key: cord-0966132-rjkvyhhd
authors: Wen, Han; Liu-Lastres, Bingjie
title: Consumers' dining behaviors during the COVID-19 pandemic: An application of the Protection Motivation Theory and the safety signal framework
date: 2022-03-21
journal: nan
DOI: 10.1016/j.jhtm.2022.03.009
sha: ef5fe042207fbee81920ec03f09b215a8e4ed505
doc_id: 966132
cord_uid: rjkvyhhd

With the long-lasting impacts of the COVID-19 pandemic, it is critically important that restaurateurs understand predictors of consumers' dining behaviors to better foster strategies to recover their revenue during the re-opening stage. Based on the safety signal framework and the Protection Motivation Theory, this study developed and tested a model investigating the combined effects of restaurateurs' measures and consumers' protective motivations on their dine-out frequencies and dine-in likelihoods. Consistent with propositions of the Protection Motivation Theory, the results confirmed that both the threat and coping appraisals influenced consumers’ dining behaviors. The coping appraisal process is affected by “access to servicescape,” “servicescape,” and “communication.” Additionally, the results of the gap analysis revealed four safety signaling strategies perceived as effective by consumers but with a low implementation rate in the restaurant industry. Theoretical and practical implications were provided to restaurateurs.

The year 2020 and moving on has been challenging for the restaurant industry in the 20 U.S.. According to the National Restaurant Association [NRA] report, the total sales were $659 21 billion in 2020, which was $240 billion lower than expected (NRA, 2021) . In addition, due to the 22 reduced visitation of customers during the COVID-19 pandemic, nearly 110,000 restaurant 23 locations were temporarily or permanently closed (Dube, Nhamo, & Chikodzi, 2021 ; NRA, 24 2021). As a result, total employment was down by almost one-fourth compared with the 25 expected number in the absence of a pandemic (NRA, 2021). Among those industries that have 26 been negatively impacted by the pandemic, the restaurant industry was hit the hardest (Gössling 27 et al., 2020) . The lockdown and the seating capacity restrictions led to reduced revenue; making 28 it especially difficult for restaurants to survive. For example, in South Carolina, researchers 29 found that nearly 25% of the restaurants they studied did not survive during the lockdown period 30 in 2020 . 31

Due to the fear of COVID-19, the need for social distancing, and the increasing health 32 and safety concerns, consumers' perceptions and behaviors have changed significantly since the 33 pandemic (Prentice et al., 2021) . According to the results of a marketing study, the majority of 34 the consumers believed that foods purchased from groceries stores were safer than those sold by 35 restaurants (Datassential, 2020) . Consequently, consumers' dine-out frequencies dropped 36 significantly at the beginning of the pandemic, with restaurant visitation declining by nearly 90% 37 (Dube et al., 2021) . In addition, in March and April 2020, there were almost no seated diners. 38 Kleinaltenkamp, 2004; Moeller, 2008) . Moreover, "servicescape" is typically manifested through 87 four aspects: "physical environment," "tangibles," "staff," and "other customers." Once 88 customers enter the service environment, the physical environment (i.e., servicescape) and the 89 moveable tangibles will trigger different safety signals to customers (Bitner, 1992) . It is further 90 pointed out that the social elements inside the service environment, such as staff and other 91 customers, may become potential sources of safety signals (Rosenbaum & Massiah, 2011) . 92

Lastly, "communication" means that service providers can communicate with customers about 93 their safety protocols (Bove et al., 2020) . 94

Hospitality researchers have conducted various studies to examine how restaurants can 95 make customers feel safer during their dining experiences, especially in COVID times. For 96 instance, through a quasi-experimental design study, Taylor (2020) found that customers 97 preferred partitions in-between tables to mannequins at tables as ways to set up social distance 98 between customer groups in restaurants. The partitions can be used as a strategy in servicescape 99 to increase consumers' intentions to dine out in restaurants (Taylor, 2020) . By conducting four 100 experiments and integrating the theories about the psychological effects of risk, Kim and Lee 101 (2020) found that consumers felt more comfortable sitting at a private dining table or in a private 102 dining room when dining in full-service restaurants during the pandemic. This study proved that 103 the availability of private dining could be part of the servicescape that reduces consumers' risk 104 perceptions when eating out (Kim & Lee, 2020) . 105

