research discussion

Journal of Hospitality & Tourism Research, Vol. XX, No. X, Month 201X, 1 –35 DOI: https://doi.org/10.1177/1096348020946383 © The Author(s) 2020

1

How To Build a BeTTer roBoT . . . For Quick-Service reSTauranTS

dina Marie v. Zemke Ball State University

Jason Tang carola raab

Jungsun kim University of Nevada, Las Vegas

Hospitality firms are exploring opportunities to incorporate innovative technologies, such as robotics, into their operations. This qualitative study used focus groups to investigate diner perspectives on issues related to using robot technology in quick- service restaurant (QSR) operations. QSR guests have major concerns regarding the societal impact of robotics entering the realm of QSR operations; the cleanliness and food safety of robot technology; and communication quality, especially voice recognition, from both native and nonnative English speakers. Participants also offered opinions about the functionality and physical appearance of robots, the value of the “human touch,” and devised creative solutions for deploying this technology. Surprisingly, few differences in attitudes and perceptions were found between age groups, and the participants were highly ambivalent about the technology. Future research may consider further exploration of robot applications in other restaurant segments, using quantitative methods with a larger sample.

Keywords: restaurants; robotics; qualitative; service encounter; smart technology; customer perception

inTroducTion

The use of technology in hospitality businesses has evolved over time. It originated in information management, where handwritten orders and folios were replaced by point-of-sale systems. The computers used to operate these systems also evolved from punch-card entry mainframes to desktop computers, laptops, pads, and smartphones. For the past 40 years, the evolution in technol- ogy has focused on data entry and data management, and much hospitality

946383 JHTXXX10.1177/1096348020946383JOURNAL OF HOSPITALITY & TOURISM RESEARCHZemke et al. / BeTTer roBoT For QUICK-serVICe resTAUrANTs research-article2020

Authors’ Note: The research team gratefully acknowledges the support of the William F. Harrah College of Hospitality in conducting this study. Jason Tang is now affiliated with Mount Royal University, Calgary, Canada.

2 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

research has focused on how to facilitate acceptance and usage of this technol- ogy, but not necessarily how to influence its creation and presentation.

Robotic technology to perform specific physical tasks has recently emerged as an option for hospitality businesses. Integrating robots into hospitality opera- tions has become more feasible due to decreasing robotic equipment costs. Currently, the majority of robots used in the hospitality industry are technologies initially developed for other industries, such as automobile and food manufac- turing, which have been modified from their original functions to perform their tasks in a hospitality setting. Examples include the work performed by robotic vacuums (now presented as housekeeping robots), information displays (now made mobile on a rolling platform), and robotic manufacturing assembly arms (which now assemble pizzas and cocktails).

The genesis of this study was the “Fight for $15” movement in the United States which focused heavily on hourly service jobs, such as those in franchised quick-service restaurants (QSRs). The minimum hourly wage would increase to at least $15 per hour, and many restaurant operators cautiously suggested that they may explore robotics as an alternative to absorbing these increased labor costs. The most vocal proponent for exploration was Andy Puzder, the former CEO for CKE Enterprises (which owns the Carl’s Jr. and Hardees restaurant brands). While he was not threatening to replace employees with robots, he forthrightly stated that all restaurant companies may need to consider this tech- nology to remain competitive.

The media’s attention to this issue raised the public’s awareness of robotics, and there is evidence of increased incorporation of robotic technology in produc- tion operations. Most of the advances in “robotics” are simply information pads or kiosks for order taking, which do not move or perform a physical task and, as such, do not qualify as robots; however, a few restaurant companies now use actual robotics. For example, the Cali-Burger chain has deployed its “Flippy” hamburger-making robot in its stores. In July 2018, McDonald’s opened its first “all-robot” restaurant in Phoenix, Arizona (which actually is not 100% robot- staffed; humans are required to ensure that the robots are functioning properly).

The current study explored the QSR customer’s perceptions of the use of robotic technology in the QSR industry. Rather than using preexisting measure- ment instruments, which may have limited the range of the guest’s perceptions, this zero-based qualitative study used a focus group technique to elicit this infor- mation from QSR patrons. The participants offered their thoughts on issues that included what robots should or should not do, what they should or should not look like, and the benefits or hazards that could be posed by using robots in QSRs. The objective of this qualitative study was to reveal a broad range of positive and negative perceptions of this technology that will inform the devel- opment of an instrument to gain perceptions from a broader swath of the popula- tion. This study’s results establish a fertile ground for future academic study as well as inform industry practitioners’ efforts to design and implement these tech- nologies into their operations.

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 3

liTeraTure review

defining robotics

International Organization for Standardization (ISO) standard 8373:2012 pro- vides definitions of common robot terminology, officially defining a robot as “an actuated mechanism programmable in two or more axes with a degree of auton- omy, moving within its environment, to perform intended tasks” (International Organization for Standardization, 2012). Robots can be segmented into two pri- mary types, industrial and service. The International Federation of Robotics (2016) defines a service robot as one that autonomously performs useful tasks for humans or equipment outside an industrial automation application without human intervention. Robots are subsequently categorized as personal service robots or professional service robots. Personal service robots are utilized in a noncommer- cial setting and include examples such as automated wheelchairs and personal mobility assistive robots, while professional service robots are utilized for com- mercial tasks, such as to make deliveries or for cleaning. In addition, professional service robots require a human operator to start, monitor, and stop the robot’s operation.

The physical appearance of robots. Past research into the physical appear- ance of robots is rooted in anthropomorphism, or how nonhuman creatures or objects can mimic the appearance of humans, and the process of disambigua- tion, which proposes a model that explains the human response to anthropomor- phic presentations of nonhuman objects and images.

Anthropomorphism. Anthropomorphism has been broadly defined as the attribution of human character and behavior to nonhuman entities (Bartneck et al., 2007). Examples of anthropomorphic representations include toys, adver- tising images, video game characters, and robotic devices that bear some resem- blance to a human form and human behavior. In the context of robotics, anthropomorphism refers to the degree to which a robot resembles a human through visual cues, movements, and communications (Murphy, Gretzel, et al., 2017). These elements of anthropomorphism are critical when evaluating how robots are used in a service context, as they assist in predicting the degree to which guests perceive robots to express moral care, concern, responsibility, trust, and social influence during social interactions (Waytz et al., 2014). These elements are particularly critical when examining human-robot interactions (Murphy, Hofacker, et al., 2017).

Companies have successfully incorporated anthropomorphism to encour- age brand attachment and loyalty (Veer, 2013). As a result, Murphy, Gretzel, et al. (2017) proposed a new construct, anthropomorphism loyalty, which could be applied to human robot interactions. Anthropomorphism loyalty sug- gests that anthropomorphism could potentially create a level of attachment and loyalty previously unattainable with inanimate digital objects. By exten- sion, anthropomorphism loyalty could be applied to animated physical objects, such as robots.

4 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

Disambiguation and the “uncanny valley hypothesis.” Disambiguation is a cognitive system of category processing through which humans make sense of, or remove the ambiguity posed by, an ambiguous image or object. It is a natural response that humans employ to make sense of language, images, or other situ- ations that are unclear. The results of this process can range from acceptance/ positive affect through repulsion/negative affect. Mori (1970, 2012) proposed the “uncanny valley hypothesis” to explain the relationship between an image’s human likeness and shinwa-kan. Here, human likeness is an example of anthro- pomorphism, the attribution of human form, character or behavior to nonhuman entities (Bartneck et al., 2007; Epley et al., 2008). Mori’s use of the Japanese term shinwa-kan was translated into English as familiarity (MacDorman, 2005; MacDorman & Ishiguro, 2006). However, as the concept of the uncanny valley grew in popularity, shinwa-kan has also been translated into other English terms, including likeability (Bartneck et al., 2007), affinity (Bartneck et al., 2007; MacDorman & Entezari, 2015; Mori, 2012), emotional response (Blow et al., 2006), and pleasantness (Seyama & Nagayama, 2007).

Mori (1970, 2012) posited that the likability of nonhuman entities increases as anthropomorphism increases. However, when anthropomorphism reaches a point where human replicas appear nearly, but not exactly, like human beings, feelings of distress, eeriness, and repulsion are elicited (Blow et al., 2006; MacDorman, 2005) and likability drastically drops. It then rapidly recovers to surpass the previous peak in likability when the nonhuman entity becomes virtu- ally indistinguishable from a human (Bartneck et al., 2007). Notwithstanding the increased attention that the uncanny valley has recently received, there has been sparse empirical support for this concept (Bartneck et al., 2007; Blow et al., 2006). The limited evidence substantiating the scale and pace of growth, decline, and subsequent recovery of likability, relative to increasing anthropo- morphism, constrains the utility of this hypothesis.

robotic Technology in Service encounters

The quality of the face-to-face interaction between the service provider and the customer is critical to service organizations, as it may exclusively determine customer satisfaction and repurchase intentions (Solomon et al., 1985). Past researchers have evaluated the quality of these interactions in terms of rapport (Gremler & Gwinner, 2000; Hennig-Thurau et al., 2006) and how well custom- ers have enjoyed interactions with service providers (Gremler & Gwinner, 2000; Rafaeli et al., 2017). An implied requirement for effective service and positive rapport during social interactions is a display of positive emotions from service providers (Rafaeli et al., 2017). If the human customer service representative is replaced by a digital or robotic version, the quality of reciprocal interactions between the robot and the customer needs to be considered. Researchers have examined the role of technology in service encounters, finding that customer perceptions of such service situations vary depending on the presence, or

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 5

absence, of employee rapport (Giebelhausen et al., 2014). Early studies of human–computer interactions have stressed the importance of building rapport between humans and nonhuman entities (Picard, 1997, 2003). In certain service situations, technology effectively enhanced social interactions between service providers and customers. A study of robot deployment in a shopping mall found that robots need to be friendly for customers to feel comfortable during interac- tions (Kanda et al., 2009).

