Communicating Food Safety via the Social Media

Abstract

This study examined the Chinese public’s use of Weibo (a microblog platform) and their cognitive, affective, and behavioral responses to a series of food safety crises. Based on a sample of 1,360 adult Weibo users across China, the study found that Weibo use contributed to cognitive and behavioral responses to food safety concerns, but access to other online and off-line news and information outlets was largely irrelevant. Emotional response toward the food safety incidents was a stronger predictor of both food safety risk perception and prevention action, relative to food safety incident awareness and factual awareness. Theoretical and social implications of study findings are discussed.

Keywords

emotions food safety microblogs risk communication social media

The issue of food safety has attracted a lot of public attention in the past decade in China, as an estimated 200,000 to 400,000 Chinese fell victim to food poisoning annually (Foods1, 2010). More recently, China saw the unfolding of a series of food safety disasters caused by the sale of melamine-tainted baby formula, gutter cooking oil, and clenbuterol-tainted pork, among others (iFeng, 2011). In the aftermath of these serious food safety incidents, public confidence in the Chinese government’s ability to regulate food industry plummeted (Foster, 2011). For some, the tainted baby formula incident served as a painful reminder of the 2003 SARS (severe acute respiratory syndrome) outbreak, when the Chinese government did not inform the public about the pandemic for 4 months and allowed the disease to spread to Hong Kong and other countries (Bradsher, 2008).

Although the government has taken a firmer stance on food safety regulation, the public is still suffering from the continuing food safety scares (MacLeod, 2012). This food safety concern is well reflected in public opinion polls. For instance, a Pew Global Attitudes Project (2012) reported that 41% of the respondents in 2012 considered food safety to be a very big problem in China, compared to only 12% who stated the same in 2008. Another 2012 consumer survey in China also indicated that, while 61% of the consumers had less confidence in domestic food supplies in the past year, 28% of them planned to purchase more imported foods/brands to replace domestic products (Chen & Xie, 2012).

Owing to government censorship on media and the Internet as well as a lack of institutional openness and initiative to communicate food safety issues in China, grassroots communication transmitted through a social media platform (or microblogs) has emerged to champion the food safety cause (Magistad, 2012a). The rise of social media as an alternative platform for sharing risk information has brought a significant change in how risk communication is conducted (Currie, 2009). For instance, Ding and Zhang (2010) found that students from a university in Beijing first reported the outbreak of the H1N1 flu via social media; this information was quickly spread online to the general public, before drawing attention from and compelling government officials to initiate preventive actions.

The question of how social media use might affect the public’s affective and cognitive responses, risk perceptions, and coping strategies—when confronted with food safety risk information—is a topic that has yet to be fully investigated. Moreover, existing literature is very limited in studying the public’s reaction to food safety outbreaks with a theoretical approach. The objective of this study was to explore whether affective (or emotional) responses played a mediator role between risk perceptions and its cognitive antecedents—risk awareness and factual awareness—to advance our theoretical understanding of the relationships between emotions and public response to an environmental risk. Specifically, the study examined how the Chinese public utilized Weibo, a social media platform similar to Twitter, to engage in food safety communication.

Risk Awareness, Perception, and Knowledge

According to cognitive learning theory (Ormrod, 2008), once information is rehearsed and practiced through elaboration in human brains, the information is stored in memory; this information can be subsequently retreated to deliver a response as a reaction to a stimulus. In line with this general thesis, past work suggests that knowledge of food safety issues influences the public’s assessment of their personal risks and risk perceptions (e.g., De Boer, McCarthy, Brennan, Kelly, & Ritson, 2005; Frewer, Shepherd, & Sparks, 1994; Shaw, 2003). For instance, a statewide mail survey conducted in California revealed that consumers considered themselves informed about food safety and most reported taking actions to minimize risk from potential food safety hazards (Bruhn & Schutz, 1999). Specifically, consumer knowledge of appropriate safeguards in the selection and cooking of selective foods (e.g., red meat, poultry, fish and seafood) helped them avoid unsafe consumption of foods.

The literature has suggested that different demographic groups have differential food safety knowledge and related behavior. In general, females tend to have more food safety knowledge than males; younger adults show a greater need for food safety education than older adults (Bruhn & Schutz, 1999; Rimal, Fletcher, McWatters, Misra, & Deodhar, 2001). Yet food storage practices among the majority of elderly people living at home do not meet recommended safety standards, even though their food safety knowledge was deemed acceptable (Johnson et al., 1998). Significant differences in food safety knowledge and perception across different education levels have also been identified, as better educated individuals engaged in safer practices of food purchase, handling, and preparation than their less educated counterpart (Unusan, 2007).

Previous research has also uncovered a significant correlation between the decline in public knowledge about food safety and the erosion of safe practices (Shaw, 2003). Specifically, the less food safety knowledge the public have, the more food safety risk they will experience. However, food safety knowledge associated with daily food handling practices seems to have little to do with food science knowledge, which reflects the understanding of (1) the physical, biological, and chemical makeup of food; (2) the concepts underlying food processing; and (3) food technology or the application of food science to the selection, preservation, processing, packaging, distribution, and use of safe food (Institute of Food Technology, 2014). This was evidenced by a study conducted in Ireland, which discovered that even though respondents seemed to possess a relatively high level of food safety knowledge, their food science knowledge was low (McCarthy et al., 2007). In particular, most of the respondents knew what they should be doing in their kitchen from a food safety perspective, but they did not always follow the best safe-practice guidelines.

The literature reviewed above painted an inconsistent picture of the relationships between food safety knowledge and safety handling process as well as between individuals of different demographic characteristics. Gaziano (1997) summarized three types of measures in assessing knowledge: (1) awareness of a topic, (2) closed-ended researcher-determined content, and (3) unlimited content. Closed-ended a priori food science knowledge measures may not match the public’s food safety literacy level, and an assessment of “unlimited” food science knowledge content would be impractical. Hence, the current study measured food safety knowledge—via awareness of food safety facts or factual awareness—instead of actual scientific knowledge. The following research question was proposed to explore the relationships between these cognitive factors and behavioral outcomes.

Research Question 1: Are food safety incident awareness, factual awareness, and risk perception related to food safety preventive action?

