Discussion

U i

K ∗

,

,

D

a

A R R A A

K F I R R F

1

a g A a t p 2 f n t

( e l r R

( (

h 0

International Journal of Hospitality Management 60 (2017) 67–76

Contents lists available at ScienceDirect

International Journal of Hospitality Management

jo u r n al homep age: www.elsev ier .com/ locate / i jhosman

nderstanding responses to posted restaurant food safety scores: An nformation processing and regulatory focus perspective

imberly J. Harris (Ed. D.) (Professor, The Bessie Morgan Marshall Professor of Hospitality Management) Lydia Hanks (Ph. D.) (Assistant Professor), Nathaniel D. Line (Ph. D.) (Assistant Professor) Sean McGinley (Ph. D.) (Assistant Professor) edman School of Hospitality Management, Florida State University, 288 Champions Way, UCB 4112, P. O. Box 30306541, Tallahassee, FL 32306-2541, USA

r t i c l e i n f o

rticle history: eceived 17 May 2016 eceived in revised form 11 August 2016 ccepted 11 September 2016 vailable online 18 October 2016

a b s t r a c t

Ensuring the safety of food served in restaurants continues to be an essential issue in the hospitality industry. An important part of the efforts to stem the outbreak of foodborne illnesses are the mandatory inspections of any entity that serves food to the public. Unfortunately, while posting food safety scores is intended to help consumers make better dining choices, interpreting these scores can often be difficult and confusing. The purpose of this study is to use information processing theory as a framework to

eywords: ood safety nformation processing theory egulatory focus estaurants ood safety inspection scores

investigate how consumers evaluate food safety inspection scores. To achieve this goal, this research provides an account of the effect of food safety concern on consumers’ attitudes toward restaurants under conditions of both positive and negative health inspection results. The results identify a moderating effect of health score in the formation of consumers’ attitudes toward restaurants. The downstream effects on expected satisfaction and behaviors are also established.

Published by Elsevier Ltd.

. Introduction

Incidents of foodborne illness outbreaks in the United States re climbing despite the increase in food safety management pro- rams and the implementation of the Food Safety Modernization ct in 2011 (FDA, 2011). The recent outbreaks of foodborne illness t Chipotle restaurants and on commercial cruise lines have served o highlight the fact that food safety is an ongoing issue in the hos- itality industry (CDC, 2016; Food Safety News, 2016). In fact, in 013 (the latest year for which data are available), the U.S. Centers or Disease Control (CDC) reported 13,360 cases of foodborne ill- ess, a 38% increase from the year before, and projections suggest hat this number will continue to grow (CDC, 2013).

As outlined in the Food and Drug Administration’s Food Code 2013), an integral part of preventing such outbreaks is the food stablishment inspection process wherein health inspectors regu-

arly inspect restaurant food handling practices and subsequently ate these establishments on adherence to food safety guidelines. esults of these inspections are public information, and many states

∗ Corresponding author. E-mail addresses: [email protected] (K.J. Harris), [email protected]

L. Hanks), [email protected] (N.D. Line), [email protected] S. McGinley).

ttp://dx.doi.org/10.1016/j.ijhm.2016.09.002 278-4319/Published by Elsevier Ltd.

require the public posting of inspection results in the form of an alphabetic letter, usually an A, B, or C (McVicar, 2011). Accordingly, the issue of food safety is an important one for restaurants, as the negative publicity associated with a poor health rating can result in loss of consumer trust, public relations problems, and legal costs (DiPietro et al., 2011).

Food safety is also of importance to a restaurant’s customer base (Fatimah et al., 2011) as consumers often look to posted food safety scores to make dining decisions (Henson et al., 2006). How- ever, while consumers can benefit from the disclosure of restaurant inspection results, interpreting posted food safety scores can often be difficult and confusing (Filion and Powell, 2011). To better under- stand this phenomenon, previous studies have investigated the delivery of food safety rating information, such as format, location, and information source (e.g., Choi et al., 2013; Filion and Powell, 2001). While such research has advanced the knowledge of con- sumer reactions to restaurant food safety in a general way, very few studies have explored (1) the cognitive traits consumers use to interpret food safety scores and (2) the ways in which food safety information is used to make assessments about the establishment. Accordingly, very little is known about how individual personal-

ity traits affect consumer’s perceptions of food safety inspection results and the interpretation of posted scores.

The purpose of this study is to bridge this gap in the literature by using information processing theory as a framework to inves-

6 of Hos

t i v t a ( t t h f s e r e i

2

t C h h t k o a i w A o 2 a

a i o t o m t c I i e

b h h a ( m i c i t s

2

C i c e

8 K.J. Harris et al. / International Journal

igate the how consumers evaluate posted restaurant food safety nspection scores. Understanding the extent to which individually arying traits affect food safety perceptions is important because he evaluation of food safety information can affect beliefs about

restaurant’s food, service expectations, and consumer behavior Henson et al., 2006). To achieve this goal, this research inves- igates the effect of food safety concern on consumers’ attitudes oward restaurants under conditions of both positive and negative ealth inspection results. The hypotheses inherent to the proposed

ramework are tested using data from a sample of restaurant con- umers in the United States. Upon establishing the moderating ffect of health score in determining consumers’ attitudes toward estaurants, the effects on expected satisfaction and behaviors are xamined. The implications for restaurant operators and those nvolved in developing food service policy are discussed.

. Literature review

Food safety inspection programs in the U.S. vary according o state, regional and city food safety management programs. odified by the Food and Drug Administration (FDA), authorized ealth agencies are approved to execute inspection programs and ave flexibility in evaluation methods and reporting nomencla- ure. Inspections are conducted by trained sanitarians, otherwise nown as health inspectors, and methods of inspection are devel- ped by the local agency. Adopted methods and evaluation criteria re used to guide the evaluation as sanitarians visit restaurants for nspections, which are usually conducted twice per year; however,

ith the recent implementation of the Food Safety Modernization ct, those entities qualifying as high risk can be inspected more ften (FDA, 2011, 2016). Inspections can be based on the 2009 or 013 adopted FDA Food Code or on the guidelines established by uthorized health agencies (FDA, 2009, 2013).

Regardless of the food safety inspection protocol, inspections re conducted and results are reported and summarized. After the nspection has been conducted, the food establishment is informed f any violations and the criticality of identified issues. Based on he level of violation(s) received, restaurant operators are notified f specific areas of concern and the actions that should be imple- ented to correct the noted violations. If a violation is considered

o pose a significant safety risk to the consumer, the establishment an be closed for a number of days until the situation is corrected. f correction is not accomplished in the timeline designated by the nspector, the establishment can be permanently closed (Waters t al., 2013).

The practice of posting food safety scores is important for a num- er of reasons. Posted inspection notices impact consumer choice, eighten the motivation of food service establishments to receive igh scores on food safety inspections, and serve as a marketing dvantage, whether overtly or covertly, for food establishments Choi et al., 2013; Tarca and Murphy, 2014). Unfortunately, the

ethods of communicating such information to the public are nconsistent in the U.S., and the results are often confusing for ustomers (Baker-White, 2014; Fatimah et al., 2011). As follows, nformation processing theory and regulatory focus theory are used o provide a framework for consumer interpretation of posted food afety inspection reports (see Fig. 1; H = Hypothesis).

