PR practitioners have a PR problem?: The effect of associating a source with public relations and client-negative news on audience perception of credibility. Journal of Public Relations Research, 13, 219–234.

Chebat, J. C., Filiatrault, P., & Perrien, J. (1990). Limits of credibility: The case of political persuasion. The Journal of Social Psychology, 130(2), 157-167.

Clarke, A. (1997). The impact of biotechnology in a changing world. -Paper based on the closing plenary lecture of the International Biotechnology Symposium (10th: 1996: Sydney). Australasian Biotechnology, 7(2), 96.

Coombs, W. T. (2014). Ongoing crisis communication: Planning, managing, and responding. Sage Publications.

Coyle, J. R., & Thorson, E. (2001). The effects of progressive levels of interactivity and vividness in web marketing sites. Journal of Advertising, 30(3), 65-77.

DiStaso, M. W., Vafeiadis, M., & Amaral, C. (2015). Managing a health crisis on Facebook: How the response strategies of apology, sympathy, and information influence public relations. Public Relations Review, 41(2), 222-231.

Escoffery, C., Miner, K. R., Adame, D. D., Butler, S., McCormick, L., & Mendell, E. (2005). Internet use for health information among college students. Journal of American College Health, 53(4), 183-188.

Food and Drug Administration (2018). Questions and answers on food from genetically engineered plants. Retrieved from:https://www.fda.gov/Food/IngredientsPackagingLabeling/GEPlants/ucm346030.htm

Frewer, L. J., Howard, C., Hedderley, D., & Shepherd, R. (1997). The elaboration likelihood model and communication about food risks. Risk Analysis, 17(6), 759-770.

Frewer, L. J., Scholderer, J., & Bredahl, L. (2003). Communicating about the risks and benefits of genetically modified foods: The mediating role of trust. Risk Analysis, 23(6), 1117-1133.

Fry, R. (2016). Millennials overtake Baby Boomers as America’s largest generation. Pew Research Center, 25.

Goldstein, D. G., Johnson, E. J., & Sharpe, W. F. (2008). Choosing outcomes versus choosing products: Consumer-focused retirement investment advice. Journal of Consumer Research, 35(3), 440-456.

Greenwood, S., Perrin, A., & Duggan, M. (2016). Social media update 2016: Facebook usage and engagement is on the rise, while adoption of other platforms holds steady. Pew Research Center.

Grewal, D., Gotlieb, J., & Marmorstein, H. (1994). The moderating effects of message framing and source credibility on the price-perceived risk relationship. Journal of Consumer Research, 21(1), 145-153.

Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion: psychological studies of opinion change. New Haven, CT: Yale University Press.

Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication effectiveness. Public Opinion Quarterly, 15(4), 635-650.

Hu, Y., & Shyam Sundar, S. (2010). Effects of online health sources on credibility and behavioral intentions. Communication Research, 37(1), 105-132.

James, C. (2016). Global status of commercialized biotech/GM Crops: 2016. ISAAA Brief, (52).

Kareklas, I., Muehling, D. D., & Weber, T. J. (2015). Reexamining health messages in the digital age: a fresh look at source credibility effects. Journal of Advertising, 44(2), 88-104.

Kelman, H. C., & Hovland, C. I. (1953). ” Reinstatement” of the communicator in delayed measurement of opinion change. The Journal of Abnormal and Social Psychology, 48(3), 327.

Kunst, H., Groot, D., Latthe, P. M., Latthe, M., & Khan, K. S. (2002). Accuracy of information on apparently credible websites: survey of five common health topics. Bmj, 324(7337), 581-582.

Loebnitz, N., Schuitema, G., & Grunert, K. G. (2015). Who buys oddly shaped food and why? Impacts of food shape abnormality and organic labeling on purchase intentions. Psychology & Marketing, 32(4), 408-421.

MacCrimmon, K. R., & Wehrung, D. A. (1990). Characteristics of risk taking executives. Management Science, 36(4), 422-435.