Other researchers have focused on the differences in group perceptions and attitudes toward robotics. Most research in this area has focused on health care robotics for consumer use (often a mobile screen) or on young children’s percep- tions of robots. Conventional wisdom suggests that younger people will have more positive perceptions and will be more accepting of robotics than older people; however, little research has explored this and the conventional wisdom may be fallible. For example, I. H. Kuo et al. (2009) found little difference between middle-aged and elderly people in their perceptions and acceptance of a health care robot, although they did find that males are more likely to accept robots than are females. Another study of owners of household robotic vacuum cleaners found no significant differences between age or gender groups, although the researchers did find that, over time, the perception and acceptance of the robot improved (Fink et al., 2011).

More recent research has found that the proper integration of smart technol- ogy can complement human service providers, eliminating the need to trade-off between service efficiency and effectiveness (Marinova et al., 2017). Smart technology has been defined as tools consisting of information, software, and hardware to facilitate learning from service encounters to coproduce value. Marinova et al. (2017) proposed that learning, facilitated by smart technology, enables both service providers and customers to retain knowledge and enhance service encounters in real time. However, past smart technology research has focused on information systems, but has not investigated robots as a form of smart technology within the context of a service encounter.

Robotics in the hospitality industry. Hospitality businesses are operated using a significant number of entry-level employees, often staffed with part-time and seasonal employees (Baum, 2006). The rapid advancements in robotic technol- ogy are expected to assist operators in mitigating difficulties related to seasonal and entry-level employment, while maximizing labor utilization (C. M. Kuo et al., 2017). Researchers have proposed the benefits of using robots as an alter- native to low-skilled employees, since robots are easier to train, are capable of providing more reliable and consistent levels of service, and do not lose interest in mundane and repetitive tasks (C. M. Kuo et al., 2017; Qureshi & Sajjad, 2017). However, the restaurant industry is expected to remain highly labor- intensive, regardless of the amount of technology implemented (National Restaurant Association [NRA], 2019). The U.S. labor force participation rate is expected to increase modestly in the decade between 2020 and 2030 (NRA, 2019). The biggest gains in labor force participation are expected to be adults

6 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

aged 65 years and older, while teenagers entering the workforce are expected to decline (Abrams & Gebeloff, 2018). Therefore, the QSR industry must address the potential for a chronic shortage of its traditional entry-level employees.

An exploratory qualitative investigation of robots in hospitality found that hoteliers value obtaining customer feedback while the guest is still physically present on the property, specifically using robots to collect guest feedback (Chung & Cakmak, 2018). One participant in Chung and Cakmak’s study envi- sioned that robots could increase direct guest feedback, as the robot would act as an effective neutral agent between the guest and the hotel. To avoid interper- sonal confrontation or hurting an employee’s feelings, guests may be reluctant to share negative experiences directly with employees; therefore, the robot might provide a neutral avenue to share these experiences. The hotel industry has begun to incorporate service robots into some locations. The best-known examples at this time are the Botlr delivery robot and the Pepper robotic con- cierge. Little data has been published about the actual cost savings or operational benefits of using one of these devices, although there is a great deal of publicity in the popular and trade press about the novelty and delight that these robots generate among hotel guests.

If a robot’s ultimate objective is to serve people through providing informa- tion or assisting with physical tasks (Zalama et al., 2014), research on how hos- pitality guests feel about robots is necessary. One recent study has attempted to identify critical components for human–robot interactions in a hospitality set- ting. Tussyadiah and Park (2018) recommended that the design of hotel service robots that appear to be more humanoid must emphasize the design of the “face” of the robot (vs. the body). They also found that robots that are currently designed as “delivery” robots, which do not look humanoid, should emphasize the intel- ligence displayed by the robot.

Robotics in restaurants. Most of the extant research on robots in hospitality has been conducted in a hotel setting, focusing on concierge and delivery robots. In contrast, very little published research has been conducted that examines the application of robotics to the restaurant industry. A few studies have explored the benefits to restaurateurs when deploying robots (C. M. Kuo et al., 2017; Zalama et al., 2014), including financial benefits such as reduced labor and training costs and operational benefits such as improved quality control and consistency.

The restaurant industry is experimenting with automation in an attempt to provide novelty for customers and to cut costs. For example, the Flippy ham- burger-cooking robot costs approximately $60,000 to purchase, plus the cost to maintain the machine (Bishop, 2019). However, many industrial service robots, such as robotic vacuums and robotic assembly arms, are leased. Purchasing a commercial robotic vacuum will cost between $7,000 and $15,000, but they are often leased at $4 to $6 per hour of operating time; the manufacturer or distribu- tor is responsible for all maintenance on the device (Ackerman, 2014). The actual costs and financial benefits are not published, as they are proprietary

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 7

information for the manufacturer or dealer. However, if the hourly wage rate at a base cost of $15.00/hour plus fringe benefits is compared with an hourly lease rate of $6.00/hour for a commercial robotic vacuum, it is likely that cost effi- ciencies are available as long as the technology works as promised. Costs for these devices is expected to continue to decline.

As the cost of robotics has decreased, some restaurants have already substi- tuted humans with robots for certain jobs (Bowen & Morosan, 2018). A study of the use of robots in restaurants found adoption in the front-of-house (FOH) in the form of tablets, for order taking and payment, and back-of-house (BOH) robots acting as chefs and engaging in cleaning functions (Ivanov et al., 2017). Robots can potentially cook hundreds of different dishes (Yu et al., 2012), with some QSRs currently using robots to prepare and cook hamburgers, limiting the human participation in the process to simply providing the finishing touches (Graham, 2018).

While the operational and financial benefits of using robotics in restaurants have received some attention, scarce research has been conducted that explores the guest’s perspectives on the benefits of a restaurant using robotics. There is scant evidence for research that examines the guest’s perspectives on the service process, the quality of the service, communication experiences while interacting with the technology, and broader societal implications. The current study explores how the use of robotics might benefit the customer or detract from his or her experience.

Purpose of the Study

This qualitative study explored the positive and negative aspects of using robotics in QSRs, from the guest’s perspective. Rather than using well-estab- lished quantitative measurement items, this qualitative technique used a phe- nomenological approach to explore the restaurant guest’s perspectives on using robotics without placing pre-determined boundaries on the range of the partici- pants’ responses. This zero-based approach (Creswell, 2015) was intended to elicit the top-of-mind impressions, concerns, advantages, and disadvantages of incorporating robotic technology in QSRs. Much of the extant literature on robotics, in both the popular press and scholarly hospitality publications, has focused on their use in hotels; less attention has been paid to their deployment in restaurants.

Very few “mom and pop” restaurants have the financial foundation to incor- porate robotics into their operations at this time and are not likely to be a signifi- cant source of technology innovation in the near future (NRA, 2019). However, the larger brands/chains are well-positioned to support incorporating robotics into their operations (NRA, 2019). Restaurants, and QSRs in particular, are exemplified by several operating characteristics that provide a fertile environ- ment for deploying robotics. First, QSR employees are engaged in highly repeti- tive tasks, which have been identified as an ideal opportunity for robotic

8 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

adaptation, as “more facets of cooking will be organized to be readable by arti- ficial intelligence” (NRA, 2019, slide 39). Most QSRs offer a highly standard- ized product and generate high-volume production (Byrd, 2015), which are also ideal attributes for the efficiencies offered by robotic technology. The QSR industry experiences a very high employee turnover rate (Abrams & Gebeloff, 2018), which has been identified as a factor that enhances the feasibility of deploying robotic technology. Moreover, when an economy experiences periods of low unemployment, a labor shortage often occurs which leads to even higher rates of employee turnover (Patton, 2019), further increasing the attractiveness of robotic technology. Evidence also shows that QSRs are increasing their hours of operation, where many locations now operate 24 hours/day, increasing their staffing needs in an already tight market. Finally, there is a well-docu- mented social and political movement in the United States to increase the mini- mum wage rate (“the fight for $15”) along with increased efforts to unionize QSR employees. An environment of rising wages provides a compelling ratio- nale for QSR owners to consider incorporating robotics into their operations (McFarland, 2016).

MeTHod

This study used a phenomenological approach to explore QSR customers’ perceptions and attitudes regarding the use of robotics in these restaurants. Specifically, this is an exploratory sequential mixed methods study; in this article, the results of the qualitative phase of the process are presented. Exploratory sequential studies are typically designed as follows: (1) the research question is formed; (2) qualitative data are collected to answer “how” or “what” questions (but not “why” questions, which require quantitative data); (3) the qualitative data are analyzed to uncover major themes; (4) the themes are used to develop questions that can be used as variables/measures in a quantitative technique; (5) quantitative data are collected and analyzed to establish reliability and validity; and (6) the validated data are examined to determine if the original questions can be explained (Creswell, 2015). The cur- rent article covers Steps 1 through 3, the qualitative phase, with recommenda- tions for the quantitative phase.

The customers’ perceptions and attitudes were obtained through a series of three focus groups (10 participants per group) that were held in a large metro- politan area in the Southwestern United States. The focus group discussion for- mat elicits not only the participants’ initial thoughts on the discussion topic but also provides a free-flowing discussion that uncovers new ideas and perspec- tives, changes opinions, uncovers conditional opinions, and permits the researcher to uncover “deeper truths” related to particularly emotional responses (Mariampolski, 2001). Ultimately, the technique’s results should reveal multiple aspects of a phenomenon, which then permits identification of specific themes for deeper exploration.

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 9

Participant recruitment and Sample characteristics

Study participants were recruited through on-campus recruiting at a local university, as well as through intercepts at a shopping mall approximately one mile away from the university. Individuals 18 years of age or older who self- identified as eating at a QSR (a.k.a., “fast food restaurant”) at least once during the previous month were eligible to participate in the study.

Participant demographic data were not formally collected, as the intent of this qualitative phase of a larger mixed methods exploratory study was only to iden- tify the phenomena surrounding the use of robotics, rather than to attempt to generalize the results to the greater population (Creswell, 2015; Weber, 1990). Subsequent phases of this study will use the qualitative data to develop a quan- titative measurement instrument that will be used to collect data from a larger sample that will permit generalizability to a larger population.