.1. Information processing theory

Information processing theory (Chaiken, 1987; Petty and

acioppo, 1986) suggests that individuals process new pieces of

nformation either heuristically or systematically. Heuristic pro- essing is based on shortcuts, clues, proxies, or stereotypes to valuate a situation and can be advantageous as it is more efficient

pitality Management 60 (2017) 67–76

and requires less cognitive energy than deeper processing (Chaiken, 1987; Petty and Cacioppo, 1986). The disadvantage of using heuris- tics is that this process is less detailed and can result in less accurate conclusions, resulting in potentially incomplete gathering of infor- mation.

Conversely, systematic processing is a deeper and more cogni- tive processing of information (Chaiken, 1987; Petty and Cacioppo, 1986). Systematic processing entails gathering a plethora of infor- mation, analyzing it, and using it to arrive at a conclusion. This process is much more time-intensive, effortful, and cognitively demanding than heuristic processing, but the result often leads to a more accurate conclusion or evaluation.

The information processing model suggests that when deter- mining whether an individual will process a new piece of information heuristically or systematically, people tend to be inher- ently predisposed to the conservation of cognitive resources (Taylor and Fiske, 1978). This conservative approach leads to heuristic pro- cessing whenever possible, unless the individuals are otherwise motivated to expend the extra time, energy, and cognitive effort on systematic processing. Previous studies have found that moti- vations to engage in systematic information processing can include importance, relevance, outcome dependency, mood, need for cog- nition, and desire for control (Neuberg and Fiske, 1987; Petty et al., 1981; Pittman and D’Agostino, 1989).

Positive cues result in heuristic processing, while negative cues often lead to systematic processing. For example, a good mood or positive environment can result in evaluations that are creative, simplified, and characterized by less attention to detail (Mackie and Worth, 1989; Schwarz et al., 1991). Positive cues send a message that “all is as it should be,” and there is no need to further evaluate the situation; thus, the heuristic-based conclusion will be adequate. Contrariwise, a bad mood, a negative environment, or an adverse occurrence often leads to more systematic processing (Bohner et al., 1994; Schwarz, 1990). Some researchers suggest that such negative cues signal threats or danger in the environment, prompting a per- son to look carefully, gather detailed information, and analyze it critically to solve problems (Frijda, 1988; Schwarz, 1990; Schwarz et al., 1991). Negative cues send a message that “something is not right,” or “possibly dangerous,” and in such cases, systematic pro- cessing is a way of minimizing the possibility that one might arrive at a wrong, and possibly dangerous, conclusion.

The present research uses information processing theory as framework to understand consumers’ reactions to posted health inspection scores in a restaurant environment. However, equally important to the proposed framework are the constructs of pro- motion and prevention focus. According to regulatory focus theory, these constructs are proposed to affect consumer attitudes and behaviors when positive or negative visual cues are presented. Thus, the premise of the present research is that individually vary- ing traits affect consumers’ perceptions of food safety inspection scores and subsequently determine whether they use heuristic or systematic processing to form attitudes about the restaurant. These specific traits are discussed as a part of the following account of regulatory focus theory.

2.2. Regulatory focus theory

The concept of regulatory focus suggests that people are moti- vated by two different kinds of goals: promotion goals and prevention goals (Higgins, 1998; Lockwood et al., 2002). Promotion goals are oriented toward maximizing positive outcomes and pre-

vention goals are oriented toward minimizing negative outcomes. Regulatory focus is an inherent, individual-level trait, but can also be primed by the situation, as in the presentation of a health inspec- tion score.

K.J. Harris et al. / International Journal of Hospitality Management 60 (2017) 67–76 69

H2

H1

H3

H7 (+)

Attitude toward Restaurant

Word of Mouth

Expected Satisfaction

H5

H6 (+)

H8 (+)

Health Score

H4

Interpersonal Influence

Food Safety Concern

Prevention Focus

Promotion Focus

Self Efficacy

Intent to Patronize

ptual

p b l f t i t r a d u a t h

H a h w s

e s f o “ t f r r t t r

H a h w n

H1-H5: (a, b)

Fig. 1. Conce

The promotion-focused individual is aspirational and seeks leasure. These individuals are focused on being the best they can e, achieving the most success possible, and gaining the highest

evel of rewards. If a promotion-focused individual is dining with riends or co-workers at a restaurant, this person is looking forward o tasty food, fun with friends, and a pleasant experience. Accord- ng to regulatory focus theory, such individuals are predisposed o process information regarding food safety heuristically, as their egulatory focus causes them to perceive the landscape as gener- lly safe, negating the need to gather information and process too eeply. Thus, it is expected that when a promotion-focused individ- al encounters a good health safety score, they will have a positive ttitude toward the restaurant. Conversely, if the rating is poor, hey will have a negative attitude toward the restaurant. Thus, it is ypothesized that:

ypothesis 1. A restaurant’s posted health score will moder- te the relationship between an individual’s promotion focus and is/her attitude toward the restaurant such that the relationship ill be (a) significantly positive when the score is good and (b)

ignificantly negative when the score is poor.

An individual with a prevention-based regulatory focus is ori- nted toward minimizing negative outcomes. Issues such as safety, ecurity, and responsibility are of concern to the prevention- ocused individual (Chitturi et al., 2008). When faced with a choice r decision, prevention-focused individuals are likely to think, How can I minimize loss or negative outcomes?” It is expected hat a person highly concerned with prevention will be heavily ocused on preventing foodborne illness, even when the restau- ant has received a positive inspection grade. However, when the estaurant receives a negative inspection score, it is expected that he prevention-focused consumer would be extremely sensitive to his threat and would thus exhibit a negative attitude toward the estaurant. Thus, it is hypothesized that:

ypothesis 2. A restaurant’s posted health score will moder-

te the relationship between an individual’s prevention focus and is/her attitude toward the restaurant such that the relationship ill be (a) insignificant when the score is good and (b) significantly egative when the score is poor.

framework.

Furthermore, a prevention-focused consumer, according to information processing theory, is concerned with minimizing or avoiding negative outcomes, such as foodborne illness. This theory suggests that a poor health inspection score acts as a negative cue, signaling that danger or threat is present, leading to deeper, more systematic processing. When systematic processing is engaged, the person seeks additional detailed information to make an evaluation about the restaurant and uses it to decide whether or not to eat there. In this study, three additional pieces of information that the consumer may assess are examined: susceptibility to interpersonal influence, level of food safety concern, and sense of self-efficacy.