Emotions and Risk Response

Existing risk communication literature has addressed the concept of risk perception and related psychological factors, including affective or emotion appraisals (Loewenstein, Weber, Hsee, & Welch, 2001). Affect is believed to be a critical stimulus in eliciting cognitive and behavioral changes in individuals (Buck, Anderson, Chaudhuri, & Ray, 2004; Jin, Pang, Cameron, 2007; Tuner, 2006). According to the appraisal tendency theory (Lerner & Keltner, 2000), emotions are related to an individual’s cognitive response to stimuli. Past research suggested that “emotions not only arise from, but also elicit specific cognitive appraisals” (Lerner, Gonzalez, Small, & Fischhoff, 2003, p. 144). Moreover, as elaborated by Frijda (1988), emotions can draw an individual’s attention to situations that further results in changes in his or her action tendencies. Hence, as emotion is an expression of an individual’s affective state, emotional response to an external stimulus can influence his or her cognitive and behavioral response to the stimulus.

Lazarus (1991) defined emotion as “organized cognitive-motivational-relational configurations whose status changes with changes in the person-environment relationships as this is perceived and evaluated (appraisal)” (p. 38). Nabi (1999) contended that negative emotions, such as sadness or fear can “motivate problem-solving activity by forcing people to focus inward, looking for possible solutions, and/or help from others” (p. 298). Izard’s (1977) differential emotions theory proposes 10 primary human emotions: interest, joy, surprise, sadness, anger, disgust, contempt, fear, shame, and guilt. A number of these discrete emotions—including sadness, anger, disgust, contempt, and fear—seem highly relevant to the present study context. Additional research further suggested that as emotions can influence an individual’s coping actions, coping actions can also affect an individual’s emotions (Lazarus, 1993; Luce, 1998).

In general, people may experience various emotions during a crisis: anger, sadness, fright, and anxiety (Jin et al., 2007). According to Jin et al., in time of crisis, the public experiences anger when encountering threats that can adversely affect their personal well-being. Three emotions—sympathy, anger, and schadenfreude (or taking joy from others’ pain)—were found to be salient factors in institutional crisis management (Coombs & Holladay, 2005). The public often makes attributions of the cause of a risk or crisis; the stronger they believe that the relevant industries or government agencies are responsible for creating the crisis in question, the more negative their affective response toward these entities would be (Coombs & Holladay, 2005). Therefore, the public’s feeling of sympathy toward an organization such as the government is negatively associated with that government’s responsibility in preventing the crisis, while the public’s anger and schadenfreude are positively associated with that responsibility.

Public perception of food safety hazards is usually attributed to the outcome of a natural occurrence or human interference (Fife-Schaw & Rowe, 1996). In the case of the latter, there is a logical connection between awareness of food safety crises and negative emotions toward those who are responsible for causing the crises. Lazarus (1993), for instance, attributes the accountability of others as an important component in eliciting negative emotions; such attribution could lead to anger toward those individuals who are responsible for creating the risk. As food safety is an environmental health issue over which the public has very little control, the emotional response associated with the individuals or institutions that fail to safeguard food safety is particularly relevant to how the public reacts to any food safety crisis.

Past research on risk communication has yet to disentangle the dynamics between affective and cognitive responses toward risk information (see Loewenstein et al., 2001). The concept of “affect heuristic”—proposed by Slovic and Peters (2006)—highlights the significance of affect (or emotion) associated with risk perception and risk-related behavior. As pointed out by a neurophysiologist, “Connections from the emotional systems to the cognitive systems are stronger than connections from the cognitive systems to the emotional systems” (LeDoux, 1996, p. 19).

The literature that focused on how people respond to perceived risks (e.g., Fife-Schaw & Rowe, 1996; Fleming, Thorson, & Zhang, 2006) has not fully addressed how emotional responses may influence risk perception and the subsequent risk-prevention action. A recent study found that compared to positive emotions, negative emotions such as anger spread most quickly and broadly in social media platforms (Fan, Zhao, Chen, & Xu, 2013). To explore the relations between emotions and the public’s cognitive and behavioral responses to food safety concern, the following hypotheses were tested.

Hypothesis 1a: Food safety incident awareness is positively related to negative emotions associated with food safety incidents.

Hypothesis 1b: Food safety–related factual awareness is positively related to negative emotions associated with food safety incidents.

Hypothesis 1c: Negative emotions associated with food safety incidents are positively related to food safety risk perception.

Hypothesis 1d: Negative emotions associated with food safety incidents are positively related to food safety preventive action.

Communicating Food Safety

Extant research has reported that media use could affect an individual’s risk perceptions about food safety issues. For instance, respondent exposure to food safety coverage in a local newspaper and television news in Missouri was found to be predictive of their perceptions about food safety, after controlling for food safety awareness and demographic variables (Fleming et al., 2006). In particular, the number of days in a week devoted to reading a local newspaper and attention to local television food safety coverage significantly predicted the public concern about food safety. Another study reported that food safety knowledge was positively predicted by an individual’s media use, apart from other personal background factors such as social class, education level, age, gender, and marital status (De Boer, McCarthy, Brennan, Kelly, & Ritson, 2004).

The literature has shown the effectiveness of using social media platforms to disseminate food safety–related information (e.g., Newkirk, Bender, & Hedberg, 2012). For instance, Mayer and Harrison (2012) evaluated a Facebook-based intervention (i.e., “Safe Eats”) for young adults to improve food safety practices and knowledge; they found that participants who spent more time on the Facebook page also showed greater improvements in food safety attitudes and practices. A content analysis on tweets posted during the 2011 European Escherichia coli bacteria outbreak revealed that crisis coping expressions via tweets seemed dynamic, flexible, and social as well as reflective of a predominance of accommodation, information-seeking, and opposition (e.g., anger) strategies (Gaspar et al., 2014). On the down side, social media as a platform for communicating food safety risks is subject to questions of information accuracy, trust, and source credibility (Rutsaert et al., 2013).

Twitter, a short-messaging or microblogging service, is unavailable in China; Weibo is its closest substitute in China. As of July 2013, China has an estimated 560 million Internet users and over 330 million Weibo users (CNNIC, 2013). The influence of Weibo as an outlet for the Chinese to discuss “hot-button” issues in society has been growing stronger and seems unstoppable (Hewitt, 2012; Magistad, 2012b). Weibo users who are active in blogging, sharing, or accessing these controversial issues tend to be college-educated middle-class young adults between 18 and 35 years old (Gan, 2011). Food safety–related issues have been one of the “hot” topics for Weibo communication (Magistad, 2012a; Richburg, 2012). The adoption of microblogging as an alternative venue for risk communication parallels an earlier event—when the Chinese public utilized the then less censored Internet venue—instead of relying on the government-controlled media to seek updated development of the deadly SARS pandemic in 2003 (Tai & Sun, 2007).