The first focus group consisted of participants who were recruited through the mall intercepts; these participants ranged in age from 25 to 70 years. The second and third groups consisted of college students who ranged in age between 20 and 29 years, with two students who self-describe as “nontraditional” (i.e., older than 30 years) students. The student participants were invited to participate in this study at the end of a class period; none of the students were current stu- dents of any member of the research team. All participants were recruited with the understanding that they would receive a $50 Amazon gift card at the end of the focus group session in consideration of their time.

Of the 30 participants, 16 were female and 14 were male. A total of 20 under- graduate students, 2 graduate students, and 8 nonstudents participated in the study. Twelve participants were categorized as “older,” which here means 29 years of age or older, and the remaining 18 students were younger than 29 years. Creswell (2015) recommends a sample size of between 3 and 10 participants for a phenomenological study, and this sample size exceeds that recommendation. Focus group recruitment often requires overrecruitment of participants, since it is common for some of the recruits to fail to show up for the focus group. In addition, while the intent of the study was to examine restaurant customer per- ceptions of robots, many of the participants reported during the focus group discussions that they either currently or had previously worked in the foodser- vice industry. This is typical of the U.S. population, where nearly 50% of all adult Americans have at one time or another worked in the foodservice industry (NRA, 2019; The Aspen Institute, 2013).

data collection

The focus groups were held in May 2018. Each focus group lasted between 60 and 75 minutes in meeting rooms on the university’s campus, providing a neutral (nonrestaurant) environment. The groups were led by an experienced moderator, using a semistructured interview protocol, which is displayed in Supplement

10 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

Figure 1 (available online). Interestingly, all study participants reported being familiar with robots used in hospitality industry settings, and nearly half of the participants report having first-hand experience with them.

The focus groups were also observed by members of the research team, who took notes of the proceedings for future reference. The discussions lasted until data saturation was reached, where each topic was exhausted and further dis- cussion would not yield additional insight (Miller & Fredericks, 1999). The discussions were digitally-recorded (audio only), and were later professionally transcribed. On completion of the focus group session, each participant was provided a $50 Amazon gift card, as promised at the time of recruitment.

analySiS and reSulTS

The focus group transcripts were analyzed manually using content analysis, via a grounded approach involving a series of immersion/crystallization cycles (Miller & Fredericks, 1999). This emergent, inductive process follows a cyclical approach where the data are reviewed, the researcher mulls over it, the categori- zation of data is reviewed and revised, and finally a set of initial codes or themes emerges (Saldaña, 2016). The research team then coded the transcripts for rela- tionships with each theme, followed by valence analysis (a form of sentiment analysis) to assess the focus group participants’ sense of whether the use of robotics in QSRs is positive, negative, or both. A summary of the results of the analysis is presented, and a full description of the results is provided in the next section. A flowchart illustrating the steps in this process is provided in Figure 1.

Stage 1: identifying Themes

First, four trained raters used a team coding approach (Saldaña, 2016). Each team member read through the transcripts independently numerous times to identify the overarching themes. Each team member created a set of written theme descriptions. The team then met to discuss the themes to clarify and generally reach agreement. Each description was transferred to a slip of paper which was color coded to reflect the team member who generated the descrip- tion; the team members organized each individual slip of paper into cohesive groups on a sheet of flipchart paper (Supplement Figure 2, available online). This method yielded the following 19 themes: customer experience, robot tasks—BOH, robot tasks—FOH, physical appearance, human touch, labor force impact, communication, novelty, benefits—efficiency, benefits—quality control, cost control, safety—robot movement, safety—sanitation, safety— food, safety—robot sanitation, motivations to visit or repatronize, trust/distrust, flexibility, and societal change.

In the next step, a random sample of 30 fragments (statements) from the tran- scripts was entered into a spreadsheet, where each comment was assigned a case number and occupied its own row. The fragments were randomized to avoid order bias. The rating was a simple binary code, where the fragment received a

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 11

“1” if it related to the theme, and a “0” if it did not relate to the theme. Each rater then independently rated each fragment on whether it was related to each of the 19 themes. The results were analyzed to determine if the raters generally agreed on each fragment’s relationship with each theme (Weber, 1990). Where there was disagreement, the raters then discussed their rationale for their rating of the item. Furthermore, the discussion led to the raters agreeing that one additional theme had emerged—Service Speed. The final set of 20 themes and their descrip- tions is shown in Table 1.

Frequency analysis. Once all team members completed their evaluations of the transcripts, the results were combined and frequency analysis was conducted using Excel. Cases that at least two out of the four team members coded as being related to the relevant theme were included. A total of 406 speech fragments were analyzed for their relationship to each of the 20 themes. The results of this analysis are shown in Table 2.

When the analysis yields a large number of themes, Creswell (2015) recom- mends focusing subsequent analysis on a smaller subset of themes, usually lim- ited to five or six themes. The coding team discussed the results of the frequency analysis, and reached consensus on selecting nine themes for subsequent analy- sis (Table 3). However, the nine selected were not automatically drawn from the themes with the highest frequencies. Two themes—Customer Experience and Benefits—were rejected for further analysis and the team identified three sub- themes related to Safety that were combined into one overarching theme, as explained next.

Figure 1 Flow chart of Methods

12 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

Table 1 initial Themes identified Through content analysis

Theme Description

Customer experience Elements of the customer experience. Robot tasks—BOH Tasks that robots might perform in the back-of-house

(BOH) areas, such as food preparation or cleaning the kitchen.

Robot tasks—FOH Tasks that robots might perform in the front-of-house (FOH) areas, such as delivering food to guests or cleaning tables.

Physical appearance The physical attributes of the robot, such as shape, material, finishes, and attachments.

Human touch The special quality that humans bring to the process, service or product.

Labor force impact The impact that robots would have on labor; most often in terms of reducing or changing employment.

Communication The experience that participants have with communicating with robots or humans; often language-based.

Novelty The unique qualities that a robot would bring to a restaurant; often related to desire to visit restaurant at least once to “check out the robot.”

Benefits—efficiency The benefits that robots would bring in terms of efficiency, such as reducing waste, reducing incorrect orders, and so on.

Benefits—quality control

The benefits that robots would bring in terms of maintaining consistent quality standards.

Cost control Generally, the use of robotics to save money, particularly in terms of labor costs.

Safety—robot movement

The safety of the robot’s movement; particularly in terms of movement of the entire unit, robotic arms, and whether or not these pose a threat to safety.

Safety—sanitation This refers to the sanitation of the restaurant itself. Safety—food This refers to food sanitation and safety. Safety—robot

sanitation This refers to how well the robot itself can be cleaned and

sanitized. Motivations to visit or

repatronize Related to the “novelty” theme, but more broadly

expressed as being motivated (or not) to visit to satisfy curiosity, get more accurate order, receive faster service, and to not have to interact with a human.

Service speed Mentioned if robotics were likely to increase the speed of service, although occasionally the robot may slow the service.

Trust/distrust Whether or not the participant trusted robotics to perform properly or deliver anticipated results.

Flexibility Whether a robot could perform different tasks or if it could be easily reprogrammed.

Societal change The effect of robotics on greater societal trends.

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 13

Table 2 Frequencies

Themea Total Percentage

Robot tasks—BOH 262 13.0 Customer experience 252 12.5 Robot tasks—FOH 226 11.2 Physical appearance 170 8.4 Human touch 140 6.9 Labor force impact 110 5.4 Communication 108 5.3 Novelty 106 5.2 Safety—food 94 4.7 Benefits—efficiency 85 4.2 Cost control 70 3.5 Safety—robot movement 69 3.4 Safety—sanitation 53 2.6 Service speed 52 2.6 Motivations to visit or repatronize 50 2.5 Benefits—quality control 46 2.3 Safety—robot sanitation 43 2.1 Trust/distrust 32 1.6 Flexibility 27 1.3 Societal change 26 1.3 Total frequencies 2,021

Note. A total of 406 fragments were analyzed. aFragments with 2 or more rater scores.

Table 3 Mean valence for nine Themes

One-Sample Statistics

N M SD SE

Communication 99 0.0253 1.10345 .11090 Human touch 164 −0.3999 1.48123 .11566 Labor impact 158 −0.5105 1.74421 .13876 Novelty 66 1.2412 1.23041 .15145 Physical appearance 87 −0.0326 0.91370 .09796 Restaurant safety 69 −0.2041 1.49970 .18054 Robot safety 406 −0.0782 0.41225 .02046 Task—BOH 216 0.6323 1.14168 .07768 Task—FOH 211 0.4364 1.21583 .08370

14 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

While the theme that received the most mentions was Customer Experience, it was rejected for further analysis. This construct was the focus of the study and the term was used heavily in participant recruitment and as prompts for the focus group discussion. In addition, a visual inspection of the data revealed that Customer Experience almost always corresponded to a more explicit theme (e.g., a comment was rated as being related to both Customer Experience and Robot—FOH Task). Thus, since the use of Customer Experience almost cer- tainly influenced the frequency of participant references to it, the team set Customer Experience aside and did not use it for further analysis.

The team also reviewed the two Benefits subthemes of efficiency and quality control. The Benefits constructs were removed from further analysis. The first reason for removal was because virtually every comment regarding Benefits was highly positive (as benefits would naturally be), and no additional insight was offered by the participants. For example, when asked about the benefits of robot- ics, the responses were usually limited to one or two words—efficiency and/or quality control; the participants did not elaborate and a mere mention of the word “efficiency” did not yield deeper insight. The second reason for removing Benefits from further analysis was that the focus group discussion included fre- quent prompting regarding the benefits of robotics. Similar to the rational for removing Customer Experience, the prompting generated frequent mentions of the benefits overall, but the resulting comments did not offer deeper insight.

Finally, the Safety theme was initially broken into three subthemes; Safety- food, Safety-robot movement, and Safety-sanitation. However, while each Safety subtheme yielded rated fragments, the frequencies were relatively low in num- ber. As Weber (1990) and Saldaña (2016) discuss, using human coders, particu- larly when they are also present during data collection, will “impose the reality of the investigator on the text” (Weber, 1990, p. 37). Each of the coding team members had attended one or more of the focus group sessions. The team mem- bers reviewed their notes taken while observing the focus groups, including recording nonverbal communications. Their recollections showed that the focus group participants exhibited high intensity and emotion when discussing the Safety themes, despite the fact that each of Safety’s subthemes yielded relatively low frequencies of fragments. Ultimately, the team combined two subthemes— Safety-sanitation and Safety-food, thus creating a new theme labeled Restaurant Safety. The robot movement subtheme was not related to food safety or sanita- tion, and was thus relabeled Robot Safety, describing the robot’s safety (from harm) in the environment and the safeness/unsafeness of the robot’s movements in the environment.