2.3. Interpersonal influence

If a customer is seeking additional information in order to make an evaluation about restaurant choice, consideration of others din- ing with the individual might influence the decision. This may or may not influence their own decision, subject to their level of susceptibility to interpersonal influence. Susceptibility to inter- personal influence is a general personality trait that exists across conditions (McGuire, 1968) and has been described as a confor- mation to expectation (Bearden et al., 1989). People may conform to influence from others in order to gain acceptance, raise status, gain rewards from others, or avoid punishments (Burnkrant and Cousineau, 1975; Park and Lessig, 1977).

The consumer behavior literature has examined the effect of this trait on consumer decision-making when others are present in the environment (Ford and Ellis, 1980; Ratner and Khan, 2002) and found that interpersonal influence plays a major role in the forma- tion of consumer attitudes and behaviors (Chaplin and John, 2010; Mourali et al., 2005). Thus, when a customer is deciding whether or not to eat at a restaurant with which they are not entirely com- fortable, they may seek the opinion of others in the dining party. It is expected that the importance of this information to their deci- sion will depend on the individual’s susceptibility to interpersonal influence, such that:

Hypothesis 3. A restaurant’s posted health score will moderate the relationship between an individual’s interpersonal influence and his/her attitude toward the restaurant such that the rela-

7 of Hos

t s

2

h c r 2 p c r p ( o f c s

H t h w n

2

( i t c i 1 b o m a s p a s

H t a i w

2

i 1 u c s m

b t p o t t b

0 K.J. Harris et al. / International Journal

ionship will be (a) insignificant when the score is good and (b) ignificantly positive when the score is poor.

.4. Food safety concern

Food safety concern is a measure of the concern an individual as regarding food safety issues including food to be consumed, leanliness, foodborne illness possibilities, and the overall restau- ant environment. Due to the proliferation of the Internet and the 4-h news cycle, outbreaks of foodborne illness and violations of roper food safety protocols are now widely publicized. This has reated a sense of concern on the part of some consumers as to the isks of eating at restaurants where the origins of the food and the rocedures by which the food is stored and handled are unknown Ortega et al., 2011; Yeung and Yee, 2012). However, even in the face f such publicity, some consumers do not have a great concern for ood safety issues. When a customer is engaging in systematic pro- essing due to a prevention focus, their level of food safety concern hould be relevant such that:

ypothesis 4. A restaurant’s posted health score will moderate he relationship between an individual’s food safety concern and is/her attitude toward the restaurant such that the relationship ill be (a) insignificant when the score is good and (b) significantly egative when the score is poor.

.5. Perceived self-efficacy

The construct of self-efficacy is an outgrowth of Bandura’s 2001) social-cognitive theory and encompasses the idea that an ndividual can determine or affect outcomes in his/her own life or heir environment (Scholz et al., 2002). Self-efficacy can be con- eptualized as an individual-level generalized personality trait that s stable across domains (Scholz et al., 2002; Sherer and Maddux, 982). General self-efficacy indicates that the individual holds a road set of optimistic self-beliefs about an ability to affect change r impact personal life outcomes. General self-efficacy is a deter- inant and motivator of attitudes and actions (Bandura, 2001) and

person’s level of self-efficacy determines how they perceive a ituation and respond to it. As it relates to foodborne illness, a erson’s sense of self-efficacy speaks to their belief that they can void contracting foodborne illness by avoiding restaurants with ubstandard food handling practices. As such, it is expected that:

ypothesis 5. A restaurant’s posted health score will moderate he relationship between an individual’s self-efficacy and his/her ttitude toward the restaurant such that the relationship will be (a) nsignificant when the score is good and (b) significantly positive

hen the score is poor.

.6. Attitude and behavior

Attitude can be defined as a function of one’s beliefs about an ssue in combination with an evaluation of that belief (Fishbein, 963). Attitude encompasses the degree of an individual’s eval- ation of an object or situation, either positive or negative; onsequently, the resulting evaluation impacts and predicts con- umer behaviors that may include loyalty, purchase, and word of outh (Ajzen and Fishbein, 1980). A consumer’s attitude about an organization can be impacted

y any number of factors. This study is concerned specifically with he extent to which regulatory focus, perceived self-efficacy, inter- ersonal influence, food safety concern, and the inspection rating

f the restaurant influence restaurant consumers’ attitudes toward he restaurant and the subsequent behavioral effects of these atti- udes. Based on the established relationships between attitudes and ehavior, it is predicted that:

pitality Management 60 (2017) 67–76

Hypothesis 6. An individual’s attitude toward a restaurant will positively affect the level of satisfaction he/she expects to experi- ence as a result of patronizing that restaurant.

Hypothesis 7. An individual’s expected satisfaction will affect his/her intentions to patronize the restaurant.

Hypothesis 8. An individual’s expected satisfaction will affect his/her intentions to spread positive word of mouth about the restaurant.

3. Methodology

3.1. Measured variables

To test hypotheses developed in the preceding section, the con- structs in the proposed model were operationalized and formatted into an electronic questionnaire. The variables were largely oper- ationalized using the seminal conceptualizations of pre-existing scales. The two regulatory focus constructs (�prevention = 0.85 and �promotion = 0.86) were operationalized according to the approach developed by Lockwood et al. (2002). Likewise, the susceptibility to interpersonal influence construct (� = 0.90) was operationalized using the normative items created by Bearden et al. (1989), and the self-efficacy construct (� = 0.92) came from Sherer and Maddux (1982).

Regarding the dependent variables, two different scaling tech- niques were used. Respondents’ attitude toward the restaurant was measured on three 7-point semantic differential scales (� = 0.99). Expected satisfaction was also measured using a semantic differ- ential approach (� = 0.99).

The behavioral constructs were measured on 7-point likelihood scales (highly unlikely/highly likely). Word of mouth intention (� = 0.99) and patronage intention (� = 0.98) were both opera- tionalized using scales previously employed by Zeithaml et al. (1996). Analyses of skewness (range: −1.16:1.20) and kurtosis (range: −1.79:2.27) across all the measurement items indicated that the data were distributed appropriately for the subsequent use of structural equation modeling (Curran et al., 1996).

Throughout the data collection process, a number of steps were taken to minimize the potential effects of common method biases (Podsakoff et al., 2003). For example, multiple scaling techniques were used to eliminate item characteristic effects and attention checks were used to eliminate fatigue effects. In order to confirm that the bias reduction methods were successful, two tests were conducted to assess the amount of shared variance potentially attributable to the measurement process. First, Harmon’s single factor test demonstrated that less than 50% of the model variance (31.5%) was explained by all measurement items modeled as a single factor. Second, the common latent factor method ( ̌ = 0.51) indicated that no more than 26.01% of the variance in the mod- els was attributable to the methods. Together, these tests indicate that common method biases were likely not a significant source of variance in the measurement model (Podsakoff et al., 2003).