Utilizing the short message format of Weibo (which transmits140 Chinese characters) for discussing hot-button issues resembles a “guerrilla tactic” to circumvent government censorship, even though it is not the most adequate channel to develop knowledge about health- or food safety–related topics (Lovejoy, Waters, & Saxton, 2012). To date, very little is known about the effectiveness of microblogging in influencing the Chinese public’s cognitive, affective, or behavioral response to the food safety crises through empirical research. The following hypotheses were intended to help assess the effects of Weibo on facilitating public awareness, factual awareness, and perceptions toward the food safety concern in China.

Hypothesis 2: Weibo use is positively related to user awareness of food safety incidents.

Hypothesis 3: Weibo use is positively related to food safety–related factual awareness.

Hypothesis 4: Weibo use is positively related to perceived food safety risks.

Since the existing literature has not addressed whether social media use is related to any food safety preventive action, a research question was posed below to explore this potential relationship. Based on the preliminary empirical evidence discussed above, another research question was also proposed to examine the relationships between demographic, media use, Weibo use, and food safety–related measures.

Research Question 2: Is Weibo use positively related to food safety preventive action?

Research Question 3: Are demographic characteristics, media use, Weibo use frequency, and food safety cognition factors positively interrelated?

Based on the study framework advanced above, a conceptual model is presented below to illustrate the potential interrelationships between the cognitive, affective, and behavioral factors of the proposed model as well as the likely connections between demographic factors, mediated channel use, and Weibo use frequency (see Figure 1).

Figure 1.

Proposed conceptual model.

Research Method

Sample recruitment and data collection for this study were contracted to an online survey research company in China, whose academic clients included a number of major U.S. universities. The study procedure and research instrument were submitted to the institutional review board for review and approval in advance of data collection. The survey company sent a recruitment e-mail to a national sample of adults across the country—a common practice for conducting large-scale public opinion polls and marketing research in the United States and elsewhere—offering participants an incentive of bonus points that can be accumulated to exchange for cash or consumer products. Overall, the survey yielded 1,400 valid survey responses, representing a 21.8% response rate. Two bilingual graduate students translated the survey into Chinese and back-translated it into English. Modifications were made to reconcile any differences in translation between the two translators before producing a Chinese-language survey for use.

The sample consisted of 46% males and 54% females; the mean age of the sample was 28.71 years (SD = 6.49). Just over half of the respondents were married (52.4%), and nearly 98.7% of the sample lived in an urban environment. The urban population has widespread wireless Internet access, which is the primary means of accessing the Internet in China. According to a recent report, an estimated 28% of the Chinese rural population had access to the Internet (Kan, 2013). Based on the most recent available Internet use–tracking statistics, 88.81% of the Internet users who are at least 19 years old are also Weibo users (Takamitsu, 2012). The Weibo users represented 97% (N = 1,360) of the current sample. Only Weibo users’ responses were utilized in the data analysis.

Measures

The demographic attributes measured include age, gender, marital status, education level, and monthly household income. Operational definitions for all other variables are as follow.

Media Use

Participants were asked to report the amount of time that they spent accessing television, newspaper, magazine, and online news—in the unit of “hours” per day—via a set of open-ended questions.

Internet Use

Participant use of the Internet (for any purposes) was measured in the unit of “hours” per day.

Weibo Use

Participants were asked to indicate how often they accessed Weibo sites via a 5-point scale, ranging from never (1) to at least once per day (5).

Incident Awareness

This variable assesses respondent awareness of a list of 12 major food safety incidents between 2008 and 2011. Participants were instructed to respond with either a Yes (1) or a No (0) to the following: (1) recycled gutter cooking oil, (2) melamine-tainted baby formula, (3) clenbuterol-tainted pork, (4) lime-tainted flour, (5) Nanjing crayfish, (6) Qingdao formalin-tainted ice fish, (7) Hebei fake red wine, (8) Sudan Red–tainted orange, (9) Nestlé baby food, (10) Shanghai dyed steamed bread, (11) beef extract, and (12) Mengniu-tainted milk. Specifically, gutter cooking oil is reprocessed waste oil and animal fat collected and/or extracted from sewers, grease traps, slaughterhouses, and the like. Responses to these 12 items were summed to create an index to indicate incident awareness level.

Food Safety–Related Factual Awareness

This concept gauges participants’ self-reported awareness of selective food safety facts. Four original items were constructed to assess participant response on a 5-point Likert-type scale, ranging from 1 (not aware at all) to 5 (very aware). Participants were asked how aware they were of the following facts: (1) Melamineis added to powder milk to increase the measured quantity of nitrogen (an indicator of protein); (2) the clenbuterol contained in pork could cause nausea, arrhythmia, and palpitations; (3) aflatoxin and benzopyrene, two of the most toxic chemicals, are found in gutter cooking oil; and (4) not all food additives are harmful to humans, therefore they should not be totally banned. The principal component factor analysis produced one factor with a scale reliability of .66.

Food Safety Risk Perception

This construct describes perceived level of risk associated with food supply and handling. The measures for this construct were adapted from the 12-item scale developed by Fleming et al. (2006). To ensure semantic equivalence, 2 of the items were separately broken into two subitems to avoid potentially double-barreled answers. These 14 items queried participants about their level of risk concern associated with the following items: (1) pesticides; (2) additives in processed foods; (3) food preservatives; (4) livestock, poultry, and food produced with growth enhancers; (5) antibiotics in livestock production; (6) antibiotics in poultry production; (7) food irradiation (i.e., X-ray); (8) food handling and storage in restaurants; (9) food handling and storage in supermarkets; (10) alteration of the nutritional value of food during processing; (11) genetically altered foods; (12) addition of products to increase milk processing; (13) food handling in the home; and (14) refrigeration storage of food in the home. A 5-point Likert-type scale ranging from 1 (not concerned at all) to 5 (very concerned) was used to measure these items. The reliability coefficient across these 14 items was .85.

Emotions

Izard’s (1977) differential emotions theory was used as the basis for assessing five discrete negative emotions that were considered relevant to the present study context. These five discrete negative emotions—anger, sadness, fright (or fear), anxiety, and resentment (or contempt)—were assessed by 15 items adopted from Allen, Machleit, and Marine’s (1988) study; Allen et al.’s scale (αs = .72-.89) was developed based on Izard’s (1977) paradigm. In particular, respondents were asked to indicate the degree that “The food safety cases made me feel angry, irritated, ‘annoyed,’ ‘sad,’ ‘downhearted,’ ‘unhappy,’ ‘scared,’ ‘fearful,’ ‘afraid,’ ‘nervous,’ ‘anxious,’ ‘worried,’ ‘disgusted,’ ‘hatred,’ and ‘resentful.’ All items were measured on a 5-point Likert-type scale, ranging from 1 (never) to 5 (all the time). The reliability coefficients for the anger, sadness, fright, anxiety, and resentment dimensions were .78, .81, .91, .83, and .85, respectively. A negative emotion scale was constructed by combining these dimensions (α = .93).