Stage 2: valence analysis

The nine themes identified in the previous step—Communication, Human Touch, Labor Impact, Novelty, Physical Appearance, Restaurant Safety, Robot Safety, Task—BOH, and Task—FOH, were then explored using valence analysis.

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 15

Valence analysis is a technique that is used to determine the level of positive or negative attitudes associated with a construct or, here, a speech fragment (Colombetti, 2005). This study used a form of magnitude coding (Saldaña, 2016) to determine the level of overall positiveness or negativeness attached to each fragment.

An initial coding scheme ranging from −3 to +3, where a −3 was very nega- tive, a +3 was very positive, and a 0 was neutral, was used (Barrett, 2005; Colombetti, 2005; Mazurenko et al., 2015). This permitted the research team to assess the relative positiveness, negativeness, or neutrality of each fragment. The goal was to triangulate the raters’ results. A trial sample was again con- ducted to clarify the coding process. After reviewing the coding trial’s results, the research team determined that further clarification was not required. The full set of randomized transcript fragments was then analyzed for valence. Each fragment was entered into an Excel spreadsheet where each column contained one of the nine themes. Each team member read each fragment and if the frag- ment was related to a theme, the coder entered his or her valence rating. If the fragment was not related to the theme, the coder left the cell blank. The mean valence for each theme was calculated, as presented in Table 3.

Once each team member completed his/her valence coding, the results were combined into a master sheet in Excel, then transferred to SPSS. The valence coding for each theme was then analyzed for interrater agreement across all four raters, using intraclass correlation analysis (Shrout & Fleiss, 1979). Intraclass correlation is an appropriate reliability test because each “target,” or transcript fragment, is analyzed by a fixed set of judges (the four raters, and not a random sample of a larger population of raters), thus creating a fixed effect. Here, the individual valence ratings of each fragment for each judge were entered into the analysis. The results are displayed for Cronbach’s alpha, which is an appropriate measure for inter-rater reliability in this situation (Shrout & Fleiss, 1979). Intraclass correlation is shown for single (individual rater) measures as well as for the average rating across all four judges. The results for interrater agreement correlations are displayed in Table 4.

For each theme, the single measure intraclass correlation was lower than the correlation for the average measures. This was expected, where there was more variability for a single rater, but when all raters’ values were combined, some of the variability was smoothed out and the correlations were higher. The Cronbach’s alpha values were between .70 and .80 for five of the themes, con- sidered “good”; all were statistically significant (p < .01). Two additional themes had Cronbach’s alpha values of .665 and .691; these do not reach the “good” level, but both are statistically significant (p < .01). The remaining two themes—Novelty and Human Touch—had low Cronbach’s alpha values of .304 and .339, respectively; neither was statistically significant. Although these results indicate that there was not perfect agreement between the raters’ assess- ments of valence of all themes, the process was successful overall in obtaining a reasonable level of agreement. As illustrated by the small number of degrees of

16

T ab

le 4

in tr

ac la

ss c

o rr

el at

io n

Th em

e C

ro nb

ac h’

s A

lp ha

In tr

ac la

ss

C or

re la

tio na

95 %

C on

fid en

ce In

te rv

al F

Te st

W ith

T ru

e V

al ue

0 F

Te st

W

ith S

ig Lo

w er

B ou

nd U

p p

er B

ou nd

V al

ue d

f1 d

f2

C om

m un

ic at

io n

.7 21

S in

gl e

m ea

su re

s .3

93 b

.1 28

.6 92

3. 58

8 13

39 .0

01

A ve

ra ge

m ea

su re

s .7

21 c

.3 71

.9 00

3. 58

8 13

39 .0

01 H

um an

t ou

ch .3

39

S

in gl

e M

ea su

re s

.1 14

b −

.0 59

.3 71

1. 51

2 20

60 .1

11

A ve

ra ge

m ea

su re

s .3

39 c

− .2

86 .7

03 1.

51 2

20 60

.1 11

La b

or im

p ac

t .7

38

S

in gl

e m

ea su

re s

.4 14

b .1

78 .6

72 3.

82 4

17 51

.0 00

A

ve ra

ge m

ea su

re s

.7 38

c .4

64 .8

91 3.

82 4

17 51

.0 00

N ov

el ty

.3 04

S in

gl e

m ea

su re

s .0

99 b

− .1

00 .4

29 1.

43 7

13 39

.1 86

A

ve ra

ge m

ea su

re s

.3 04

c −

.5 71

.7 50

1. 43

7 13

39 .1

86 P

hy si

ca l a

p p

ea ra

nc e

.6 65

S in

gl e

m ea

su re

s .3

32 b

.1 53

.5 38

2. 98

7 30

90 .0

00

A ve

ra ge

m ea

su re

s .6

65 c

.4 20

.8 23

2. 98

7 30

90 .0

00

(c on tin ue d)

17

Th em

e C

ro nb

ac h’

s A

lp ha

In tr

ac la

ss

C or

re la

tio na

95 %

C on

fid en

ce In

te rv

al F

Te st

W ith

T ru

e V

al ue

0 F

Te st

W

ith S

ig Lo

w er

B ou

nd U

p p

er B

ou nd

V al

ue d

f1 d

f2

R es

ta ur

an t

sa fe

ty .7

91

S

in gl

e m

ea su

re s

.4 87

b .1

98 .7

76 4.

79 1

11 33

.0 00

A

ve ra

ge m

ea su

re s

.7 91

c .4

97 .9

33 4.

79 1

11 33

.0 00

R ob

ot s

af et

y .6

91

S

in gl

e m

ea su

re s

.3 58

b .3

06 .4

12 3.

23 1

40 4

12 12

.0 00

A

ve ra

ge m

ea su

re s

.6 91

c .6

38 .7

37 3.

23 1

40 4

12 12

.0 00

Ta sk

— B

O H

.7 90

S in

gl e

m ea

su re

s .4

84 b

.3 55

.6 14

4. 75

5 60

18 0

.0 00

A

ve ra

ge m

ea su

re s

.7 90

c .6

88 .8

64 4.

75 5

60 18

0 .0

00 Ta

sk —

FO H

.7 68

S in

gl e

m ea

su re

s .4

52 b

.2 68

.6 44

4. 30

5 29

87 .0

00

A ve

ra ge

m ea

su re

s .7

68 c

.5 94

.8 79

4. 30

5 29

87 .0

00

N ot

e: T

w o-

w ay

m ix

ed e

ff ec

t m

od el

w he

re p

eo p

le e

ff ec

ts a

re r

an d

om a

nd m

ea su

re s

ef fe

ct s

ar e

fix ed

. a T

yp e

C in

tr ac

la ss

c or

re la

tio n

co ef

fic ie

nt s

us in

g a

co ns

is te

nc y

d ef

in iti

on . T

he b

et w

ee n-

m ea

su re

s va

ria nc

e is

e xc

lu d

ed fr

om t

he d

en om

in at

or v

ar ia

nc e.

b Th

e es

tim at

or is

t he

s am

e, w

he th

er t

he in

te ra

ct io

n ef

fe ct

is p

re se

nt o

r no

t. c

Th is

e st

im at

e is

c om

p ut

ed a

ss um

in g

th e

in te

ra ct

io n

ef fe

ct is

a b

se nt

, b ec

au se

it is

n ot

es

tim ab

le o

th er

w is

e.

T ab

le 4

. (c

o nt

in ue

d )

18 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

freedom, the limited and fixed number of raters, combined with a relatively small set of fragments that aligned with many of the themes, could result in a rough but relatively robust assessment of reliability for qualitative items.

Stage 3: interaction Between Frequency and valence

Strictly relying on either frequency or valence provides limited insight (Colombetti, 2005). For example, some fragments that received frequent theme associations yield common sense results (e.g., a discussion topic of “customer experience” yielded numerous statements related to the customer’s experience). Alternatively, some fragments may receive a highly positive or negative valence value, but are related to topics that are infrequently mentioned. The comment may simply be an outlier, or it may suggest that the issue merits further discus- sion. Each of the nine themes is discussed in detail in the next section of this article.

diScuSSion

This study elicited diner perspectives on a variety of issues related to the use of robotic technology in QSR operations. Major findings of this study indicated that guests have concerns regarding the societal impact of robotics entering the realm of QSR operations; the cleanliness and food safety of robotic technology; and the quality of communication, especially voice recognition, from both native and non-native English speakers. Diners also voiced compelling ideas about the functionality and physical appearance of robots; expressed strong bipolar com- ments relating to the value of the “human touch” in fast food restaurants; and offered creative solutions for the deployment of this technology.

communication

The Communication theme received numerous mentions (99) with a slightly positive valence of 0.0253. The focus group participants generally approved of using robots for basic communication functions, such as touchpads or kiosks for placing orders. However, these devices do not constitute robotics, as they do not perform physical tasks. The participants also discussed verbally interacting with devices. As the nearly neutral valence shows, the participants were divided in their thoughts on communication with robots. Some mentioned it as a positive, as a way to avoid having to interact with other people. As one participant (male, mid-60s) stated,

I have a problem hearing, like in a crowded restaurant . . . yesterday, at [restaurant name], I tried to put in [an order]. First of all, I come up and she goes “gibberish.” “Ah, I’m sorry, what?” “Ah, welcome to [restaurant]. Can I help you today?” Well, you know, I mean, she had to slow down and say it louder. And then . . . I told her my order and then she said something else that I couldn’t hear . . . so personally,

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 19

I’d rather not interact with people. If . . . a robot could take my order the first time and make it faster . . . and smoother, I would be interested in that. Easier.

Another positive aspect of Communication was the potential to offer communi- cation options in virtually any language. One participant (female, early-20s) suggested that the potential capability of robots to understand foreign languages could be a major advantage when communicating with customers, “But what if it knows all the languages? Like what if it knows your natural tongue?”