3.2. Manipulations

Because the purpose of this research was to identify the effect that a restaurant’s health score exerts on consumers’ attitudes toward the restaurant, it was necessary to create a manipulation of the health score of the referent restaurant. The health score vari- able was manipulated by presenting participants with a scenario about a hypothetical restaurant selection experience. In the sce-

nario, respondents were told to imagine a situation regarding the decision to eat lunch at a restaurant. The scenario read as follows: “You are on your lunch break from work, and you decide to try a new place to eat for lunch. The place you have chosen is located

of Hos

i r s n r i

a w i s n r s L a d s ( h

o m r t p r o i ( c r a s s m ( c ( c

3

f z c p a S b d

c w d w d c d G m m t o

K.J. Harris et al. / International Journal

n a strip mall shopping area near a collection of shops, stores, and estaurants. You have shopped in the area and you have eaten at ome of the other establishments in the strip mall, but you have ever been to this one before. You park and walk up to the restau- ant. As you get to the front door, you see the following sign posted n the window:”

Following the scenario, respondents were shown an image of formal health score posting document. Half of the respondents ere shown an image of an ‘A’ rating for meeting the overall san-

tary requirements while the other half of the respondents were hown the image of a ‘C’ rating. In order to ensure that the sce- ario and the manipulated health score did not bias participants’ esponses to the individual trait variables, the scenario was pre- ented following the measurement of the independent variables. ikewise, to ensure that sample demographics would not confound ny potential moderation effects, a series of ANOVAs were con- ucted. The results of these analyses suggest that there were no ignificant differences among the measured demographic variables gender, ethnicity, education, age, income) across the high and low ealth score conditions.

Following the measurement of the behavioral variables, a series f manipulation checks were conducted to ensure a successful anipulation of the health score variable in terms of content and

ealism. Regarding the scores, both a subjective and an objec- ive manipulation check were included. In the objective check, articipants were asked to respond to the statement: “This restau- ant complies with state mandated sanitary regulations.” Results f an ANOVA (F = 3,365.7, p < 0.001) indicated that respondents n the A score condition had significantly higher mean responses MAScore = 6.22) to this statement than respondents in the C score ondition (MBScore = 2.40). In the subjective manipulation check, espondents were asked to indicate the extent to which they greed with the statement: “This restaurant got a good score on anitation.” Again, respondents in the A score condition reported ignificantly higher mean responses (MAScore = 6.27) to this state- ent than respondents in the C score condition (MBScore = 2.13)

F = 3,786.2, p < 0.001). Finally, responses to realism checks indi- ated that sll respondents found the scenarios to be realistic M = 5.41). The results of the manipulation checks indicated a suc- essful manipulation of the health score variable.

.3. Data collection

The data used to test the proposed framework were collected rom a sample of restaurant consumers in the United States. Ama- on Mechanical Turk was used as the sampling frame for data ollection. The use of Mechanical Turk as a sampling frame was urposeful, as users of this service have been demonstrated as rel- tively representative of consumer demographics in the United tates (Mason and Suri, 2011). Additionally, Mechanical Turk has een found to be an effective sampling tool when demographic iversity is desirable (Buhrmester et al., 2011).

As an incentive to participate, respondents were given a $.50 redit to their Mechanical Turk account. A total of 1547 responses ere recorded in this manner. To eliminate potentially spurious ata resulting from poor attention, three attention check items ere embedded within the survey (e.g., “please choose somewhat isagree. . .”). A total of 203 responses failed at least one of these hecks and were subsequently omitted from further analysis. After eleting these data, a total of 1324 usable responses remained. iven the number of estimated parameters in the measurement

odel (n = 113), the non-nested structural model (n = 96), and the oderated structural model (n = 192), the final sample size meets

he required threshold of free parameters to observations as rec- mmended by Kline (2005).

pitality Management 60 (2017) 67–76 71

Regarding the characteristics of the sample, respondents were relatively evenly distributed across several important demographic variables. In terms of age, 18.6% of the respondents were between 18 and 25 years old, while another 38.9% were between 26 and 35 years old. A further 18.6% were between 36 and 45 years old with the remainder of the sample (23.9%) reporting an age of 45 or older. Additionally, while a majority of the respondents were Caucasian (n = 1044, 78.9%), minority groups were rela- tively well represented (e.g., African American (n = 88); Hispanic (n = 66); and Asian (n = 90)). Regarding gender, there were slightly more female respondents (n = 737, 55.7%) than male respondents (n = 582, 44.0%).

Concerning respondents’ education, 40.6% reported having a high school degree or attending at least some college. Another 59.5% reported the completion of either an undergraduate or graduate degree. Finally, the respondents were also relatively evenly dis- tributed in terms of their reported household income. Forty one percent of the sample reported a household income of $40,000 or less. Another 36.4% reported a household income of between $40,000 and $80,000. The remaining respondents (22.1%) reported an annual household income of more than $80,000.

Finally, regarding dining behavior, respondents showed a rela- tively wide range of responses in answer to the question “How often do you eat outside of your home?” This information was captured on a per-week scaling basis with 29.2% of respondents reporting that they eat out less than once per week. The largest category of respondents (57.9%) indicated they eat out at least one to three times per week, while a further 10.8% indicated eating meals out four to eight times per week. These results suggest that the sam- ple had sufficient experience with the phenomenon in question to accurately respond to the survey content.

4. Results

4.1. Measurement model

To assess the validity of the constructs in the hypothesized framework, a measurement model was specified using struc- tural AMOS equation modeling software. The resultant fit indices indicated that the measurement model was a good fit to the data (�2 = 1,752.97, df = 553; RMSEA = 0.04; CFI = 0.98; NFI = 0.97; TLI = 0.98). Table 1 provides a full psychometric account of each construct. Means and composite reliabilities are reported along with the standardized loadings for each reflective indicator and the associated error variances.

The average variance extracted (AVE) of each latent variable was calculated to assess the validity of each of the constructs in the measurement model (Fornell and Larcker, 1981). The AVE is presented in Table 2 along with the critical ratio (CR), maximum shared squared variance (MSV), and average shared squared vari- ance (ASV) for each construct. As seen in the table, the critical ratio is greater than the average variance extracted for all constructs in the model. Additionally, all instances of AVE are greater than 0.50. These results provide support for the convergent validity of the constructs (Hair et al., 2006).

To establish discriminant validity, the square root of the AVE was calculated for each operational construct. This value was then compared to the correlations between the construct under consid- eration and all other constructs in the measurement model (Fornell and Larcker, 1981). Table 2 shows the correlations among all con- struct pairs in the measurement model. Additionally, the square

root of the AVE is presented on the diagonal. As seen in the table, the shared variance between each construct pair is less than the AVE, providing strong evidence of discriminant validity among the constructs (Hair et al., 2006).

72 K.J. Harris et al. / International Journal of Hospitality Management 60 (2017) 67–76

Table 1 Scale properties.

Constructs and Indicators Mean: A Mean: C Std. Est. Std. Err.

Promotion Focus: Likert (� = 0.85) 5.37a 5.37a 0.59b

I typically focus on the success I hope to achieve in the future.

5.50 5.44 0.83* 0.029

I often think about how I will achieve success. 5.28 5.31 0.83* 0.030 My major goal right now is to achieve my ambitions. 5.22 5.25 0.73* 0.033 I often imagine myself experiencing good things that I hope will happen to me.