Food Safety Risk Preventive Action

This variable depicts the degree to which participants were engaged in taking precautions against food safety risks. A set of 10 original items was developed to measure this variable on a 5-point Likert-type scale (ranging from not at all frequently to very frequently): (1) When I purchase food products, I avoid those products with a history of safety recalls; (2) when I purchase food products, I avoid those products that are being recalled; (3) I avoid eating in restaurants or public dining halls, if I am aware of any ongoing food-borne illnesses; (4) when I purchase food products, I check the safety of the products by examining the integrity of packaging, expiration date, appearance (and/or smell), and other safety indicators; (5) I reduce or avoid consuming those food products that are related to any ongoing food-borne illnesses; (6) I discard food products that I have purchased, if I am aware of any ongoing food safety recalls related to those products; (7) I try to follow safe food-handling practices when I prepare food for consumption; (8) when I consume food products, I avoid those products whose safety may be in question; (9) I try to follow safe food preservation practices (e.g., observing expiration dates, setting the correct refrigerator temperature, etc.) when I store food products in my home; and (10) I purchase food products only from those retailers that I trust. The reliability coefficient generated for this scale is .80.

Results

As shown in Table 1, the mean age of the sample was 28.71 years (SD = 6.49). The sample’s age breakdown showed that 49.7% of the respondents were between 18 and 27 years, 49.6% between 28 and 54, and another 0.7% between 55 and 67. The average monthly income was in the range of 6,000 to 10,000 RMB Yuan (or about US$951-US$1,585), reflecting a middle-class income level. Respondents’ education level revealed a highly educated group, with 74% reporting having received at least a 4-year college degree. Weibo use time per day was around 1.18 hours (SD = 1.84). The average Weibo use frequency was 4.31 times (SD = .91) or at least several times per week.

Table 1.

Means and Standard Deviations of All Variables.

	M	SD	n
Age, years	28.71	6.49	1,360
Income^a	6.22	2.11	1,360
Education^b	3.80	0.72	1,360
Internet use time, hours	6.13	3.99	1,356
Weibo use frequency	4.31	0.91	1,360
Online news use, hours	0.61	0.90	1,357
TV news use, hours	0.94	3.09	1,356
Newspaper use, hours	0.39	1.07	1,359
Magazine use, hours	0.40	0.73	1,360
Incident awareness^c	8.88	2.24	1,360
Factual awareness^d	3.82	0.71	1,360
Risk perception^d	4.05	0.53	1,360
Emotion^d	3.19	0.82	1,360
Preventive action^d	3.92	0.56	1,360

Scale = 1-10 (1 = 1,000 yuan or below; 10 = above 20,000 yuan).

Scale = 1-5 (1 = middle school or lower; 5 = graduate and above).

Scale = 0-12.

Scale = 1-5.

On an average day, respondents spent about (1) 6.13 hours (SD = 3.99) on the Internet, (2) less than 1 hour on Internet news (M = 0.61, SD = 0.90) or television news viewing (M = 0.94, SD = 3.09), and (3) less than ½ hour on newspaper (M = 0.39, SD = 1.07) or magazine (M = 0.40, SD = 0.73) reading. Respondents were aware of approximately 9 of the 12 major food safety incidents (M = 8.88, SD = 2.24). On a 5-point scale, the means for the level of factual awareness (M = 3.82, SD = 0.71), risk perception (M = 4.05, SD = 0.53), preventive action (M = 3.92, SD = 0.56), and negative emotions (M = 3.19, SD = 0.82) were moderate or moderately strong.

A zero-order correlation test was conducted to ascertain the interrelationships between demographic, media use variables, and Weibo use frequency as well as to screen their potential contributions in order to help explain psychometric measures included in the proposed conceptual model (see Table 2). The correlation results indicate that income (r = .15) and education (r = .15) were significant correlates of Weibo use frequency. Except for Internet use time (r = .12), none of the media use variables—including time spent with newspapers, magazines, TV news, and online news—were correlated with Weibo use frequency. Age was found to be positively and significantly associated with food safety risk perception (r = .17), incident awareness (r = .23), negative emotion (r = .15), and preventive actions (r = .18). Both income and education level were significantly but weakly associated with food safety risk perception (r = .20 and r = .10, respectively), incident awareness (r = .27 and r = .12, respectively), negative emotion (r = .16 and r = .06, respectively), and preventive actions (r = .19 and r = .09, respectively). Internet use time was marginally related to incident awareness (r = .05, p < .10), risk perception (r = .08, p < .01), and emotion (r = .07, p < .05) but unrelated to factual awareness or preventive action. Among the off-line media use variables, only newspaper use was found to be a significant correlate of risk perception (r = .07, p < .05); however, newspaper use was unrelated to incident awareness, factual awareness, negative emotion, or preventive actions. Based on these correlation results, only age, income, education, and Internet use time were included in model testing.

Table 2.

Zero-Order Correlation Coefficients Between Key Variables.

	1	2	3	4	5	6	7	8	9	10
1. Weibo frequency	—	—	—	—	—	—	—	—	—	—
2. Incident awareness	.20***	—	—	—	—	—	—	—	—	—
3. Risk perception	.09**	.34***	—	—	—	—	—	—	—	—
4. Factual awareness	.15***	.31***	.24***	—	—	—	—	—	—	—
5. Emotion	.07*	.19***	.39***	.20***	—	—	—	—	—	—
6. Preventive action	.13***	.24***	.34***	.31***	.37***	—	—	—	—	—
7. Income	.15***	.27***	.20***	.19***	.16***	.19***	—	—	—	—
8. Education	.15***	.12***	.10***	.19***	.06*	.09**	.20***	—	—	—
9. Age	−.02	.23***	.17***	.17***	.15***	.18***	.27***	.00	—	—
10. Internet use	.12***	.05^†	.08**	.03	.07*	.00	.01	−.02	−.06*	—

***

p < .001. **p < .01. *p < .05. ^†p < .10.

Correlations between the variables associated with the proposed theoretical links are also shown in Table 2. Specifically, incident awareness (r = .20), risk perception (r = .09), factual awareness (r = .15), and preventive action (r = .13) were all significantly but weakly correlated (p < .05) with Weibo use frequency. Incident awareness, risk perception, factual awareness, and preventive action were weakly to moderately correlated with each other, with coefficients ranging from .24 to .34 (p < .05). Emotion was weakly related to incident awareness (r = .19, p < .001) and factual awareness (r = .20, p < .001) but moderately associated with risk perception (r = .39, p < .001) and preventive actions (r = .37, p < .001).