On the other hand, participants also viewed their past communication expe- riences in a negative light, primarily driven by deficiencies in the current technology. The main concern was related to deficiencies in voice recognition technology. While some of the participants who were non-native English speak- ers expressed frustration with not being understood by English language-only systems, voice recognition was a concern for both the nonnative and the native English speakers. For example, one participant (male, mid-20s) stated, “ . . . but to be fair, Siri doesn’t understand me when I speak, like, Californian.” There was consensus among participants that the technology should not be incorporated until the communication technology is perfected, as deficiencies in the technol- ogy will lead to user frustration and decreased customer satisfaction.

Human Touch

Strong bipolar sentiments arose regarding the value of providing the “human touch” versus not wanting to interact with humans in a QSR setting. In this con- text, the “human touch” refers to the special quality that humans bring to the process, service, or product. One participant (female, early-20s) indicated that most robots she has encountered in service situations are missing the human touch: “And a lot of the complaints that they’ll get, no matter where they are in the world . . . a lot of people feel like it’s too impersonal. And I know it’s a robot, but it looks like me, and I don’t like that.”

Some participants were in favor of freely implementing robots throughout the BOH areas. As one participant (male, mid-60s) commented, “They should do everything, no humans in there,” although one area of caution involved con- cerns about robots handling knives. On the other hand, others mentioned the importance of having their meals prepared by human cooks and that customer complaints, or “escalations,” should be handled by humans, not by robots.

Participants also indicated that they would not miss the human touch in a QSR but required human touch in other restaurant segments. As one participant (male, late-60s) stated,

I have no objection to it, and I’m just talking about fast food, I think if I went to a bar, I’d rather talk to a human bartender because usually you get to have a conversation with them and that’s a big, a good part about the experience. So I’m not sure a robot could do that. Maybe it can.

20 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

Others stated that customers would miss the human interaction, “Eventually people will feel like its missing the part of the connection, people-to-people.”

As the participants actively deliberated the value and role of the “human touch” in QSRs, they offered creative solutions for the use of information tech- nology, such as facial recognition for regular customers. For example, one par- ticipant (male, mid-30s) suggested that the robot could greet the customer at the door and say,

“Hi, [name], how are you today? How was work today? Are you gonna come in and have your Number 1 (combination meal)?” . . . that would really be amazing. Like, have face recognition software. They could just, “Hi, [name], do you want your Number 1? You want your fries, right?” And you’re like, “Yeah.” By the time you get to the counter, it’s already waiting for you.”

Another participant (male, mid-60s) piggybacked on this comment by adding, “It should take your order in the parking lot, so when you walk in the door, it’s already at the table.”

labor and Societal impacts

One theme—Societal Change—was mentioned relatively infrequently and was not selected for valence analysis. Another theme—Labor Impact—was mentioned frequently and displayed a negative valence. Both themes are dis- cussed here because they are closely related from a conceptual basis. The par- ticipants suggested an inevitability about the deployment of robotic technology. Participants pointed out that QSRs are the first employer for many people, leading to concerns regarding adverse societal effects arising from the use of robots. Many participants were concerned that an increased presence of robotic technology in QSRs will lead to a decline in the employment of young adults as well as to the subsequent negative effects on society. For example, one par- ticipant (male, mid-20s) voiced the following concern, “(I) do not like the idea, simply because it’s going to hurt the economy as we reduce the basic labor force.” Another participant (male, mid-60s) proposed that “It’s a bit of a philosophical issue. And that is, are we replacing humans for automation?” While some participants held negative sentiments regarding this chain of effects, others acknowledged that such changes in society are not uncommon and that over time, major changes are inevitable. One participant (female, early-20s) stated,

And actually, even if we accommodate robotic systems at fast food restaurants, we aren’t really taking jobs from people because it’s like, energy is not created but it’s only transformed into different forms of energy. It’s the same with jobs. There will be some other jobs (that will) appear and those will be available to the workers at fast food restaurants.

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 21

novelty

Generally speaking, focus group participants had very positive impressions of the ability of the Novelty of robotic technology to entice guests to visit the location at least once, although they were unsure whether the robotics would sufficiently overcome average food or service to entice them to return to the restaurant a second time. This is consistent with past examples of restaurant concepts that provided a highly unique experience but suffered from the reputa- tion that the guest would visit once because of the “experience” but would not return because the food was too expensive and/or the food quality or service provided a poor value overall.

Physical appearance

The Physical Appearance theme was mentioned frequently but had a very slightly negative valence. Participants stressed that robots should not appear menacing or scary (e.g., “ . . . there’s a lot of people that are scared of clowns”), but rather held strong preferences for connecting a robot’s physical appearance to its intended functionality. Examples include “I don’t think they should look humanoid. I think that’d be weird” (female, late-20s), and “ . . . they should look like they have a purpose . . . designed for a specific purpose that they’re intended to do . . . utilitarian” (male, mid-20s). As one participant (male, late-60s) stated in particular, “I don’t like the humanoid ones very much. I think if they’re robots, they should be good robots and not try to be humans.” These sentiments are explained, in part, by disambiguation. Participants in this study had a preference for robots to appear plain, dull, and uninteresting rather than take the form of a humanoid, corresponding to the section of the likeability graph that appears just prior to the uncanny valley (Mori, 1970). Functionally, the robots should have “no sharp edges,” “smooth surfaces,” “ability to switch tools,” and should appear “not tacky.”

restaurant Safety

Strong concerns about restaurant cleanliness and food safety arose through- out the discussions. Several participants commented on the benefits of using robotics to improve food safety, citing the opportunity to have robots that can check food temperature, have sensors to check the quality of the ingredients, and to remove some of the risk of human contamination. Interestingly, multi- ple participants in all three groups mentioned the benefit of not worrying about a robot spitting in the food. For example, one participant (female, early-20s) stated that

Like, for example, I know some people—I’m not saying all people do that, when there’s a personal feeling involved, some people mess the food up that’s supposed

22 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

to go to the customer. Like, for example, they’ll spit on the food and no one will know. (The) customer will not know, but if you have (a) robot that cooks for people, you will probably not . . . have that kind of issue.

However, many participants also raised concerns about robots creating food safety problems. In particular, the participants were concerned a robot that pre- pares food items not being able to identify if food ingredients were safe for consumption. One participant (female, early-20s) expressed this concern: “ . . . some of the food gets, like, bad and then the robot couldn’t recognize that and just cook it. And it’s like there’s mold on it and they don’t know, cook it.” There were also concerns about the integrity of the robot’s components and ensuring that pieces of the robot did not contaminate the food product. For example, one participant (male, mid-60s) described his past experiences with finding foreign objects in his food, although he then stated that “I’d rather have a nut or bolt (in the food) than, say, a couple of cockroaches that I once found in a restaurant (in town).”

Many of the comments focused on the ability to keep the robot clean to avoid contaminating food and service items. For example, one participant (male, mid- 30s) stated, “They should have rust prevention. If a robot is a dishwasher, do not let it rust. Well, the robots have to be cleaned. There’s gonna be build-ups.” Therefore, if robots are in direct contact with food, it should be possible to clean and disinfect the robots properly. One participant also mentioned her concerns about air-borne oil and grease from the grill and deep fryers. The airborne oils settle on all surfaces and can be difficult to clean, according to one participant (female, late-20s):

. . . like inside the machine somehow. It’ll fly in there and it’ll somehow get to, like, anywhere (in) the whole robot. And what if they’re not getting cleaned, like, for the whole week? Am I gonna eat the fries or burger . . . after they touched it? I don’t know how long that oil’s been there, staying on the body. And when it’s working in a kitchen, it’s always hot. . . . And the oil might melt on them and just drop back to the food.

There are some measures already available in foodservice industry to ensure the hygienic design of such robots. For example, “wash-down robots” feature a sanitary design with a smooth surface to prevent foreign substances from remaining on the robot arm, and they can be easily sanitized during the antisep- tic cleaning process. They can resist harsh chemicals and intense water pressure, making it easy to maintain proper cleanliness for food and other applications that require wash down capabilities (Edelbrock, 2012). Some manufacturers have opted for protective coatings such as epoxide; and others have made their robots entirely out of stainless steel that does not react with cleaning agents, acids, or alkalis. It is also suggested that the lubricants used on these robots be food-grade certified (NSF H1).

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 23

robot Safety

The potential dangers posed by using robotics, either from the perspective of the robot physically harming people or being harmed by people, was discussed. A major concern focused on whether robots should wield sharp objects, particu- larly knives. Another concern was whether the robot should be mobile, either moving through the space or having robotic arms that move. As one participant (male, early-20s) explained,

Participant: I don’t think the robot should be moving around in the kitchen. Moderator: Why should they not move around the kitchen? Participant: Sometimes it’s chaos in the kitchen . . . it’s always chaos in the

kitchen, so people moving around with knives and stuff. So, we don’t want to trip over a robot.

One participant (female, late-50s) was concerned about the robot running into guests: “I’d be afraid they’d run over somebody in the front of the house.” Another participant (female, mid-20s) stated,

. . . if they’re small or if they’re big, somebody will find something to complain about. Or to sue them about. Like, “oh, this robot . . . ,” “I tripped over this robot because it was small,” or “I ran into this robot and busted my head open because it was too big.” I just feel like . . . if it’s moving around, it will be an issue some way or another.

Another participant suggested that robots should only operate in a designated area to avoid liability problems, or not a “free range robot” (male, early-30s). One recommendation for what amounts to a robotic “train” to deliver food or other items was mentioned in each group. A description from a participant (female, early-20s) is as follows:

The thing is not touching the food and it’s on its own train (track) . . . so they won’t . . . fly out or something. So basically, in the front of house . . . they can move and won’t . . . injure other people. But if it’s a host robot and it’s running (around the dining room), then that might (lead to injury).

Concerns for the robot’s safety were also expressed. As one participant (male, mid-20s) stated, “ . . . I guarantee you, my brother, who’s younger than me, would totally mess with that (robot), and push it to its limits.”