5.47 5.48 0.68* 0.032

Prevention Focus: Likert (� = 0.86) 3.95a 4.16a 0.55b

I am anxious that I will fall short of my responsibilities and obligations.

4.12 4.36 0.70* 0.044

I often think about the person I am afraid I might become in the future.

3.48 3.73 0.75* 0.045

I often worry that I will fail to accomplish my goals. 4.06 4.29 0.81* 0.043 I often imagine myself experiencing bad things that I fear might happen to me.

3.60 3.83 0.81* 0.044

I frequently think about how I can prevent failures in my life.

4.51 4.60 0.63* 0.041

Interpersonal Influence: Likert (� = 0.92) 2.44a 2.52a 0.69b

I rarely purchase the newest products until I am sure my friends will approve of them.

2.50 2.56 0.64* 0.038

It is important that others like the products and brands I buy.

2.35 2.36 0.85* 0.033

When buying products, I generally purchase those brands that I think others will approve of.

2.33 2.44 0.95* 0.031

If other people can see me using a product, I often purchase the brand they expect me to buy.

2.24 2.36 0.92* 0.031

I like to know what brands and products make good impressions on others.

2.78 2.88 0.76* 0.040

Food Safety Concern: Likert (� = 0.95) 4.28a 4.52a 0.84b

I am concerned that the food served to me in restaurants is not safe.

4.16 4.39 0.91* 0.034

I am concerned that the food served to me in restaurants has a good chance of making me sick.

4.02 4.26 0.96* 0.035

I am concerned that the food served to me in restaurants might be contaminated.

4.18 4.45 0.94* 0.035

I am concerned that the food served to me in restaurants might not be handled safely by employees.

4.75 4.97 0.86* 0.035

Self-Efficacy: Likert (� = 0.90) 5.32a 5.20a 0.59b

It is easy for me to stick to my aims and accomplish my goals.

5.08 4.94 0.66* 0.033

Thanks to my resourcefulness, I know how to handle unforeseen situations.

5.30 5.22 0.83* 0.029

I can solve most problems if I invest the necessary effort.

5.62 5.50 0.75* 0.026

I can remain calm when facing difficulties because I can rely on my coping abilities.

5.26 5.13 0.80* 0.031

When I am confronted with a problem, I can usually find several solutions.

5.21 5.12 0.74* 0.027

I can usually handle whatever comes my way. 5.42 5.30 0.83* 0.027

Attitude: Semantic Differential (� = 0.99) 6.49a 1.76a 0.99b

Negative: Positive 6.49 1.80 0.99* 0.050 Bad: Good 6.49 1.79 0.99* 0.050 Unfavorable: Favorable 6.50 1.69 0.99* 0.051

Expected Satisfaction: Semantic Differential (� = 0.99) 6.33a 1.93a 0.96b

Displeased: Pleased 6.39 1.84 0.99* 0.049 Disgusted: Contented 6.32 1.97 0.98* 0.048 Dissatisfied: Satisfied 6.33 1.91 0.98* 0.049 Unhappy: Happy 6.28 2.01 0.98* 0.048

Patronage Intention: Likelihood (� = 0.98) 6.15a 1.64a 0.96b

How likely is it that you would eat at this restaurant on your lunch break?

6.30 1.62 0.99* 0.050

How likely is it that you would choose this restaurant for a future dining experience?

5.99 1.66 0.97* 0.049

Word of Mouth: Likelihood (� = 0.99) 5.35a 1.73a 0.97b

How likely is it that you would say positive things about this restaurant to other people.

5.36 1.85 0.98* 0.041

K.J. Harris et al. / International Journal of Hospitality Management 60 (2017) 67–76 73

Table 1 (Continued)

Constructs and Indicators Mean: A Mean: C Std. Est. Std. Err.

How likely is it that you would recommend this restaurant to someone who seeks your advice.

5.42 1.68 0.99* 0.042

How likely is it that you would encourage friends and relatives to patronize this restaurant.

5.27 1.67 0.98* 0.042

Fit: �2 = 1752.97, df = 553; RMSEA = 0.04; CFI = 0.98; NFI = 0.97; TLI = 0.98. a = Construct mean. b = AVE.

Table 2 Validity assessment criteria and correlation matrix.

CR AVE MSV ASV PRE PRO II FSC SE ATT SAT INT WOM

PRE 0.86 0.55 0.20 0.04 0.74a

PRO 0.85 0.59 0.36 0.05 −0.09 0.77a

II 0.92 0.69 0.04 0.01 0.21 −0.02 0.83a

FSC 0.95 0.84 0.03 0.01 0.18 0.09 0.05 0.92a

SE 0.90 0.59 0.36 0.07 −0.44 0.60 −0.10 −0.01 0.77a

ATT 0.99 0.99 0.96 0.35 −0.06 0.02 0.00 −0.08 0.07 0.99a

SAT 0.99 0.96 0.96 0.35 −0.05 0.05 0.01 −0.09 0.07 0.98 0.98a

INT 0.98 0.96 0.95 0.35 −0.06 0.04 0.02 −0.08 0.08 0.97 0.98 0.98a

WOM 0.99 0.97 0.90 0.33 −0.05 0.06 0.05 −0.07 0.08 0.92 0.93 0.95 0.98a

C al infl r mouth

4

m r t C c h s H a i A d s i m t

t w t t ( a

f s t w R i h p s r H a c

R = critical ratio; PRE = prevention focus; PRO = promotion focus; II = interperson estaurant; SAT = expected satisfaction; INT = patronage intention; WOM = word of

a = square root of AVE.

.2. Nested structural model

After confirming the validity of the constructs in the measure- ent model, a structural model was specified according to the

elationships hypothesized in Fig. 1. The fit of the structural model o the data was relatively good (�2 = 2,064.7, df = 570; RMSEA = 0.05; FI = 0.98; NFI = 0.97; TLI = 0.97). However, as this research is con- erned with the extent to which the hypothesized relationships old true across both high and low health scores, a nested model pecification was necessary to test for the moderation specified in ypotheses 1–5. To test for the moderating effect of health score,

two group model was specified split between the respondents n the A score condition and respondents in the C score condition. s expected, the fit of the two-group specification (�2 = 2,815.5, f = 1140; RMSEA = 0.03; CFI = 0.96; NFI = 0.93; TLI = 0.96) was a ignificantly better fit to the data than the non-nested model spec- fication (��2 = 750.8; �df = 570; p < 0.001). As such, the nested

odel specification was accepted over the non-nested specifica- ion.

After determining that the nested model was a significantly bet- er fit to the data, the parameter estimates of the specified paths ere compared between the two groups to determine the extent

o which the direction and significance of the estimates were struc- urally analogous across the manipulated health score conditions Table 3). As expected, the estimates were significantly different cross the parameters reflecting each of the first five hypotheses.