These correlation results indicate that there was no multicollinearity concern. A confirmatory factor analysis tested the interrelationships between all the variables contained in the proposed model. The analysis yielded an acceptable measurement model fit for a large national sample, where CMIN (minimum of discrepancy function)/df (degrees of freedom) = 2.81, p < .001, comparative fit index = .94, root mean square error of approximation = .036.

Research Questions and Hypotheses

The research questions and hypotheses were tested by the proposed conceptual model via a path analysis. The original model did not fit the data well. Hence, the following modifications were made: (1) removing the paths from media (or Internet) use to factual awareness and incident awareness; (2) adding separate paths from age, income, and education to factual awareness and incident awareness; and (3) removing the path from Weibo use frequency to risk perception. The final model statistics did show a good fit for a very large national sample (CMIN/df = 2.74, p = .002, comparative fit index = .989, root mean square error of approximation = .035), and the model explained 23% of the total variance in food safety preventive action (see Figure 2). Results from testing the Research Questions and Hypotheses are reported below.

Figure 2.

Revised conceptual model.

Research Question 1 explores whether food safety incident awareness, factual awareness, and food safety risk perception are related to food safety preventive action. Based on the path model, food safety incident awareness (β = .06, p < .05), factual awareness (β = .19, p < .001), and risk perception (β = .18, p < .001) were significant and positive predictors of food safety preventive action. Research Question 2 explores whether Weibo use would be significantly related to food safety preventive action. The finding also shows a positive relationship between Weibo use frequency and preventive action (β = .07, p < .01).

Research Question 3 queries whether demographic characteristics, media use behavior and Weibo use frequency, food safety incident awareness, or factual awareness are positive related to each other. The path analysis suggests that Internet use significantly predicted Weibo use frequency (β = .12, p < .001). Age (β = −.06, p < .05), income (β = .16, p < .001), and education (β = .12, p < .001) were significant predictors of Weibo use frequency. Similarly, age (β = .17, p < .001), income (β = .18, p < .001), and education (β = .07, p < .01) also significantly predicted food safety incident awareness. Moreover, age (β = .09, p < .001), income (β = .07, p < .05), and education (β = .16, p < .001) were all significant predictors of factual awareness as well.

Hypotheses 1a and 1b presume that food safety incident awareness and factual awareness are related to negative emotions. The path analysis results demonstrate that both food safety incidence awareness (β = .12, p < .001) and factual awareness (β = .14, p < .001) were positive predictors of negative emotions, providing support for Hypotheses 1a and 1b. Hypotheses 1c and 1d assert that negative emotions are related to food safety risk perception and preventive action. The findings reveal that negative emotions significantly predicted food safety risk perception (β = .32, p < .001) and preventive action (β = .25, p < .001), validating Hypotheses 1c and 1d.

Hypotheses 2 to 4 test whether the relationships between Weibo use and (1) food safety incident awareness, (2) factual awareness, and (3) food safety risk perception are positive and significant. The path analysis results indicate that Weibo use frequency was a significant and positive predictor of food safety incident awareness (β = .18, p < .001) and factual awareness (β = .08, p < .01) but not risk perception. These findings supported Hypotheses 2 and 3 but not Hypothesis 4.

Discussion

This study aimed to examine the Chinese public’s use of microblogs and their cognitive, affective, and behavioral responses to the continuing food safety crises in China. According to the descriptive statistics reported in Table 1, as 74% of the Weibo users in the sample had at least a college degree, this finding was not significantly different from a report that indicated at least 70% of Weibo users were college educated (China Internet Watch, 2014). The same report also suggested that 98% of Weibo users were born post-1970 (i.e., ≤44 years old); this percentage is similar to the 97.4% of Weibo users between 18 and 44 years in the current sample. The self-reported incident awareness among Weibo users was relatively high, with respondents indicating being aware of nearly 9 out of the 12 food safety incidents in recent years. Findings also show a relatively strong level of risk perceptions, alongside a moderate level of food safety–related factual awareness and preventive actions. These findings are largely consistent with the results reflected in the recent public opinion polls (Chen & Xie, 2012; Pew Global Attitudes Project, 2012).

As for the influence of mediated-channel use, it is interesting to see that aside from the general Internet use indicator, time spent with TV news, newspapers, magazines, and online news was not correlated with food safety incident awareness or factual awareness. These findings contradicted the food safety risk research (e.g., De Boer et al., 2004; McCarthy et al., 2007) conducted before the social media era. Instead, they provided support for social media (e.g., Weibo) as an effective channel for food safety communication in China. Likewise, these findings are also consistent with the prediction made about how social media could become an additional venue for conducting risk communication (Currie, 2009; Ding & Zhang, 2010). More important, the Chinese citizens have utilized Weibo to engage in seeking and exchanging potentially controversial risk information and communication—by bypassing the traditional media outlets and the Internet websites—that are more closely censored by the Chinese government (Magistad, 2012a).

The path analysis findings confirmed that more Internet use time, younger age, higher income, and more formal education were all significant but weak predictors of Weibo use frequency. More frequent Weibo users who were younger, more affluent, and better educated were also more aware of the food safety incidents and perceived themselves as having greater factual awareness of food safety knowledge. These findings suggest that Weibo was a good outlet for disseminating food safety information to generate crisis awareness and factual awareness of risk, reaffirming the positive influence of mediated communication on the public’s cognitive learning of food safety information (De Boer et al., 2004; Fleming et al., 2006; McCarthy et al., 2007).

Weibo use frequency also had a significant albeit weak impact on food safety prevention action but not on food safety risk perception. A possible explanation for these results could be that having access to food safety–related information was not sufficient in generating a significant change in the public’s risk perception. However, learning about the prevention actions adopted by Weibo bloggers might have prompted other Weibo users to do the same—due to their increased incident awareness and factual awareness—without having significantly elevated their risk perceptions.

This scenario seems consistent with the findings that showed how a greater degree of incident awareness and factual awareness was also related to a greater level of risk perception and preventive action. These findings are congruent with the limited existing literature generated prior to the social media era, which suggests that the public’s knowledge about food safety issues can influence their evaluation of personal susceptibility to the risk and risk perceptions related to food safety concerns (De Boer et al., 2005; Frewer et al., 1994; Shaw, 2003). They also affirmed past poll results (Pew Global Attitudes Project, 2012), which show that the food safety incidents have heightened the Chinese public’s perception of food safety risks.