As the restaurant industry returns to normal operation, previously studied measures such 106 as partitions and private dining rooms may not always be applicable. Thus, guided by the safety 107 signal framework, this study proposed testing the impacts of six broad categories of safety 108 signals on tourists' perceptions. In addition, the Protection Motivation Theory (PMT) suggests 109 that response efficacy, which is a key element in the coping appraisal process, reflects people's 110 evaluation of the perceived effectiveness of protection measures (Roger, 1975 H1a: Access to servicescape has a positive impact on restaurants' response efficacy. 115

H1b: Servicescape has a positive impact on restaurants' response efficacy. 116

H1c: Tangibles have a positive impact on restaurants' response efficacy. 117

H1d: Technology has a positive impact on restaurants' response efficacy. 118

H1e: Other customers have a positive impact on restaurants' response efficacy. 119

H1f: Communication has a positive impact on restaurants' response efficacy. 120 121

The Protection Motivation Theory (PMT) was proposed by Roger (1975) to explain how 123 emerging health issues affect individuals' attitudes and behavioral changes Dunn, 1997). According to PMT, an individual's selection and decision of behaviors can be 125 impacted by fear-arousing communications (Roger, 1975) . Although, initially, PMT was 126 designed for health-related issues; nowadays, PMT has been widely applied in studies of various 127 disciplines, including the hospitality and tourism management field. In the PMT research model, 128 individuals typically reach their decisions through two cognitive mediating processes (Floyd et 129 al., 2000) . One is the threat appraisal process, while the other is the coping appraisal process 130 the risk, then determine its severity and their vulnerability to the risk (Harris et al., 2018; Harris 133 et al., 2014) . 134 Furthermore, in the coping appraisal process, response efficacy describes an individual's 135 beliefs in the effectiveness of a recommended behavior or a protective measure, that is, how 136 likely the threat will be removed (Roger, 1975) . The outcome of the PMT model, therefore, 137 involves individuals' intention or decision to either initiate, continue, or inhibit the corresponding 138 adaptive responses (Floyd et al., 2000) . Based on propositions of the PMT, the following 139 hypotheses were proposed: Researchers also noticed the role of trust, whose impacts varied by context. By building trust into 166 the research model, researchers identified that government and social trusts mediated the 167 relationship between customers' response efficacy and their intention to stay in hotels (Hsieh et 168 al., 2021) . In Turkey, researchers found that the relationship between motivational factors and 169 consumers' visiting intentions to upscale restaurants was moderated by their risk perceptions 170 toward COVID-19 and their trust toward the government (Dedeoğlu & Boğan, 2021) . 171

According to the literature mentioned above, risk perceptions and trust both played 172 important roles in shaping consumers' behaviors during the pandemic. In the context of the 173 current study, the output behavior is defined as consumers' dine-out frequency and the likelihood 174 to eat inside the restaurant. Therefore, based on the two cognitive mediating processes of PMT 175 (Floyd et al., 2000) and previous literature, the current study proposed that both risk perceptions 176 and trust toward the restaurant had significant impacts on consumers' dine-out frequencies and Further, participants were asked to indicate their intention to dine in the restaurant again 225 in the next three months, from 1 being "Extremely Unlikely" to 7 being "Extreme Likely." In 226 terms of dine-out frequencies, participants were asked to indicate their frequencies of dining in 227 full-service restaurants during the last three months. Their answers were coded into 1 being 228 "never," 2 being "monthly," and 3 being "weekly." This study primarily focused on the context 229 of full-service restaurants, where participants' dining behaviors and safety signals were all 230 positioned in full-service restaurants with sit-down, in-person services. 231 232

All data in this research were collected by a survey sampling company, where the link to 234 the survey was distributed to their panel members who belong to our target population -235 restaurant consumers (age greater than 18 years old) in the U.S. A pilot test (N=30) was 236 conducted through the survey sampling company before the formal data collection. The 237 reliability of survey constructs and the clarity of survey instructions were checked at the pilot 238 test. As all constructs showed Cronbach Alpha levels higher than 0.7 (Nunally, 1978) According to Peng and Lai (2012), the minimum sample size for using the SEM-PLS method is 257 ten times the largest number of indicators in the constructs. Therefore, the minimum sample size 258 of this study is 90, as the largest number of indicators is 9, for dine-in likelihood. As 