The results of the nested model analysis indicated full support or Hypothesis 1 and for Hypotheses 3–5. Hypothesis 2 was partially upported. Broadly speaking, the results confirm the hypothesis hat health scores identified as good are processed heuristically hile health scores identified a poor are processed systematically. egarding Hypothesis 1, a promotion focus was found to be a signif-

cant predictor of participants’ attitude toward the restaurant. As ypothesized, when the score was good, a promotion focus was a ositive predictor of attitude (� = 0.24, p < 0.001) while when the core was poor, a promotion focus was a negative predictor of estaurant attitude (� = −0.19, p < 0.001). Likewise, in support of

ypothesis 2a, a prevention focus was not a significant predictor of ttitude in the good score condition (� = −0.07, p = 0.19). However, ontrary to what was hypothesized for Hypothesis 2b, a prevention

uence congruence; FSC = food safety concern; SE = self-efficacy; ATT = attitude to .

focus was identified as a positive predictor of restaurant attitude in the poor health score condition (� = 0.13, p = 0.01). As this path was hypothesized to be negative based on the tenets of information processing theory, only partial support accrues to Hypothesis 2.

Similar moderated effects were identified for Hypotheses 3–5. Susceptibility to interpersonal influence was not a significant pre- dictor of attitude in the good score condition (� = −0.07, p = 0.11) but was significantly positive in the poor score condition (� = 0.24, p < 0.001). Likewise, food safety concern was not a significant pre- dictor of attitude in the good score condition (� = 0.06, p = 0.16) but was a significantly negative predictor in the poor score condition (� = −0.17, p < 0.001). Finally, self-efficacy was not a significant pre- dictor of attitude in the good score condition (� = −0.01, p = 0.84) but was significantly positive in the poor score condition (� = 0.19, p = 0.004). These results provide full support for Hypotheses 3–5.

Regarding the behavioral constructs, no moderation was estab- lished. Regardless of the health score, there was a positive effect of attitude toward the restaurant on expected satisfac- tion (�AScore = 0.86, p < 0.001; �BScore = 0.87, p < 0.001). These results support Hypothesis 6. Likewise there was a positive effect of expected satisfaction on patronage intention (�AScore = 0.79, p < 0.001; �BScore = 0.90, p < 0.001) and on word of mouth inten- tion (�AScore = 0.60, p < 0.001; �BScore = 0.79, p < 0.001). Again, these results were consistent regardless of the reported health score. Thus, the results provide full support for Hypotheses 7 and 8 as well. The implications of these results are discussed in detail in the following section.

5. Discussion

Food safety is of paramount concern in the hospitality indus- try, and an integral part of preventing outbreaks of foodborne illness is the food establishment inspection process. While, many states require the posting of inspection results within the restau- rant (McVicar, 2011), few studies have explored the mechanisms through which the consumer interprets this information. Accord-

ingly, the ways in which consumers use this information to make assessments about the establishment is not well understood (Filion and Powell, 2011). To address these issues, this study used information processing theory as a framework to investigate the

74 K.J. Harris et al. / International Journal of Hospitality Management 60 (2017) 67–76

Table 3 Results of nested model analysis.

Parameter Group: A Group: C

̌ S.E. ̌ S.E.

Promotion Focus-Attitude to Restaurant 0.242** 0.044 −0.193** 0.060 Prevention Focus-Attitude to Restaurant 0.069 0.041 0.133* 0.054 Interpersonal Influence-Attitude to Restaurant −0.066 0.033 0.240** 0.041 Food Safety Concern-Attitude to Restaurant 0.056 0.031 −0.166** 0.041 Self Efficacy-Attitude to Restaurant −0.013 0.049 0.189** 0.068 Attitude to Restaurant-Expected Satisfaction 0.863** 0.031 0.874** 0.027 Expected Satisfaction-Patronage Intention 0.794** 0.043 0.896** 0.032 Expected Satisfaction-Word of Mouth 0.602** 0.046 0.788** 0.030

F

m s c t c

a p e f p b a c e r r t w q r r a

s w a e t r a t p p h t m A e s b t i o

c r i r s

it: �2 = 2815.53, df = 1140; RMSEA = 0.033; CFI = 0.96; NFI = 0.93; TLI = 0.96. ** p < 0.001. * p = 0.01

oderating impact of restaurant inspection scores on the relation- hips among regulatory focus, interpersonal influence, food safety oncern, and self-efficacy. Upon establishing the nomological rela- ionships between these constructs, the downstream effects on onsumer attitudes and behaviors were established.

Based on information processing theory (Chaiken, 1987; Petty nd Cacioppo, 1986), this research suggested that customers would rocess the information contained in the food safety rating score ither heuristically or systematically, based on their regulatory ocus. Individuals with a promotion focus were proposed to be redisposed to processing food safety information heuristically, ecause a promotion focus causes them to perceive the landscape s generally safe, negating the need to gather information and pro- ess too deeply. Thus, it was hypothesized that (1) when a person ncounters a good health safety score, indicating that the restau- ant is a safe place to eat, he will have a positive attitude toward the estaurant, and (2) if the rating is poor, he will have a negative atti- ude toward the restaurant. The results supported this prediction. It ould appear that promotion-focused consumers do indeed make

uick, heuristic judgments without thinking too deeply. When the ating was an A, the participants had a good attitude toward the estaurant, and when the rating was a C, they exhibited a negative ttitude toward the restaurant.

Conversely, it was expected that for prevention-focused con- umers, the predisposition to prevent or avoid negative outcomes ould cause them to process information about food safety system-

tically, in an effort to prevent foodborne illness. In this case, it was xpected that the relationship between prevention focus and atti- ude toward the restaurant would be insignificant when the safety ating was an A, but negative and significant when the rating was

C. While the former proposition was supported (Hypothesis 2a), he latter proposition (Hypothesis 2b) was not. Hypothesis 2b pro- osed that when a prevention-focused person, who was already redisposed to preventing negative outcomes, was faced with poor ealth inspection for the restaurant, this would cause them to hink more critically and arrive at the conclusion that they were

ore likely to contract a foodborne illness at this establishment. ccordingly, this was expected to lead to a negative attitudinal valuation of the restaurant. However, this hypothesis was not upported. When the food safety score was a C, the relationship etween prevention focus and attitude was significant and posi- ive. The explanation for this unexpected outcome potentially lies n the additional information the customer may have gathered in rder to make an evaluation (i.e., Hypotheses 3–5).

The relationships between interpersonal influence, food safety oncern, self-efficacy and attitude were significant only in the poor

ating condition (when the restaurant received a grade of C). It s notable that these relationships were insignificant when the estaurant received a rating of A, but when the C rating prompted ystematic processing, these relationships all became significant.

These results lend support to the idea that information processing is an important factor in how consumers assimilate information about restaurant inspection ratings.

Regarding the lack of support for Hypothesis 2b, it is possi- ble that when the customer evaluated the additional information regarding interpersonal influence, food safety concern, and self- efficacy, the integration of this information shifted his attitudinal evaluations in a way that was not expected. For example, per- haps the prevention-focused customer did note the C rating, but was influenced by his friends and co-workers’ attitudes about the restaurant. It is also possible that this consumer was well aware that a rating of C was not good, but upon further examination, decided that he was really not all that concerned about food safety issues in general. Waters et al. (2013) found that food safety inspection results have little predictive value in whether or not a foodborne illness will occur, and Henson et al. (2006) reported that consumers use much more than inspection results to determine whether a restaurant is a safe choice, including visible cues of food, staff and their overall personal inspection once inside. Thus, while this is not a definitive explanation, the results do suggest such a possibility, and provide an interesting area for future research.