Weibo users who were more aware of food safety incidents and had greater factual awareness also reported stronger negative food safety–related emotions. These results echoed the general thesis of appraisal tendency theory, which suggests that an individual’s emotion is influenced by his or her response to cognitive stimuli (Lerner & Keltner, 2000). As negative emotions related to food safety incidents were a significant predictor of food safety risk perception, these findings also validated the general assumptions of appraisal tendency theory, which identify affect and emotion as correlates to environmental risk perception (Lerner & Keltner, 2001; Loewenstein et al., 2001). Likewise, as negative emotions were also a significant antecedent of participant adoption of preventive actions, these results further affirmed the theoretical assumptions of affect heuristic (Slovic, Finucane, Peters, & MacGregor, 2007)—as well as past work that found how emotions could draw an individual to focus on assessing the risk at hand—and influence their action/decision tendencies (Watson & Spence, 2007).

Combined, these results suggest that emotional response could be instrumental in shaping an individual’s food safety risk perception as well as motivating his or her adoption of preventive actions. As a case in point, since a majority of the Chinese public had low confidence in the safety of their food supply, over a quarter of those surveyed in a national study indicated that they would purchase more (expensive) imported foods instead (Chen & Xie, 2012). This national survey finding was broadly supported by the present study, which indicates that younger, more educated, and wealthier Weibo users were also more engaged in taking prevention actions.

As existing research has yet to integrate relevant cognitive, affective, and behavioral factors to explore an environmental risk phenomenon such as a food safety crisis, the findings reported here can provide the baseline for guiding additional empirical research and help inform future theory-building efforts. More important, these findings present the first set of empirical evidence for explaining the theoretical relations between food safety cognitions, affects, and behaviors; they also reflect the first empirical effort for studying those relations in the context of social media use. Furthermore, the current study confirms the contention from previous research in risk analysis, which emphasizes how individuals go through an emotional process in perceiving risks, along with a rational cognitive process (e.g., Nabi, 1999, 2002). The social significance of the association between emotions and food safety incidents is noteworthy. This is because the public feels powerless in protecting their own safety; they also believe that such incidents could have been avoided, circumvented, or quickly controlled, if adequate policies, safeguards, or human interventions have been put in place or implemented (Buckley, 2013).

It should be noted that Weibo represents a much more condensed form of communication in Chinese than its counterpart (Twitter) in English, as it usually takes fewer Chinese characters to convey the same communication content than Twitter does in English (Chan, Wu, Hao, Xi, & Jin, 2012). As Weibo permits the entry of only 140 Chinese characters, the use of this social network modality to share and discuss food safety issues involves dealing with topics that are rather complex, technical, and scientific in nature. It is likely that Weibo users who choose to utilize this alternative platform to gather the facts about food safety problems from other Weibo users also trust the information provided by other Weibo users more than any government, industry, or institutional sources. Trusting the food safety–related information shared via social media does not warrant the transmission of truthful, scientifically accurate, or objective information. Nonetheless, past research has validated this pattern of trust toward other social media users in conducting political discourse about controversial issues via microblogs in China (e.g., Mou, Atkin, & Fu, 2011).

Study Limitations

There are several limitations associated with the study results. The first limitation concerns the overrepresentation of Weibo users (97%) in the study sample. As 88.81% of the Internet users in China are Weibo users (Takamitsu, 2012), a study that focuses on comparing Weibo and non–Weibo users will need to stratify the sample size from each group. Second, rural residents were underrepresented in the sample (1.3%), since an estimated 28% of the Internet users resided in rural China (Kan, 2013). A more accurate representation of the urban and rural residents in the sample would have better reflected a cross-sectional study. Third, the present study was a one-shot study. Since food safety concerns are a persistent problem in China, a longitudinal study would provide information on the evolution and changes of public beliefs, risk perceptions, knowledge, and prevention practices related to different and evolving food safety crises. Fourth, the study relied on self-reported data based on a survey instrument that was not pilot tested. Food safety research tends to involve surveys instead of experiments to assess public reactions toward actual food safety–related issue, instead of manipulating the food safety risk level to gauge hypothetical outcomes. Last, a social network analysis that detects the communication patterns on Weibo—as well as the online discourse that influences Weibo user cognitions, affects, and behaviors related to food safety issues—would be highly valuable. This type of social network analysis could also further demonstrate the social relations between users and the pattern of information flow (Scott, 2007; Valente, Gallaher, & Mouttapa, 2004).

Conclusion

The present study is among the first to integrate food safety cognitions, affects, and behaviors in the context of risk communication via a social media platform, alongside key demographic variables. Study findings show that cognitive learning of food safety risk—an environmental risk over which the public has little control—can elicit negative emotions; these negative affects can influence an individual’s risk evaluation, perception, and preventive action. Specifically, results indicate that microblog or Weibo use in China was related to public awareness of a series of food safety incidents and factual awareness, in addition to preventive actions associated with food safety risks. Emotional response toward the food safety incidents was a stronger predictor of both food safety risk perception and prevention action, relative to food safety incident awareness and factual awareness. Among Weibo users, access to other online and off-line news and information outlets was largely irrelevant in informing and influencing the public response to the food safety concerns in question.

Even though the Chinese government has adopted additional scrutiny with microblogging activities (Buckley & Lee, 2011), some active bloggers continue to test the government’s censorship boundary (Magistad, 2012a, b; Richburg, 2012). As Weibo users in China have now reached more than 300 million, Weibo is bringing social change to the Chinese society in an unprecedented manner, especially with the public’s desire to express their opinions (Hewitt, 2012). In the present study context, Weibo might have performed this “watchdog” duty—a task usually performed by the mainstream media in a country such as the United States—and provided active bloggers a platform to voice critical views about the government’s handling of food safety crises.

As trust is an important dimension in effective utilization of social media platforms for disseminating risk information, future research could explore how misinformation and rumors work to enhance or undermine the objective of effective environmental risk communication through social media. Additional work could further validate the present study framework in other countries or with other public health crises, including the recent food-borne illnesses outbreaks involving Salmonella, Escherichia coli, and Listeria monocytogenes in the United States (Centers for Disease Control and Prevention, 2014).

It would be useful for future work to compare food safety incident awareness and food safety–related knowledge as well as food safety risk perception and practices between Internet versus non–Internet users as well as social media users and nonusers. As mentioned above, conducting an actual social network analysis by recruiting social media users who are “friends” with each other would be highly desirable. This type of social network analysis could be a great way to study the actual social network interaction patterns and how such patterns may help facilitate or change public perceptions, attitudes, and behaviors in relation to food safety or other environmental/health concerns.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Doctoral Dissertation Fellowship Award from the University of Connecticut (USA).