Tasks—BoH

The participants identified numerous BOH tasks that robots could perform. First, participants expressed strong support for tasks that humans do not like to do, such as dishwashing. Participants in all three focus groups repeated that no

24 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

one likes to work in the dishwashing area, due to the exposure to hot water, heat, humidity, and food waste. Robots could perform this function, to the point where one participant suggested that the dishwasher be a robot. While current dish- washers are machines (and may be perceived as robotic devices), the dishwash- ing robot could load itself, clean and sanitize the service items, and then move through the BOH area to put the items away.

Food preparation. Participants also strongly supported the idea of using robots that perform basic food preparation such as cutting vegetables, cutting potatoes for French fries, and making dough. They also supported robots per- forming basic cooking tasks, such as hamburger patties and French fries. The robots could have sensors that check temperature and doneness, or sensors that provide a visual examination of the product and possibly sample air quality around the food item to test if it was still fresh and/or safe to use. However, several participants then suggested that the final finish work should be per- formed by a human, as explained by one participant (male, mid-20s) who stated,

From a guest perspective, I don’t like the idea of my food being—I don’t mind it being prepped, so like vegetable cutting, dough making, all of that. Perfectly okay with that. You’re saving money, you’re saving steps. I don’t like the idea of my food after that point being made by a robot. I like the human element. I like knowing that someone put effort into my food.

Inventory. Other participants suggested using robots to manage inventory, illus- trated by the following exchange between two younger participants:

C (female, late-20s): Inventory actually would probably be a good (idea). . . . If you can program one that’s complex enough to manage inventory, you lose the human lying elements and you keep consistency, which is helpful.

M (male, mid-20s): Yeah, there’s an algorithm about how much in a box equates to what, just by looking at it.

C: Yeah. Or looking at barcodes. So it scans which product it is. Kinda like what they do in big warehouses.

However, the participants did not believe that robots should perform tasks requiring fine attention to detail, such as finishing plating a meal—although this task is less common in QSRs than in fast-casual or more formal dining situ- ations. The participants were also more likely to prefer a human to complete these tasks if the guest interacts with the cooking staff; the belief is that human cooks know the customer and can better customize an order the way the guest likes.

Integrating robots into existing equipment. One participant suggested that maintaining safe temperatures for food product and in coolers could be improved by incorporating a robot that could move materials in and out of a

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 25

cooler without opening doors. The benefit was described by a participant (male, mid-20s),

. . . robotic service to grab ingredients from the fridge where the employee doesn’t have to go through a door, let all that air escape. That should be awesome, because you would save time and save electricity as well.

This recommendation would work well for restaurant companies that desire to enhance their corporate social responsibility initiatives, since it will serve to reduce the environmental footprint. It will also reduce energy costs, which could defray the cost of implementing the robotic technology.

Other tasks. Numerous additional tasks that would be useful in the BOH areas included equipment maintenance, such as “changing out oils in fryers, sharpening knives, sanding down cutting boards. . . . And those are the things that everyone I’ve ever worked with hate doing, because they are very disruptive to your traditional schedule,” according to one participant (male, mid-20s). Other tasks included forming burger patties, scraping the grill between uses, and general kitchen sanitation.

Tasks—FoH

The participants frequently suggested that robotics would be appropriate for FOH tasks such as order-taking, food delivery, and FOH sanitation.

Order-taking. While discussing the opportunities for robots to take food orders, with face-to-face or on the telephone, numerous participants mentioned their concerns about the quality of communication. Strong negative sentiments focused on the accuracy and efficiency of voice recognition technology, as dis- cussed earlier in the Communication theme section.

Finally, while the participants did acknowledge the utility of using robots in order-taking, they unanimously agreed that if a conflict with a guest arises, the situation should be resolved through human intervention. In part, the partici- pants repeated their concerns regarding the current deficiencies in voice recog- nition technology. However, they also focused on their beliefs that until robotic technology can better interpret nonverbal prompts, the use of a robot during a service conflict/escalation would only exacerbate the problem. Regardless of the attractiveness of the physical appearance of the robotic device, the tension that forms when a customer is dissatisfied may erode any initial acceptance of the robotic technology and could exacerbate a negative service experience. Human intervention will be necessary to manage the subtleties of conflict resolution with unhappy customers.

Food delivery. Participants also suggested using robotics for food delivery systems, such as delivering food to a table located next to a wall by sending the food out on a track, similar to a train. The participants preferred that robots not move freely around the restaurant dining room, due to the potential for collisions

26 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

with guests. The participants were also concerned about the potential guests to “tamper with” the robot, such as the risk that small children or immature adults might “mess with it,” as suggested by one participant (male, mid-30s). These creative solutions primarily focused on increasing efficiency of QSRs to subse- quently enhance the guest experience.

Interestingly, some of these suggestions have already been implemented in the foodservice industry. For example, Alibaba recently opened a restaurant in the Hema supermarket in Shanghai, China, which uses a mobile app, QR codes, and robots to provide a high-tech dining experience without losing the “human touch.” Customers interact with employees when they select fresh seafood in the supermarket. At the entrance of the restaurant, the app informs customers where to sit, and the app can be used for customers to order and pay for meals. Human cooks prepare the food and place it inside small pod-like robots. Then the robots travel along tracks to deliver the food directly to customers.

FOH sanitation. The participants also offered thoughts on how robots could be used to clean the FOH areas during service periods. Most of the suggestions involved using robots to clear tables and sweep/mop the floor. However, while these functions come with attendant risks, the participants were very concerned about the potential risks for a robotic device moving through the dining area and creating a trip/fall hazard for customers. A way to incorporate robotics in the FOH areas may include limiting the robot’s activities to one portion of the din- ing area at a time or providing a dining room layout that permits adequate space between tables that will allow traffic circulation for both robots and humans.

Security. The focus group participants almost unanimously agreed that robotic devices should not be used for security functions in QSRs. While secu- rity robots are increasingly common in larger public spaces, such as airports, this study’s participants thought a security robot “would freak people out,” accord- ing to a younger female participant and an older male participant. There would likely also be highly negative reactions from members of the public, who may have privacy concerns.

unexpected results

Some unexpected results arose during the focus group discussions. The research team expected differences in perceptions and expectations of robot deployment in QSRs across age groups. Surprisingly, all participants, regardless of age, exhibited approximately equal parts of enthusiasm and skepticism with respect to robotic technology. A large percentage of participants with past or current restaurant work experience added an unexpected dimension to the discussions, as participants were able to relate how the inclusion of robots into QSR operations would affect both employees and guests. Last, there was a high level of resignation about the inevitability of QSRs incorporating robots. This finding is similar to the accept- ability of routine societal change. Participants felt that the incorporation of robotic technology is a question of when, rather than a question of if.

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 27

contributions of the study

This study’s results contribute to the extant literature by enhancing the cur- rent body of knowledge regarding multiple facets of the QSR customer’s per- ceptions of both FOH and BOH applications of robotic technology. In addition, this study revealed several factors that robotics manufacturers and QSR opera- tors should consider when designing and/or implementing this technology. While the study focused on the QSR industry, many of the positive and negative attitudes can easily translate to the potential incorporation of robotic technology in other service industries.

contributions to the extant literature

This study makes three contributions to the extant literature. First, the current study adds to previous research by addressing restaurant customers’ points of view. Previous research concentrated mainly on the benefits realized by opera- tors when using robots. Since a robot’s ultimate task is to serve people, the study greatly contributed by revealing crucial customer concerns. This study also enhances research on the interaction of smart technology and human service providers (Marinova et al., 2017) by revealing how robots can complement ser- vice providers and improve the customer experience in QSRs.

Extant published research has suggested that further research is needed in the area of developing and incorporating robotic technologies in the service industry (Tung & Law, 2017). The current study begins this work in the context of the QSR industry to determine the future directions for these technologies. Furthermore, the current study advanced Murphy, Gretzel, et al.’s (2017) and Murphy, Hofacker, et al.’s (2017) research by significantly enhancing an under- standing of the customer’s acceptance of robots in the hospitality industry.

This study’s results also augment Veer’s (2013) and Murphy, Gretzel, et al.’s (2017) research on anthropomorphism loyalty, by providing detailed customer perception information of the advantages and disadvantages of using robots in QSRs, therefore suggesting how to potentially create successful robot human interactions. The results also support previous research (Blow et al., 2006; MacDorman, 2005; Mori, 1970, 2012), which found that when anthropomor- phism reaches a point where human replicas appear nearly like human beings, feelings of distress are experienced and likability drastically drops. On the other hand, as the nonhuman entity becomes exactly humanlike (Bartneck et al., 2007), likability drastically increases. Participants in this study expressed the same sentiments, explaining their belief that the current state-of-the-art technol- ogy cannot provide a humanlike machine that will overcome the feelings of creepiness or eeriness.

Moreover, this study enhances previous research by Giebelhausen et al. (2014), who found that successful implementation of smart technology depends on maintaining an appropriate employee–customer rapport and finding the right balance between technology and human employees. Little prior research

28 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

examined the need to experience or to provide the “human touch,” or desire for interaction, in hospitality (Ko, 2017) and for building rapport between the cus- tomer and the service provider (Gremler & Gwinner, 2000; Ivanov et al., 2017; Sutton & Rafaeli, 1988). This study’s participants expressed concerns about the lack of “human touch,” revealing it as an important theme that merits further exploration. The NRA forecasts that a “Backlash against automation of all kinds could create a “return to artisanal” movement—predicated on humans being the center of all parts of the food and beverage process” (NRA, 2019, slide 39). While it is unlikely that most QSR brands will create significantly “artisanal” experiences, they should remain aware of this potential preference among their customers.

Finally, the study enhances previous research by revealing 20 major themes of customers’ perceptions of the use of robotics in restaurants. These themes lend themselves to further investigation using quantitative techniques to poten- tially yield new constructs of interest in this area, thus enhancing the literature in this field.