In the survey instrument, participants were asked to rate their perceived effectiveness of 323 preventive measures in restaurants and whether the restaurant of their latest visit had 324 implemented these methods. A gap analysis (mirror the "Importance Performance Analysis 325

[IPA]" by Martilla and James, 1977) was performed to find out the areas for improvement for 326 restaurants. Figure 3 illustrates the locations of these strategies in the four-quadrant matrix. The mean 332 value of perceived effectiveness is 3.52, while the mean value of implementation is 39.4%. 333

These mean values were plotted in the figure as red dashes horizontally and vertically. Quadrant 334 I ("Keep up the Good Work") included seven strategies, which refer to those strategies that are 335 perceived as effective by consumers and widely implemented in restaurants. Quadrant III ("Low 336

Priority") refers to those strategies that are not perceived as highly effective but also not 337 frequently used by restaurateurs. A total of nine strategies fell into this quadrant. Two strategies 338 were in Quadrant IV ("Possible Overkill"), which refers to those strategies that were widely 339 adopted by restaurants but not being perceived as very effective by consumers. 340

Quadrant II ("Concentrate Here") comprised the strategies perceived as very effective by 341 consumers but not widely implemented in restaurants. A total of four strategies are located in this 342 quadrant, including one item from "Servicescape" (outdoor or balcony tables), two items from 343 "Tangibles" (protective equipment for customers; single-use utensils and dishes), and one item 344 from "Technology" (digital menus). Quadrant II is the most critical area from the gap analyses as 345 it provided us with information about how restaurants can improve in the future to make 346 customers feel safe in an epidemic environment. 347 348 Figure 3 . Effectiveness-Implementation Matrix 349 350

Regarding consumers' perceptions and decision-making, the findings of this study 352 provided direct support to the key propositions of PMT (Floyd et al., 2000) , indicating that the 353 processes of threat and coping appraisals both affected consumers' dining behaviors. More 354 specifically, the results showed that participants' perceived threat of COVID-19 negatively 355 influenced their intention to dine out in restaurants, which is consistent with the prediction of the 356 threat appraisal. On the other hand, the results revealed that response efficacy, which is the key 357 element in the coping appraisal, also affected the sample's dining behaviors. Despite the 358 potential negative impacts of the threat appraisal, it is widely believed that coping appraisal 359 assumes a more critical role in driving people's behaviors in both health and hospitality contexts 360 (Floyd et al., 2000; Liu et al., 2016) . 361

In addition to the support of the effects of these two appraisals processes, the findings of 362 this study suggest that these two processes might be intertwined. For example, the results of this 363 study showed that participants' perceived risk of dining out is shaped by both the pandemic and 364 restaurant management efforts. This finding is evidenced by the statistical relationships between 365 perceived threat, response efficacy, and perceived risk. It is encouraging since this particular 366 finding suggests that consumers' perceived risk of dining out, although greatly influenced by the 367 pandemic, can be re-shaped by having effective strategies in place. This particular finding is also 368 similar to previous studies (e.g., Liu et al., 2016) , which suggest that efficacy might moderate the 369 relationships between threat and the outcome variable in a hospitality context. behaviors and perceptions during the COVID-19 pandemic but also expanded the scope beyond 377 consumers' risk perception and purchasing intentions. More importantly, this study showed that 378 restaurants' active response to the pandemic and newly adopted safety measures could greatly 379 enhance consumer trust, which affects their actual dining activities. These findings present a 380 more comprehensive interpretation of the dynamic between the hospitality industry and 381 consumers. The inclusion of new variables also sheds light on future studies, which can consider 382 replicating this study in a global context and identify potential cross-cultural differences. 383