McCroskey, J. C. (1966). Scales for the measurement of ethos.

Mangold, W. G., & Faulds, D. J. (2009). Social media: The new hybrid element of the promotion mix. Business Horizons, 52(4), 357-365.

Munro, K., Hartt, C. M., & Pohlkamp, G. (2015). Social Media Discourse and Genetically Modified Organisms. The Journal of Social Media in Society, 4(1), 38-65.

National Academies of Sciences, Engineering, and Medicine. (2016). Genetically engineered crops: experiences and prospects. National Academies Press.

Noussair, C., Robin, S., & Ruffieux, B. (2004). Do consumers really refuse to buy genetically modified food?. The Economic Journal, 114(492), 102-120.

Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3), 39-52.

Ohanian, R. (1991). The impact of celebrity spokespersons’ perceived image on consumers’ intention to purchase. Journal of Advertising Research, 31(1), 46-54.

Phillips, D. M., & Hallman, W. K. (2013). Consumer Risk Perceptions and Marketing Strategy: The Case of Genetically Modified Food. Psychology & Marketing, 30(9), 739-748. doi:10.1002/mar.20642 

Schouten, B. C. (2008). Compliance behavior and the role of ethnic background, source expertise, self-construals and values. International Journal of Intercultural Relations, 32(6), 515-523.

Shapiro, R. B. (1999). How genetic engineering will save our planet. The Futurist, 33(4), 28.

The Case of Genetically Modified Food. Psychology & Marketing, 30(9), 739-748.

Stavri, P. Z., Freeman, D. J., & Burroughs, C. M. (2003). Perception of quality and trustworthiness of internet resources by personal health information seekers. In AMIA Annual Symposium Proceedings (Vol. 2003, p. 629). American Medical Informatics Association.

Stewart Jr, W. H., & Roth, P. L. (2001). Risk propensity differences between entrepreneurs and managers: a meta-analytic review. Journal of Applied Psychology, 86(1), 145.

Thelen, S. T., Yoo, B., & Magnini, V. P. (2011). An examination of consumer sentiment toward offshored services. Journal of the Academy of Marketing Science, 39(2), 270-289.

Umeh, K. (2012). Does a credible source also need a fearful audience?. Journal of Applied Social Psychology, 42(7), 1716-1744.

United States Department of Agriculture, Economic Research Service. (2017). Recent trends in GE adoption. Retrieved from: https://www.ers.usda.gov/data-products/adoption-of-genetically-engineered-crops-in-the-us/recent-trends-in-ge-adoption.aspx

Uzogara, S. G. (2000). The impact of genetic modification of human foods in the 21st century: A review. Biotechnology Advances, 18(3), 179-206.

Weber, E. U., Blais, A. R., & Betz, N. E. (2002). A domain-specific risk-attitude scale: Measuring risk perceptions and risk behaviors. Journal of Behavioral Decision Making, 15(4), 263-290.

Whitehead Jr, J. L. (1968). Factors of source credibility. Quarterly Journal of Speech, 54(1), 59-63.

Witteman, H. O., & Zikmund-Fisher, B. J. (2012). The defining characteristics of Web 2.0 and their potential influence in the online vaccination debate. Vaccine, 30(25), 3734-3740.

World Health Organization (2014). Frequently asked questions on genetically modified foods. Retrieved from: http://www.who.int/foodsafety/areas_work/food-technology/faq-genetically-modified-food/en/

Zhang, M., Chen, C., Hu, W., Chen, L., & Zhan, J. (2016). Influence of Source Credibility on Consumer Acceptance of Genetically Modified Foods in China. Sustainability, 8(9), 899.

Zheng, Z., Gao, Y., Zhang, Y., & Henneberry, S. (2017). Changing attitudes toward genetically modified foods in urban China. China Agricultural Economic Review, 9(3), 397-414.