Managerial implications

The study also revealed important customer concerns regarding the use of robots in QSRs that owners and/or operators should acknowledge, as the adop- tion of this technology requires significant capital investment. Restaurant cus- tomers and health code agencies demand that restaurants serve food that is safe for human consumption. This study’s participants expressed deep concerns about potential problems related to cleaning and sanitizing the robots to avoid food contamination. They are also concerned about a robot’s inability to detect unsafe or undesirable food ingredients through visual inspection or detecting bad odors, functions which are currently served by trained employees. While all consumers expect food that is fit for consumption, the significant portion of the population that has worked in the foodservice industry heightens their aware- ness of restaurant sanitation protocols. Participants suggested that if a robot comes into direct contact with food, the company that develops or uses the robot needs to pay high attention to sanitation and cleanliness. QSR brand managers and their franchisees need to make sure that robots selected for use in the restau- rant are designed for easy sanitization and ensure that all lubricants are of food- grade quality. QSR brand managers and restaurant managers should also actively educate customers about food safety and the sanitation processes used with the robots. Providing this education should help alleviate concerns about the robots and contaminated food product. In fact, it could be an opportunity to emphasize the enhanced food quality and safety that a robotic device can offer versus the risks posed by human failure.

This study also revealed which attributes of physical appearances of robots are acceptable by QSR customers. For example, respondents rejected the notion of robots that are somewhat human in appearance and stated that they would

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 29

only feel comfortable with humanlike robots that are identical to humans. Until that technology is perfected, manufacturers and managers should avoid human- like appearances and should instead create devices that more clearly resemble the function they are designed to perform. If used in the FOH to interact with guests, the robots should also not rely on voice communication until the technol- ogy is perfected.

Next, while many studies point to the efficiencies gained through robotics, the experiences reported by this study’s participants suggest that future research is required to measure how well these technologies work in QSRs. For example, participants mentioned how slowly a pizza-making robot operates versus the participant’s own ability to complete preparation of a pizza in a fraction of the time. Similarly, the research team visited a bar that features two bartending robots, which are industrial robotic arms that have been adapted to make cock- tails. The research team members include experienced bartenders; they are able to prepare drinks much faster than the bartending robots can. Custom-designed robotics may be more effective at this time than robotic technology adapted from other industries.

QSR managers can also use the information provided in this study to create suc- cessful human–robot interactions and use the successful implementation of robots to encourage brand attachment and loyalty previously unattainable. For example, managers can create an environment that provides a synergy of novelty and enter- tainment, efficiency of food and beverage production, and service excellence.

Surprisingly, this study revealed very few generational differences in the per- ceptions of or attitudes toward robotics in the restaurant industry, with approxi- mately equal levels of enthusiasm and skepticism about robotic technology. Nevertheless, despite few generational differences, managers still need to under- stand their guests’ demographics to detect any differences in perceptions and attitudes that may not have been evident in this study.

Finally, managers should emphasize that the use of robots in the restaurant industry can complement human service providers (Marinova et al., 2017). This study’s participants were concerned about the potential reduction of employ- ment and its effects on society, along with the potential lack of “human touch” in customer interactions. QSR managers should explain that savings in labor costs will translate into stable or reduced prices for customers. They could also emphasize the labor opportunities created by robots in their restaurants in order to counteract such concerns. The NRA (2019) proposed that a future job cate- gory in the restaurant industry is the “food engineer.” Operators also need to ensure that employees who work side-by-side with the robots have excellent service management skills so that the core product—food and beverage—is delivered with that still-essential human touch. In addition, the use of robots may have the potential to raise the quality of life of employees and customers if robots are doing a good job serving both groups. This may also allow employees to train for jobs that require more advanced skills by leaving menial tasks to robots, therefore advancing society as a whole.

30 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

limitations

It should be noted that the qualitative method used in this study cannot render data or insights that are generalizable across the population, nor are the data suit- able for analysis of reliability or validity. The research team’s interpretation of the results could have been influenced by their first-hand observations of the focus groups and their own experiences in the foodservice and greater hospital- ity industry. The very nature of qualitative research introduces a level of subjec- tivity into the analysis and interpretation of the results, as Weber (1990) suggests. In addition, a high percentage of focus group participants had either past or current work experience in the restaurant industry. While each participant quali- fied for focus group participation by being a current QSR customer, their knowl- edge of the restaurant industry itself may have introduced bias into their perceptions.

Next, the study’s participants were solicited via mall intercepts and on-cam- pus recruitment at a local university; thus, the study’s findings may be influ- enced by the effects of self-selection bias, as the individuals who agreed to participate in focus groups may share implicit characteristics that might not be representative of the general population.

Suggestions for Future research

Future research could use the themes and insights gleaned from this study to create a quantitative instrument that could be used to obtain data from a broader population to gain deeper, more objective insight into customer perceptions about using robotics in QSRs, full-service restaurants, and throughout the hospi- tality industry. Numerous mentions of the desire for human interaction or to provide or experience the “human touch” suggested that researchers should fur- ther examine constructs related to the desire for interaction with humans (Ko, 2017). The results of this study also pointed to further refinement of the indus- try’s understanding of design, maintenance, sanitation, and high-functioning interactions between robots and humans, particularly in BOH operations.

concluding SuMMary

To date, most robotic devices used in the hospitality industry are machines that were developed for other industries, which were then altered to fit a hospi- tality task. Today’s restaurant industry owners and managers face economic and labor availability challenges that may make incorporating robotic technology into their operations attractive. The most logical sector of the restaurant industry to deploy this technology is the QSR segment. This study sought to understand QSR customers’ perspectives on using robotics in the QSR industry to identify positive and negative aspects of the technology, including how/where robots can be most appropriately used, how they should not be used, what they should or should not look like, and other societal, safety, and cost-related concerns.

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 31

The study used focus groups consisting of QSR customers to elicit this infor- mation. Key findings of the study indicate that the study’s participants believe that robots are coming to the QSR industry, whether society is ready for them or not. Many participants also wistfully discussed the potential loss of the “human touch” in the restaurant industry, indicating that they still thought it was critical to maintain in some way.

Smart robotics developers and the companies deploying this technology should ensure that the robot, whether it is for the FOH or BOH, will not merely be a robot developed for another industry that is slightly modified to “fit” in a restaurant setting. The robotics should be designed with aesthetics suitable to the audience; this study’s participants had a strong preference for robots that look like machines and not like humans at this time. The robots must operate safely and must be able to be properly sanitized. Furthermore, the participants were not only restaurant customers, but many of them have also worked in the restaurant industry—similar to a large percentage of adult Americans—and could easily visualize what might go wrong. Current restaurant industry workers should be heavily involved in the design and deployment, partly because of their current knowledge of the industry, but also because they will be the people who will still continue to provide the important “human touch” going into the future.

orcid id

Dina Marie V. Zemke https://orcid.org/0000-0002-0766-2328

SuPPleMenTal MaTerial

Supplemental material for this article is available online.

reFerenceS

Abrams, R., & Gebeloff, R. (2018, May 3). A fast-food problem: Where have all the teenagers gone? The New York Times. https://www.nytimes.com/2018/05/03/upshot/ fast-food-jobs-teenagers-shortage.html

Ackerman, E. (2014, April 30). Avidbots wants to automate commercial cleaning with robots. IEEE Spectrum. https://spectrum.ieee.org/automaton/robotics/industrial- robots/avidbots-want-to-industrialize-robot-cleaning

Barrett, L. F. (2005). Valence is a basic building block of emotional life. Journal of Research in Personality, 40(1), 35-55. https://doi.org/10.1016/j.jrp.2005.08.006

Bartneck, C., Kanda, T., Ishiguro, H., & Hagita, N. (2007). Is the uncanny valley an uncanny cliff? RO-MAN 2007: The 16th IEEE International Symposium on Robot and Human Interactive Communication, 2007, 368-373. https://doi.org/10.1109/ ROMAN.2007.4415111

Baum, T. (2006). Human resource management for tourism, hospitality and leisure: An international perspective. Thomson.

Bishop, S. (2019, June 25). The cost of the restaurant robot. QSR automations. https:// www.qsrautomations.com/blog/restaurant-technology/restaurant-robot/

32 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

Blow, M., Dautenhahn, K., Appleby, A., Nehaniv, C. L., & Lee, D. C. (2006). Perception of robot smiles and dimensions for human-robot interaction design. RO-MAN 20067: The 15th IEEE International Symposium on Robot and Human Interactive Communication, 2006, 469-474. https://doi.org/10.1109/ROMAN.2006.314372

Bowen, J., & Morosan, C. (2018). Beware hospitality industry: The robots are coming. Worldwide Hospitality and Tourism Themes, 10(6), 726-733. https://doi.org/10.1108/ WHATT-07-2018-0045

Byrd, E. (2015, July). Making wages work: How three brands are getting ahead of mini- mum wage increases. QSR Magazine. https://www.qsrmagazine.com/employee- management/making-wages-work?page=2

Chung, J.-Y. C., & Cakmak, M. (2018). “How was your stay?” Exploring the use of robots for gathering customer feedback in the hospitality industry. RO-MAN 2018: The 27th IEEE International Symposium on Robot and Human Interactive Communication, 2018, Article 25604. https://doi.org/10.1109/ROMAN.2018.8525604

Colombetti, G. (2005). Appraising valence. Journal of Consciousness Studies, 12(8-10), 103-126.

Creswell, J. W. (2015). A concise introduction to mixed methods research. Sage. Edelbrock, G. (2012, December 20). EPSON robots: Wash down robots for food, medi-

cal applications. Packaging World. https://www.packworld.com/article/food/pro- duce/epson-robots-wash-down-robots-food-medical-applications

Epley, N., Waytz, A., Akalis, S., & Cacioppo, J. T. (2008). When we need a human: Motivational determinants of anthropomorphism. Social Cognition, 26(2), 143-155. https://doi.org/10.1521/soco.2008.26.2.143

Fink, J., Bauwens, V., Mubin, O., Kaplan, F., & Dillenbourg, P. (2011). People’s per- ception of domestic service robots: Same household, same opinion? In International Conference on Social Robotics (pp. 204-213). Springer.