Furthermore, this study highlights the significance of crafting effective crisis 384 management strategies from practitioners. Response efficacy is a key component in the coping 385 appraisal in PMT and is directly related to individuals' behavioral responses. When it comes to 386 this study, the results showed that participants' level of response efficacy is related to the 387 following components in the safety signal framework: "access to servicescape," "servicescape," 388 and "communication" had significantly predicted the response efficacy of restaurants. In other 389 words, customers' assessment of how safe a restaurant is can be visual and largely depend on (1) 390 how restaurants restrict and control customers access, (2) how restaurants reorganize the layout 391 to facilitate social distancing, and (3) signs and messages used by restaurant operators to 392 communicate COVID-19 safety protocols with incoming customers. 393

To a large extent, these notions are consistent with previous studies, which suggest that 394 changes in servicescape such as using partitions in between tables and setting out private dining 395 facilities are essential in attracting and comforting customers to dine out during pandemic times 396 (Kim & Lee, 2020; Taylor, 2020) . Compared with a previous study conducted by Hsieh, Chen, 397 and Wang (2021), which found that individuals' response efficacy positively influenced 398 consumers' intentions to stay in hotels during the pandemic, the current study revealed that 399 restaurants' response efficacy could also affect consumers' dining-related behaviors through trust 400 and perceived risks. Thus, the findings of this study advance our understanding of the roles 401 response efficacy could play during customers' decision-making in hospitality and tourism. The diversified portfolio will give restaurant operators more options and flexibility to cope with 476 future challenges, and it has become a post-pandemic trend in the restaurant industry. 477

In terms of providing outdoor patio seating or balcony tables, the re-opening guideline 478 published on the New York governor website clearly defines outdoor seating servicescape 479 requirements such as seating capacity and floorplan layout (Govonor.NY.Gov, 2021). However, 480 such a program in New York also received critiques, as it encouraged the privatization of 481 sidewalks and further impacted the public's access to sidewalks (Yang, 2021) . Therefore, 482 restaurant operators should strive to find out a balance between accommodating customers 483 without over-use public spaces. Using partitions to create private dining spaces within the dining 484 room may be a better alternative. In addition, this finding also offers directions for new 485 restaurant developers or those operators who plan to conduct renovation projects. An outdoor 486

patio or rooftop open space is highly recommended for new restaurant projects as it creates 487 flexibility for business owners to deal with future crises (Killifer, 2021) . 488

Technology not only drives changes in the restaurant industry but also offers 489 opportunities for restaurants to better cope with challenges imposed by the global pandemic. 

Consumers' return intentions towards a restaurant with 522 foodborne illness outbreaks: Differences across restaurant type and consumers' dining 523 frequency

Servicescapes: The impact of physical surroundings on customers and 525 employees

Restrict, clean, and protect: signaling consumer safety during 527 the pandemic and beyond

Independent restaurant 529 operator perspectives in the wake of the COVID-19 pandemic

Independent restaurant 532 operator perspectives in the wake of the COVID-19 pandemic

Restaurants and COVID-535 19: What are consumers' risk perceptions about restaurant food and its packaging during 536 the pandemic?

COVID-19 Report 2: Fear and Response (3/17/20)

Dunkin' moves to limited service and removes tables and chairs from its 541 over 8,500 stores in the U.S. Nation's Restaurant News

The motivations of visiting upscale restaurants during the 545

COVID-19 pandemic: The role of risk perception and trust in government

COVID-19 cripples global restaurant and 548 hospitality industry

Blueprinting the service company: Managing service 550 processes efficiently

A meta-analysis of research on 552 protection motivation theory

Structural equation models with unobservable variables and 554 measurement error: Algebra and statistics

Pandemics, tourism and global change: a rapid 557 assessment of COVID-19

Re-opening New York: Outdoor and Take-Out/Delivery Food 559

Services -Guidelines for Employers and Employees

People at risk of flooding: Why some residents take 563 precautionary action while others do not

Effects of COVID-19 pandemic on hospitality industry: 565 Review of the current situations and a research agenda

COVID-19 Study 2 Report: Restaurant and Hotel 568 Industry: Restaurant and hotel customers' sentiment analysis. Would they come back? If 569 they would, WHEN?