The above results have important theoretical implications for the hospitality literature, as this study represents a novel approach to understanding the ways in which customers integrate informa- tion about a restaurant’s safety inspections. The findings suggest that consumer reaction to the information contained in food safety ratings is not as simple as it may appear at first glance. This study adds to the extant research of information processing and regula- tory focus theory to demonstrate how these two frameworks work together to influence consumers’ attitudes and behaviors in reac- tion to new pieces of information. The results also extend the body of literature on food safety and restaurant inspection scores by indicating that individual-level traits have a profound impact on customer reactions to posted food safety ratings.

5.1. Practical applications

The results of this study can serve to inform public policy. Not all states have legislated requirements that foodservice establish- ments post their food safety scores on site in a clear manner for their patrons. However, the results of this study clearly demon- strate that food safety signs can influence customer perceptions of a restaurant. Strengthening consumer protections by requiring clearly visible signs for restaurants can help promote food safety in states due to the countervailing economic force of consumer choice.

Additionally, regulating where information is posted should be

important to those involved food safety inspection programs and restaurant companies. In states where signs are posted physically at a restaurant, many consumers may order food online or even make their choice of where to dine by scanning information on websites

of Hos

s o f o t n

o i t h c p s h h m

5

n n f w t s r

o u r c W p n r

b t l a e H t i

R

A

B

B

B

B

B

B

C

K.J. Harris et al. / International Journal

uch as GrubHub, Penny Deliveries, and Yelp, and may not be aware f the inspection score. In 2014, $600 million was invested in online ood ordering companies (Kim, 2015), suggesting that online food rdering is quickly gaining popularity. Forcing food safety scores o be displayed prominently on food ordering websites may be the ext generation of regulations.

Finally, restaurateurs may use the results of this study for their wn benefit. Not only do the results of the study suggest that t is important to have a good food safety score, but that posi- ive food safety scores stimulate consumers to process information euristically. Consumers may not systematically evaluate menu hoices, seating choices, or other decisions when dining due to a ositive food safety score. Thus, offering easy to access inspection cores, presenting customers with staff who practice high personal ygiene, and maintaining a clean environment could effectively elp consumers efficiently and positively process the various ele- ents of their restaurant experience.

.2. Limitations and future research

While this research makes several important contributions, a umber of questions remain unanswered. First, it is important to ote that this study was conducted in the United States where

ood safety norms are often quite different from other places in the orld. As such, the results of the research should not be prema-

urely generalized outside of this cultural context. Future research hould continue to examine the relationships proposed in this esearch in other cultural and national contexts.

Second, it is also important to acknowledge that the posting f health scores in an alphabetical format is not the only method sed by inspection agencies to communicate food safety inspection esults. Reporting norms vary by state and may feature numeri- al scores or color-coded cards to communicate inspection results.

hile the use of an A versus C score in the present research was urposeful to establish two-group moderation, the findings do not ecessarily correspond to all forms of inspection reporting. Future esearch should expand the framework to other forms of reporting.

Finally, the boundary conditions imposed by the scenario should e mentioned. For example, in the methods, respondents were told hat they would be making a lunch decision with a group of col- eagues. The decision to include relevant others was purposeful to llow for the exploration of susceptibility to interpersonal influ- nce and dining out is, more often than not, a shared experience. owever, because an “others” group was included in the methods,

he results do not necessarily reflect how individuals would process nformation if they were dining alone.

eferences

jzen, I., Fishbein, M., 1980. Understanding Attitudes and Predicting Social Behavior. Prentice-Hall, Englewood Cliffs, NJ.

aker-White, A., 2014. The ABCs of Restaurant Inspections: Online Disclosure and Grading Systems. The Network for Public Health Law, https://www. networkforphl.org/the network blog/2014/04/04/435/the abcs of restaurant inspections online disclosure and grading systems. accessed 22.04.16.

andura, A., 2001. Social cognitive theory: an agentic perspective. Annu. Rev. Psychol. 52, 1–26.

earden, W.O., Netemeyer, R.G., Teel, J.E., 1989. Measurement of consumer susceptibility to interpersonal influence. J. Consum. Res. 15 (4), 473–481.

ohner, G., Chaiken, S., Hunyadi, P., 1994. The role of mood and message ambiguity in the interplay of heuristic and systematic processing. Eur. J. Social Psychol. 24 (1), 207–221.

uhrmester, M., Kwang, T., Gosling, S.D., 2011. Amazon’s Mechanical Turk a new source of inexpensive, yet high-quality, data? Perspect. Psychol. Sci. 6 (1), 3–5.

urnkrant, R.E., Cousineau, A., 1975. Informational and normative social influence

in buyer behavior. J. Consum. Res. 2, 206–215.

enters for Disease Control, 2013. Surveillance for Foodborne Disease Outbreaks United States, 2013: Annual Report, Retrieved from http://www.cdc.gov/ foodsafety/pdfs/foodborne-disease-outbreaks-annual-report-2013-508c.pdf. accessed 09.04.16.

pitality Management 60 (2017) 67–76 75

Centers for Disease Control, 2016. Centers of Disease Control Vessel Sanitation Program Report (2016), Retrieved from http://www.cdc.gov/nceh/vsp/surv/ gilist.htm#2016. accessed 09.04.16.

Chaiken, S., 1987. The heuristic model of persuasion. In: Zanna, M.P., Olson, J.M., Herman, C.P. (Eds.), Social Influence: The Ontario Symposium, vol. 5. Lawrence Erlbaum Associates, Inc, Hillsdale, NJ, pp. 3–39.

Chaplin, L.N., John, D.R., 2010. Interpersonal influences on adolescent materialism: a new look at the role of parents and peers. J. Consum. Psychol. 20 (2), 176–184.

Chitturi, R., Raghunathan, R., Mahajan, V., 2008. Delight by design: the role of hedonic versus utilitarian benefits. J. Mark. 72, 48–63.

Choi, J., Miao, L., Almanza, B., Nelson, C.D., 2013. Consumers’ responses to restaurant inspection reports: the effects of information source and message style. J. Foodserv. Bus. Res. 16 (3), 255–275.

Curran, P.J., West, S.G., Finch, J.F., 1996. The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis. Psychol. Methods 1 (3), 16–29.

DiPietro, D.B., Parsa, H.G., Gregory, A., 2011. Restaurant QSC inspections and financial performance: an empirical investigation. Int. J. Contemp. Hosp. Manage. 23 (7), 982–999.

Food and Drug Administration, 2009. Food Code 2009, http://www.fda.gov/Food/ GuidanceRegulation/RetailFoodProtection/FoodCode/ucm2019396.htm. accessed 22.04.16.