Giebelhausen, M., Robinson, S. G., Sirianni, N. J., & Brady, M. K. (2014). Touch ver- sus tech: When technology functions as a barrier or a benefit to service encounters. Journal of Marketing, 78(4), 113-124. https://doi.org/10.1509/jm.13.0056

Graham, J. (2018, May 28). Hamburger-making robot Flippy is back at Calif. Chain. USA Today. https://www.usatoday.com/story/tech/talkingtech/2018/05/28/hamburger- making-robot-flippy-back-serving-300-burgers-day/649370002/

Gremler, D. D., & Gwinner, K. P. (2000). Customer-employee rapport in service relationships. Journal of Service Research, 3(1), 82-104. https://doi.org/10.1177 /109467050031006

Hennig-Thurau, T., Groth, M., Paul, M., & Gremler, D. D. (2006). Are all smiles created equal? How emotional contagion and emotional labor affect service relationships. Journal of Marketing, 70(3), 58-73. https://doi.org/10.1509/jmkg.70.3.58

International Federation of Robotics. (2016). World Robotics Service Robots 2016. Retrieved from https://ifr.org/img/office/Service_Robots_2016_Chapter_1_2.pdf

International Organization for Standardization. (2012). Robots and robotic devices: Vocabulary (Standard No. 8373). https://www.iso.org/standard/55890.html

Ivanov, S., Webster, C., & Berezina, K. (2017, May 17-19). Adoption of robots and service automation by tourism and hospitality companies [Paper presentation]. INVTUR Conference, Aveiro, Portugal. https://www.researchgate.net/publication /316702026_Adoption_of_robots_and_service_automation_by_tourism_and_ hospitality_companies

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 33

Kanda, T., Shiomi, M., Miyashita, Z., Ishiguro, H., & Hagita, N. (2009). An affective guide robot in a shopping mall. RO-MAN 2009: The 4th ACM/IEEE International Conference on Human Robot Interaction, 2009, 173-180. https://doi.org/10. 1145/1514095.1514127

Ko, C.-H. (2017). Exploring how hotel guests choose self-service technologies over ser- vice staff. International Journal of Organizational Innovation, 9(3), 16-27. https:// files.transtutors.com/cdn/uploadassignments/2802178_2_yes-xploring-how-hotel- guests-choose-self-service.pdf

Kuo, C. M., Chen, L. C., & Tseng, C. Y. (2017). Investigating an innovative service with hospitality robots. International Journal of Contemporary Hospitality Management, 29(5), 1305-1321. https://doi.org/10.1108/IJCHM-08-2015-0414

Kuo, I. H., Rabindran, J. M., Broadbent, E., Lee, Y. I., Kerse, N., Stafford, R. M. Q., & MacDonald, B. A. (2009). Age and gender factors in user acceptance of health- care robots. RO-MAN 2009: The 18th IEEE International Symposium on Robot and Human Interactive Communication, 2009, 214-219. https://ieeexplore.ieee.org/ abstract/document/5326292

MacDorman, K. F. (2005). Androids as an experimental apparatus: Why is there an uncanny valley and can we exploit it. In Toward social mechanisms of android sci- ence (pp. 106-118). John Benjamins.

MacDorman, K. F., & Entezari, S. O. (2015). Individual differences predict sensitivity to the uncanny valley. Interaction Studies, 16(2), 141-172. https://doi.org/10.1075/ is.16.2.01mac

MacDorman, K. F., & Ishiguro, H. (2006). The uncanny advantage of using androids in cognitive and social science research. Interaction Studies, 7(3), 297-337. https://doi. org/10.1075/is.7.3.10mac

Mariampolski, H. (2001). Qualitative market research: A comprehensive guide. Sage. Marinova, D., de Ruyter, K., Huang, M.-H., Meuter, M. L., & Challagalla, G. (2017).

Getting smart: Learning from technology-empowered frontline interactions. Journal of Service Research, 20(1), 29-42. https://doi.org/10.1177/1094670516679273

Mazurenko, O., Zemke, D., Lefforge, N., Shoemaker, S., & Menachemi, N. (2015). What determines the surgical patient experience? Exploring the patient, clinical staff, and administration perspectives. Journal of Healthcare Management, 60(5), 332-346. https://doi.org/10.1097/00115514-201509000-00007

McFarland, M. (2016, May 25). Ex-McDonald’s CEO says raising the minimum wage will help robots take jobs. The Washington Post. https://www.washingtonpost.com/ news/innovations/wp/2016/05/25/ex-mcdonalds-ceo-says-raising-the-minimum- wage-will-help-robots-take-jobs/

Miller, S. I., & Fredericks, M. (1999). How does grounded theory explain? Qualitative Health Research, 9(4), 538-551. https://doi.org/10.1177/104973299129122054

Mori, M. (1970). The uncanny valley. Energy, 7(4), 33-35. Mori, M. (2012). The uncanny valley. IEEE Robotics & Automation Magazine, 19(2),

98-100. https://doi.org/10.1109/MRA.2012.2192811 Murphy, J., Gretzel, U., & Hofacker, C. (2017, May). Service robots in hospitality and

tourism: Investigating anthropomorphism. In 15th APacCHRIE Conference. http:// heli.edu.au/wp-content/uploads/2017/06/APacCHRIE2017_Service-Robots_paper- 200.pdf

34 JOURNAL OF HOSPITALITY & TOURISM RESEARCH

Murphy, J., Hofacker, C., & Gretzel, U. (2017). Dawning of the age of robots in hospital- ity and tourism: Challenges for teaching and research. European Journal of Tourism Research, 15, 104-111.

National Restaurant Association. (2019). Restaurant industry 2030: Actionable insights for the future. https://restaurant.org/Downloads/PDFs/Research/Restaurant2030.pdf

Patton, L. (2019, May 19). Fast-food workers have a new job perk: Finish a shift, get cash and go. Fortune. https://fortune.com/2019/05/19/fast-food-franchise-employees-pay- labor-shortage/

Picard, R. (1997). Affective computing. MIT Press. Picard, R. W. (2003). Affective computing: Challenges. International Journal of Human-

Computer Studies, 59(1-2), 55-64. https://doi.org/10.1016/S1071-5819(03)00052-1 Qureshi, M. O., & Sajjad, R. (2017). A study of integration of robotics in the hospitality

sector and its emulation in the pharmaceutical sector. Health Science Journal, 11(1), 1-6. https://doi.org/10.21767/1791-809X.1000483

Rafaeli, A., Altman, D., Gremler, D. D., Huang, M.-H., Grewal, D., Iyer, B., Parasuraman, A., & de Ruyter, K. (2017). The future of frontline research: Invited commentaries. Journal of Service Research, 20(1), 91-99. https://doi.org/10.1177/1094670516679275

Saldaña, J. (2016). The coding manual for qualitative researchers (3rd ed.). Sage. Seyama, J. I., & Nagayama, R. S. (2007). The uncanny valley: Effect of realism on

the impression of artificial human faces. Presence: Teleoperators and Virtual Environments, 16(4), 337-351. https://doi.org/10.1162/pres.16.4.337

Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86(2), 420-428. https://doi-org.proxy.bsu.edu /10.1037/0033-2909.86.2.420

Solomon, M. R., Surprenant, C., Czepiel, J. A., & Gutman, E. G. (1985). A role theory perspective on dyadic interactions: The service encounter. Journal of Marketing, 49(1), 99-111. https://doi.org/10.1177/002224298504900110

Sutton, R. I., & Rafaeli, A. (1988). Untangling the relationship between displayed emotions and organizational sales: The case of convenience stores. Academy of Management Journal, 31(3), 461-487. https://doi.org/10.5465/256456

The Aspen Institute. (2013). Reinventing low wage work: Ideas that can work for employ- ees, employers and the economy. https://www.aspeninstitute.org/series/reinventing- low-wage-work/?state=past

Tung, V. W. S., & Law, R. (2017). The potential for tourism and hospitality experi- ence research in human-robot interactions. International Journal of Contemporary Hospitality Management, 29(10), 2498-2513. https://doi.org/10.1108/IJCHM-09- 2016-0520

Tussyadiah, I. P., & Park, S. (2018). Consumer evaluation of hotel service robots. In B. Stangl & J. Pesonen (Eds.), Information and communication technologies in tourism 2018 (pp. 308-320). Springer. https://doi.org/10.1007/978-3-319-72923-7_24

Veer, E. (2013). Made with real crocodiles: The use of anthropomorphism to promote product kinship in our youngest consumers. Journal of Marketing Management, 29(1- 2), 195-206. https://doi.org/10.1080/0267257X.2012.759990

Waytz, A., Cacioppo, J., & Epley, N. (2014). Who sees human? The stability and impor- tance of individual differences in anthropomorphism. Perspectives on Psychological Science: A Journal of the Association for Psychological Science, 5(3), 219-232. https://doi.org/10.1177/1745691610369336

Zemke et al. / BETTER ROBOT FOR QUICK-SERVICE RESTAURANTS 35

Weber, R. P. (1990). Basic content analysis (2nd ed., Sage University Paper Series on Quantitative Applications in the Social Sciences, Series no.49). Sage.

Yu, Q., Yuan, C., Fu, Z., & Zhao, Y. (2012). An autonomous restaurant service robot with high positioning accuracy. Industrial Robot: An International Journal, 39(3), 271-281. https://doi.org/10.1108/01439911211217107

Zalama, E., García-Bermejo, J. G., Marcos, S., Domínguez, S., Feliz, R., Pinillos, R., & López, J. (2014). Sacarino, a service robot in a hotel environment. In ROBOT2013: First Iberian Robotics Conference (pp. 3-14). Springer.

submitted July 15, 2020 Accepted April 3, 2020 refereed Anonymously

dina Marie V. Zemke, PhD (e-mail: [email protected]), is an associate professor of residential property management at Department of Applied Business Studies, Miller College of Business, Ball State University. Jason Tang, PhD (e-mail: jason.tang@ pm.me), is a senior lecturer at Department of Accounting and Finance, Faculty of Business and Communication Studies, Mount Royal University. Carola raab, PhD (email: [email protected]), is a professor at Department of Food and Beverage, Meetings, and Event Management, William F. Harrah College of Hospitality, University of Nevada, Las Vegas. Jungsun Kim, PhD (email: [email protected]), is an associate professor at the Department of Resort, Gaming and Golf Management, William F. Harrah College of Hospitality, University of Nevada, Las Vegas.