A primer on partial least squares 572 structural equation modeling (PLS-SEM)

PLS-SEM: Indeed a silver bullet

Foodborne illness outbreaks in restaurants and patrons' 576 propensity to return

Critical food safety violations 579 in Florida: Relationship to location and chain vs. non-chain restaurants

Government and social trust vs. hotel response 582 efficacy: A protection motivation perspective on hotel stay intention during the COVID-583 19 pandemic

Designing for Flexibility Now and Later

The effect of sensation seeking and 587 emotional brand attachment on consumers' intention to consume risky foods in 588 restaurants

Effects of COVID-19 on preferences for private dining facilities in 590 restaurants

Hope, fear, and consumer behavioral change 592 amid COVID-19: Application of protection motivation theory

Consumer Studies

Source market perceptions: 595 How risky is Jordan to travel to?

Importance-performance analysis

Customer integration-a key to an implementation perspective of service 600 provision

Restaurant Industry Facts at a Glance

Psychometric Theory: 2d Ed

The restaurant industry in recovery

Antecedents and consequences of panic buying: 608 The case of COVID-19

A protection motivation theory of fear appeals and attitude change1. The 610 journal of psychology

Handbook of health behavior research 1: Personal and social determinants

An expanded servicescape perspective

The socially distant servicescape: An investigation of consumer 617 preference's during the re-opening phase. International journal of hospitality 618 management

The effects of crowdedness and safety measures on 620 restaurant patronage choices and perceptions in the COVID-19 pandemic

Risk reduction and adventure 623 tourism safety: An extension of the risk perception attitude framework (RPAF)

Outdoor Dining during COVID? Analyzing the Impact of New York City's 626

Open Restaurants Program (Doctoral dissertation, Columbia University)

Afraid to travel after COVID-19? Self-protection, 628 coping and resilience against pandemic 'travel fear

This study offered various practical implications to managers or operators in the 450 restaurant industry. These implications were essential for the restaurant industry because nearly 451 70% of restaurants in the U.S. are single-unit operations (NRA, 2021), which have minimal 452 access to information or resources circulated in large chain foodservice operations. The study 453 results demonstrated that, even though consumers were hesitant to dine out in restaurants due to 454 the fear of COVID-19, restaurant operators, through adopting effective management measures, 455can increase their response efficacy, which further enhances their level of trust, reduces the 456 perceived risk of dining-out, and results in a higher frequency of dining out. 457Restaurant managers or operators should focus on creating a safe servicescape and 458 utilizing communications strategies as these efforts will significantly improve consumers' 459 perceptions of response efficacy. For capacity management, restaurants can consider removing 460 some of the furniture in the dining room to avoid congregation and promote a mood of social 461 distancing. In fact, Dunkin's, McDonald's, and many other large restaurant chains had already 462 utilized this strategy and removed all furniture in their public space during the pandemic 463 (Dawson, 2020) . This strategy is more applicable to fast foods restaurants or limit-service 464 restaurants where customers would be encouraged to get the order to-go with limited seating 465 available. 466Additionally, restricted seating or reduced seating capacity has imposed challenges for 467 restaurant operators, especially the owners of small independent restaurants . 468 Therefore, restaurant operators can diversify their product offerings to gain revenue from other 469 areas to cover the loss. For example, restaurants can sell meal kits with instructions to help 470 customers quickly prepare foods they like with easy steps. In addition, they can also consider 471 expanding to-go options by adding additional take-home-friendly food items such as customized 472 provide solutions by offering a platform for restaurants to create their online menu through Q.R. 496 codes. 497 498

The current study was not exempted from limitations. First, an online survey instrument 500 was developed and used for data collection that participants may be impacted by the social 501 desirability bias when filling out the online survey. Future studies may use other data collection 502 methods, such as second-hand sales/revenue data, or observations, to gather insights. Second, 503 this study employed the safety signal framework to study consumers' behaviors. There may be 504 other safety measures used by the restaurants that were not included in the current study. developments (e.g., the invention of the vaccine) and regions. The current study did not link 512 consumers' behaviors to these factors. Thus, future research may take it into consideration and 513 explore the impact of regional differences on consumers' behaviors. Forth, the participants of 514 this study were recruited through online panels where no access was provided regarding subjects 515 who refused to take the survey. Future studies should consider using probability-based sampling 516 strategies and conducting non-response bias tests. Last but not least, the COVID-19 pandemic is 517 still ongoing, and consumers' behaviors are changing simultaneously. Longitudinal studies are 518