Food and Drug Administration, 2011. FDA Food Safety Modernization Act, http:// www.fda.gov/Food/GuidanceRegulation/FSMA/. accessed 22.04.16.

Food and Drug Administration, 2013. Food Code 2013, http://www.fda.gov/Food/ GuidanceRegulation/RetailFoodProtection/FoodCode/ucm374275.htm. accessed 22.04.16.

Fatimah, U.Z.A.U., Boo, H.C., Sambasivan, M., Salleh, R., 2011. Foodservice hygiene factors-the consumer perspective. Int. J. Hosp. Manage. 30 (1), 38–45.

Filion, K., Powell, D., 2011. Designing a national restaurant inspection disclosure system for New Zealand. J. Food Prot. 74 (11), 1869–1874.

Fishbein, M., 1963. An investigation of the relationship between beliefs about an object and the attitude toward that object. Hum. Relat. 16, 233–240.

Food Safety News, 2016. Retrieved from http://www.foodsafetynews.com/2016/ 02/cdc-declares-chipotle-e-coli-outbreaks-over-cause-unknown/#. VwlHsPkrJD8. accessed 09.04.16.

Ford, J.D., Ellis, E., 1980. A re-examination of group influence on member brand preference. J. Mark. Res. 17, 125–132.

Fornell, C., Larcker, D.F., 1981. Evaluation of structural equation models with unobservable variables and measurement error. J. Mark. Res. 28, 39–50.

Frijda, N.H., 1988. The laws of emotion. Am. Psychol. 43, 349–358. Hair Jr., J.F., Black, W.C., Babin, B.J., Anderson, R.E., Tatham, R.L., 2006. Multivariate

Data Analysis, 6th ed. Pearson Prentice Hall, Upper Saddle River, NJ. Henson, S., Majowicz, S., Masakure, O., Sockett, P., Jones, A., Hart, R., Carr, D.,

Knowles, L., 2006. Consumer assessment of the safety of restaurants: the role of inspection notices and other information cues. J. Food Saf. 26, 275–301.

Higgins, E., 1998. Promotion and prevention: regulatory focus as a motivational principle. Adv. Exp. Social Psychol. 30, 1–46.

Kim, E., 2015. A Secular Shift to Online Food Ordering. Crunch Network, Retrieved from http://techcrunch.com/2015/05/07/a-secular-shift-to-online-food- ordering/ accessed 02.04.16.

Kline, R.B., 2005. Principles and Practice of Structural Equation Modeling, 2nd ed. The Guilford Press, New York, NY.

Lockwood, P., Jordan, C.H., Kunda, Z., 2002. Motivation by positive or negative role models: regulatory focus determines who will best inspire us. J. Pers. Soc. Psychol. 83, 854–864.

Mackie, D.M., Worth, L.T., 1989. Processing deficits and the mediation of positive affect in persuasion. J. Pers. Soc. Psychol. 57, 27–40.

Mason, W., Suri, S., 2011. Conducting behavioral research on Amazon’s Mechanical Turk. Behav. Res. Methods 44 (1), 1–23.

McGuire, W.J., 1968. Personality and susceptibility to social influence. In: Borgatta, E.F., Lambert, W.W. (Eds.), Handbook of Personality Theory and Research. Rand McNally, Chicago, pp. 1130–1187.

McVicar, B., 2011. Some States Grade Restaurant Based on Health Inspections, Retrieved from http://www.mlive.com/news/muskegon/index.ssf/2011/08/ some states grade restaurants.html. accessed 02.04.16.

Mourali, M., Laroche, M., Pons, F., 2005. Individualistic orientation and consumer susceptibility to interpersonal influence. J. Serv. Mark. 19 (3), 164–173.

Neuberg, S.L., Fiske, S.T., 1987. Motivational influences on impression formation: outcome dependency, accuracy-driven attention, and individuating processes. J. Pers. Soc. Psychol. 53 (3), 431–444.

Ortega, D.L., Wang, H.H., Wu, L., Olynk, N.J., 2011. Modeling heterogeneity in consumer preferences for select food safety attributes in China. Food Policy 36 (2), 318–324.

Park, C.W., Lessig, P., 1977. Students and housewives: differences in susceptibility to reference group influence. J. Consum. Res. 4, 102–110.

Petty, R.E., Cacioppo, J.T., 1986. Communication and Persuasion: Central and Peripheral Routes to Attitude Change. Springer, New York, NY.

Petty, R.E., Cacioppo, J.T., Goldman, R., 1981. Personal involvement as a determinant of argument-based persuasion. J. Pers. Soc. Psychol. 41 (5), 847–855.

Pittman, T.S., D’Agostino, P.R., 1989. Motivation and cognition: control deprivation and the nature of subsequent information processing. J. Exp. Soc. Psychol. 25 (6), 465–480.

7 of Hos

P

R

S

S

S

J. Environ. Health 75 (10), 8–12. Yeung, R., Yee, W.M., 2012. Food safety concern: incorporating marketing

6 K.J. Harris et al. / International Journal

odsakoff, P.M., MacKenzie, S.B., Lee, J.-Y., Podsakoff, N.P., 2003. Common method biases in behavioral research: a critical review of the literature and recommended remedies. J. Appl. Psychol. 88 (5), 879–903.

atner, R., Khan, B.E., 2002. The impact of private versus public consumption on variety seeking behavior. J. Consum. Res. 29, 246–257.

cholz, U., Doña, B.G., Sud, S., Schwarzer, R., 2002. Is general self-efficacy a universal construct? Psychometric findings from 25 countries. Eur. J. Psychol. Assess. 18 (3), 242–251.

chwarz, N., Bless, H., Bohner, G., 1991. Mood and persuasion: affective states influence the processing of persuasive communications. Adv. Exp. Social

Psychol. 24, 161–199.

chwarz, N., 1990. Feelings as information: informational and motivational functions of affective states. In: Higgins, E.T., Sorrentino, R.M. (Eds.), Handbook of Motivation and Cognition: Foundations of Social Behavior, vol. 2. Guilford Press, New York, pp. 527–561.

pitality Management 60 (2017) 67–76

Sherer, M., Maddux, J.E., 1982. The self-efficacy scale: construction and validation. Psychol. Rep. 51, 663–671.

Tarca, S., Murphy, K.S., 2014. Food safety inspections and storefront grade posting: the case of NYC and LA. J. Hosp. Tour. Case Stud. 3 (4), 21–26.

Taylor, S.E., Fiske, S.T., 1978. Salience, attention, and attribution: top of the head phenomena. Adv. Exp. Social Psychol. 11, 249–288.

Waters, B., VanDerslice, J., Porucznik, C.A., Kim, J., deLegge, R., Durrant, L., 2013. Impact of internet posting of restaurant inspection scores on critical violations.

strategies into consumer risk coping framework. Br. Food J. 114 (1), 40–53. Zeithaml, V.A., Berry, L.L., Parasuraman, A., 1996. The behavioral consequences of

service quality. J. Mark. 60 (2), 31–46.