Author Biographies

Yi Mou is an assistant professor of the Faculty of Humanities and Arts at Macau University of Science and Technology, Macau. Her research interest includes health communication and new media adoption and effects. Examples of her research work include environmental health communication on social media and utilizing new media tools in achieving health objectives.

Carolyn A. Lin (carolyn.lin@uconn.edu) is a professor in the Department of Communication at University of Connecticut. She was the founder of the Communication Technology Division at the Association of Education in Journalism and Mass Communication and a recipient of a Distinguished Research Faculty award. Her research interests include the content, uses, and effects of health communication, social marketing, and e-health technology.

References

Allen

C. T.

Machleit

K. A.

Marine

S. S.

(1988). On assessing the emotionality of advertising via Izard’s differential emotions scale. Advances in Consumer Research, 15, 226-231.

Bradsher

(2008, September 12). China begins inquiry into tainted baby formula. The New York Times. Retrieved from http://www.nytimes.com/2008/09/13/world/asia/13milk.html

Bruhn

C. M.

Schutz

H. G.

(1999). Consumer food safety knowledge and practices. Journal of Food Safety, 19, 73-87.

Buck

Anderson

Chaudhuri

Ray

(2004). Emotion and reason in persuasion: Applying the ARI model and the CASC scale. Journal of Business Research, 57, 647-656. doi:10.1016/S0148-2963(02)00308-9

Buckley

(2013, May 3). Rat meat sold as lamb highlights fear in China. The New York Times. Retrieved from http://www.nytimes.com/2013/05/04/world/asia/rat-meat-sold-as-lamb-in-china-highlights-fears.html?_r=0

Buckley

Lee

(2011, October 26). China vows stricter controls on social media. Retrieved from http://in.reuters.com/article/2011/10/26/us-china-internet-idINTRE79P0OJ20111026

Centers for Disease Control and Prevention. (2014). List of selected multistate foodborne outbreak investigations. Retrieved from http://www.cdc.gov/foodsafety/outbreaks/multistate-outbreaks/outbreaks-list.html

Chan

Hao

Jin

(2012). Microblogging, online expression, and political efficacy among young Chinese citizens: The moderating role of information and entertainment needs in the use of weibo. Cyberpsychology, Behavior, and Social Networking, 15, 345-349. doi:10.1089/CYBER.2012.0109

Chen

Xie

(2012, August 20). How has Internet use, devices and Chinese netizens evolved in the first half of 2012? Understanding Chinese Consumer Perceptions towards food safety in China. China Pulse. Retrieved from http://www.ipsos.com/sites/ipsos.com/files/27-China%20Pulse-August%202012.pdf

10.

China Internet Watch. (2014, January 9). Sina Weibo user demographics analysis in 2013 (4 part series). Retrieved from http://www.chinainternetwatch.com/5568/what-weibo-can-tell-you-about-chinese-netizens-part-1/

11.

CNNIC. (2013). The statistical report of Chinese Internet development. Retrieved from http://www1.cnnic.cn/IDR/ReportDownloads/

12.

Coombs

W. T.

Holladay

S. J.

(2005). An exploratory study of stakeholder emotions: Affect and crises. Research on Emotions in Organizations, 18, 263-280.

13.

Currie

(2009). Public health leaders using social media to convey emergencies. Nation’s Health, 39(6), 1-30.

14.

Ding

Zhang

(2010). Social media and participatory risk communication during the H1N1 flu epidemic: A comparative study of the United States and China Media Research, 6(4), 80-91.

15.

De Boer

McCarthy

Brennan

Kelly

A. L.

Ritson

(2004). Differences in the perception of food risk issues and food risk messages between various demographic groups on the island of Ireland (Interim report). Cork, Ireland: Safefood, the Food Safety Promotion Board.

16.

De Boer

McCarthy

Brennan

Kelly

A. L.

Ritson

(2005). Public understanding of food risk issues and food risk messages on the island of Ireland: The view of food safety experts. Journal of Food Safety, 25, 241-261.

17.

Fan

Zhao

Chen

(2013). Anger is more influential than joy: Sentiment correlation in Weibo. Retrieved from http://arxiv.org/abs/1309.2402

18.

Fife-Schaw

Rowe

(1996). Public perception of everyday food hazards: A psychometric study. Risk Analysis, 16, 487-500.

19.

Fleming

Thorson

Zhang

(2006). Going beyond exposure to local news media: An information-processing examination of public perceptions of food safety. Journal of Health Communication, 11, 789-806. doi:10.1080/10810730600959705

20.

Foods1. (2010, March 17). About 200,000-400,000 food poison victims each year; consumers shouldn’t be trial target. Retrieved from http://spzx.foods1.com/show_909315.htm (Article in Chinese)

21.

Foster

(2011, April 27). Top 10 Chinese food scandals. The Telegraph. Retrieved from http://www.telegraph.co.uk/news/worldnews/asia/china/8476080/Top-10-Chinese-Food-Scandals.html

22.

Frewer

L. J.

Shepherd

Sparks

(1994). The interrelationship between perceived knowledge, control and risk associated with a range of food-related hazards targeted at the individual, other people and society. Journal of Food Safety, 14, 19-40.

23.

Frijda

N. H.

(1988). The laws of emotion. American Psychologist, 43, 349-58.

24.

Gan

(2011). Weibo user profile. Retrieved from http://www.weiboyx.com/weiboyx/20110908715.html

25.

Gaspar

Gorjao

Seibt

Lima

Barnett

Moss

Wills

(2014). Tweeting during food crises: A psychosocial analysis of threat coping expressions in Spain, during the 2011 European EHEC outbreak. International Journal of Human-Computer Studies, 72, 239-254.

26.

Gaziano

(1997). Forecast 2000: Widening knowledge gaps. Journalism & Mass Communication Quarterly, 74, 237-264.

27.

Hewitt

(2012, July 31). Weibo brings change to China. BBC World Service. Retrieved from http://www.bbc.co.uk/news/magazine-18887804

28.

iFeng. (2011). Big food safety incidents during 2008-2011. Retrieved from http://finance.ifeng.com/news/special/shipinanquan/20110505/3986786.shtml (Article in Chinese)

29.

Institute of Food Technology. (2014). Learn about food science. Retrieved from http://www.ift.org/knowledge-center/learn-about-food-science.aspx

30.

Izard

C. E.

(1977). Human emotions. New York, NY: Plenum Press.

31.

Jin

Pang

Cameron

G. T.

(2007). Integrated crisis mapping: Towards a publics-based, emotion-driven conceptualization in crisis communication. Sphera Publica, 70, 81-96.

32.

Johnson

A. E.

Donkin

A. J. M.

Morgan

Lilley

J. M.

Neale

R. J.

Page

R. M.

Silburn

(1998). Food safety knowledge and practice among elderly people living at home. Journal of Epidemiology & Community Health, 52, 745-8.

33.

Kan

(2013, July 17). China’s Internet users grow to 591 million. PC World. Retrieved from http://www.pcworld.com/article/2044521/chinas-internet-users-grow-to-591-million.html

34.

Lazarus

R. S.

(1991). Emotion and adaptation. New York, NY: Oxford University Press.

35.

Lazarus

R. S.

(1993). From psychological stress to the emotions: A history of changing outlooks. Annual Review of Psychology, 44, 1-21.

36.

LeDoux

(1996). The emotional brain. New York, NY: Simon & Schuster.

37.

Lerner

J. S.

Gonzalez

R. M.

Small

D. A.

Fischhoff

(2003). Effects of fear and anger on perceived risks of terrorism: A national field experiment. Psychological Science, 14, 144-150.

38.

Lerner

J. S.

Keltner

(2000). Beyond valence: Toward a model of emotion-specific influences on judgment and choice. Cognition & Emotion, 14, 473-93.

39.

Lerner

J. S.

Keltner

(2001). Fear, anger, and risk. Journal of Personality and Social Psychology, 81, 146-159.

40.

Loewenstein

G. F.

Weber

E. U.

Hsee

C. K.

Welch

(2001). Risk as feelings. Psychological Bulletin, 127, 267-286.

41.

Lovejoy

Waters

R. D.

Saxton

G. D.

(2012). Engaging stakeholders through Twitter: How nonprofit organizations are getting more out of 140 characters or less. Public Relations Review, 38, 313-318. doi:10.1016/j.pubrev.2012.01.005

42.

Luce

M. F.

(1998). Choosing to avoid: Coping with negatively emotion-laden consumer decisions. Journal of Consumer Research, 24, 409-433.

43.

MacLeod

(2012, May 28). Chinese despair at endless food-safety scares. USA Today. Retrieved from http://usatoday30.usatoday.com/news/world/story/2012-05-28/china-food-safety/55252482/1

44.

Magistad

M. K.

(2012a, November 13). Blogger Wu Heng, A champion of food safety in China. The PRI’s World. Retrieve from http://www.theworld.org/2012/11/food-safety-china/

45.

Magistad

M. K.

(2012b, August 9). How Weibo is changing China. Yale Global. Retrieved from http://yaleglobal.yale.edu/content/how-weibo-changing-china

46.

Mayer

A. B.

Harrison

J. A.

(2012). Safe Eats: An evaluation of the use of social media for food safety education. Journal of Food Protection, 8, 1453-1463.

47.

McCarthy

Brennan

Kelly

A. L.

Ritson

De Boer

Thompson

(2007). Who is at risk and what do they know? Segmenting a population on their food safety knowledge. Food Quality and Preference, 18, 205-217. doi:10.1016/j.foodqual.2005.10.002

48.

Mou

Atkin

(2011). Predicting political discussion in a censored online environment. Political Communication, 28, 341-356. doi:10.1080/10584609.2011.572466

49.

Nabi

R. L.

(1999). A cognitive-functional model for the effects of discrete negative emotions on information processing, attitude change, and recall. Communication Theory, 9, 292-320.

50.

Nabi

R. L.

(2002). Anger, fear, uncertainty, and attitudes: A test of the cognitive-functional model. Communication Monographs, 69, 204-216.

51.

Newkirk

R. W.

Bender

J. B.

Hedberg

C. W.

(2012). The potential capability of social media as a component of food safety and food terrorism surveillance systems. Foodborne Pathogens and Disease, 9, 120-124.

52.

Ormrod

(2008). Human learning (5th ed.). Upper Saddle River, NJ: Pearson Education.

53.

Pew Global Attitudes Project. (2012, October 16). Growing concerns in China about inequality, corruption. Retrieved from http://www.pewglobal.org/2012/10/16/growing-concerns-in-china-about-inequality-corruption/

54.

Richburg

K. B.

(2012, June 26). China’s bloggers are taking risks and pushing for change, one click at a time. The Washington Post. Retrieved from http://www.washingtonpost.com/world/asia_pacific/chinas-bloggers-are-taking-risks-and-pushing-for-change-one-click-at-a-ime/2012/06/26/gJQANgmZ3V_story.html

55.

Rimal

Fletcher

S. M.

McWatters

K. H.

Misra

S. K.

Deodhar

(2001). Perception of food safety and changes in food consumption habits: a consumer analysis. International Journal of Consumer Studies, 25(1), 43-52.

56.

Rutsaert

Regan

Pieniak

McConnon

Moss

Wall

Verbeke

(2013). The use of social media in food risk and benefit communication. Trends in Food Sciences & Technology, 30(1), 84-91.

57.

Scott

(2007). Social network analysis: A handbook. Thousand Oaks, CA: Sage.

58.

Shaw

(2003). Public understanding of food risks: Expert and lay views. Retrieved from http://library.wur.nl/WebQuery/clc/1675485

59.

Slovic

Finucane

M. L.

Peters

MacGregor

D. G.

(2007). The affect heuristic. European Journal of Operational Research, 177, 1333-1352.

60.

Slovic

Peters

(2006). Risk perception and affect. Current Directions in Psychological Science, 15, 322-325.

61.

Tai

Sun

(2007). Media dependencies in a changing media environment: the case of the 2003 SARS epidemic in China. New Media & Society, 9, 987-1009.

62.

Takamitsu

(2012, July 27). 88.81% of over 19 years old internet user are using Weibo in China. Retrieve from http://socialmediainasia.blogspot.hk/2012/07/8881-of-over-19-years-old-internet-user.html

63.

Tuner

M. M.

(2006). Using emotion in risk communication: The anger activism model. Public Relations Review, 33, 114-119. doi:10.10016/j.pubrev.2006.11.013

64.

Unusan

(2007). Consumer food safety knowledge and practices in the home in Turkey. Food Control, 18, 45-51.

65.

Valente

T. W.

Gallaher

Mouttapa

(2004). Using social networks to understand and prevent substance use: A transdisciplinary perspective. Substance Use & Misuse, 39(10), 1685-1712.

66.

Watson

Spence

M. T.

(2007). Causes and consequences of emotions on consumer behaviour: A review and integrative cognitive appraisal theory. European Journal of Marketing, 41, 487-511.