Project Title: Robots in hospitality and tourism: Can they create memorable experiences?
Description of the theoretical and practical importance of the research
With the development of technology, robots have been applied in various sectors of the hospitality and tourism industry (e.g., airlines, hotels, restaurants); they are used as service agents to perform customer services (Berezina et al., 2019; Ivanov et al., 2019). This research is focusing on the embodied service robots powered by artificial intelligence (AI) that deliver service to customers at the front line. Service robots powered by AI can interact with the customer as regular frontline employees (Lu et al., 2019). Given the trend of the aging population, using service robots as frontline employees is expected to be a norm in hospitality in the near future (Webster & Ivanov, 2020).
Thus, hospitality and tourism managers should have a tool to measure customer experiences with service robots to be able to enhance the overall service experience and increase satisfaction and loyalty (Gallarza et al., 2015). Furthermore, the service robot developers also need to have instruments to measure customer experience with service robots in hospitality settings to develop industry-specific robots. However, to date, there is no scale developed for measuring customer experience with service robots. Thus, this research aims to fill the void and develop a scale by exploring the measurement items and factors of customer experience with service robots in hospitality and tourism.
Methods and Procedures
The scale development is based on grounded theory. This research will follow a widely accepted psychometric procedure for scale development by Churchill (1979) and Gerbing and Anderson (1988). The scale development consists of four separate stages: 1) item generation, 2) item purification, 3) item reduction and dimensionality, and 4) scale validation. We plan to use data collected via Prolific to perform the last two stages of scale development. The data collection for item generation and purification stages are funded by a grant provided by the University of Mississippi; thus we will not explain these stages here.
The hospitality and tourism customers who had experience with service robots will be surveyed with a questionnaire with the items left after the item purification stage. The participants will be recruited through the Prolific online panel among the US and UK residents who are fluent in English. As Prolific does not allow screening questions in the questionnaire, first, we will run a short initial screening survey that contains the questions to check if the respondents experienced service from robots in hospitality and tourism settings. Then we will collect Prolific IDs of participants who experienced service robots and create a “Custom Allowlist” with their IDs. Next, the questionnaire with the items left after the item purification stage will be distributed only to those respondents who are in the Allowlist.
Exploratory factor analysis (EFA) will be performed to reduce the number of scale items and understand the underlying dimensionality of the scale with remaining items and assessments of the scale reliability (Hair et al., 2019). Then, confirmatory factor analysis (CFA) will be performed to validate the measurement scale (Hair et al., 2019). These analyses require different samples (Hair et al., 2019); thus, the collected Prolific data will be split into two samples randomly.
Sample size estimation
The sample sizes for the last two stages of the scale development (i.e., 3) item reduction and dimensionality and 4) scale validation) will be determined based on the number of the remaining scale items after the item purification stage (Hair et al., 2019).
The number of scale items for further factor analyses (EFA and CFA) was determined based on the previous academic research in the field of robots, automation, and artificial intelligence. Ivanov et al. (2021) applied EFA to 49 measurement items to identify the factors that impact the fear of automation among Bulgarian employees. In the scale development of a service robot integration willingness scale by Lu et al. (2019), the instrument contained 51 items remaining for EFA. Thus, for this project, we estimate 50 measurement items will remain before EFA. This research will use ten responses per item that is the conservative threshold for determining the sample size for the reliability of factor analysis (Hair et al., 2019). Thus, the sample size for the surveys via Prolific online panel is 500 responses for each EFA and CFA, totaling 1000 responses. It is expected that some data will be lost due to data cleaning. The minimum effective total sample size for both factor analyses is 500, which would provide 250 for each analysis and five responses per one measurement item, and still offer an acceptable sample size for the factor analyses (Hair et al., 2019).
Description of the study costs
Prolific reports a “40-50% response rate from eligible participants” (Prolific, 2018). Thus, to get 1,000 responses for the scale development, we need to identify 2,500 individuals who have had experience with service robots while traveling (or 3,000 to be on the safe side after data cleaning). Literature suggests that about 12-15% of people may have already experienced robots in hospitality (Abufele et al., 2018; Cain & Berezina, 2021). Thus, we need to screen 25,000 people to get 3,000 on the AllowList to get 1,000 qualified responses for the scale development. Therefore, first, the short screening survey questionnaire will be distributed to 25,000 Prolific participants at a £7.80 hourly rate. The screening questionnaire will require not more than 1 minute to respond. Thus, the total cost of the screening survey estimated by the Prolific price estimator (Prolific | Pricing) is £4,333.33.
The main survey with estimated 50 measurement items will be distributed to 1,000 respondents from the Allowlist who experienced service robots in hospitality and tourism at a £7.50 hourly rate. The main questionnaire will require around 20 minutes to respond to. Thus, the total cost of the main survey estimated by the Prolific price estimator (Prolific | Pricing ) is £3,333.33.
Overall, we need £7,666.66 to complete the project.
Open Science Commitment
The study is pre-registered at AsPredicted.org. The pre-registration is available for peer-review. All data, materials, and analysis results will be made available on open-access repositories (e.g., OSF, PsyArxiv, author’s website). We intend to publish in an open-access journal (e.g., Journal of Tourism Futures) or with a publisher that allows a pre-print version of the manuscript to be freely available on the open-access repositories without any restrictions.
Abufele, S., Yu, W. G., Chen, C., & Booras, T. (2018, February 28). What is yet to come: Robots taking over the hospitality industry. Hospitality Upgrade. Retrieved June24, 2021, from Hotel Technology Blog | Tech Talk on Hospitality Upgrade
Berezina, K., Ciftci, O., & Cobanoglu, C. (2019). Robots, artificial intelligence, and service automation in restaurants. In S. Ivanov & C. Webster (Eds.), Robots, Artificial Intelligence, and Service Automation in Travel, Tourism and Hospitality (pp. 185-219). Emerald Publishing Limited. https://doi.org/10.1108/978-1-78756-687-320191010
Cain, L., & Berezina, K. (2021). Robo-Tipping: Are Customers Game?. In Information and Communication Technologies in Tourism 2021 (pp. 222-227). Springer, Cham.
Churchill Jr, G. A. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16 (1), 64-73.
Gallarza, M. G., Arteaga, F., Del Chiappa, G., & Gil-Saura, I. (2015). Value dimensions in consumers’ experience: Combining the intra-and inter-variable approaches in the hospitality sector. International Journal of Hospitality Management, 47, 140-150. Redirecting
Gerbing, D. W., & Anderson, J. C. (1988). An updated paradigm for scale development incorporating unidimensionality and its assessment. Journal of Marketing Research, 25 (2), 186-192.
Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2019). Multivariate Data Analysis (8th Ed.). Cengage.
Ivanov, S., Gretzel, U., Berezina, K., Sigala, M. and Webster, C. (2019). Progress on robotics in hospitality and tourism: a review of the literature. Journal of Hospitality and Tourism Technology, 10 (4), 489-521. https://doi.org/10.1108/JHTT-08-2018-0087
Ivanov, S., Kuyumdzhiev, M., & Webster, C. (2020). Automation fears: Drivers and solutions. Technology in Society, 63, 101431. Redirecting
Lu, L., Cai, R., & Gursoy, D. (2019). Developing and validating a service robot integration willingness scale. International Journal of Hospitality Management, 80, 36-51. Redirecting
Prolific (2018, September 12). Using our demographic filters to prescreen participants. Using our demographic filters to prescreen participants – Prolific
Webster, C. & Ivanov, S. (2020). Demographic change as a driver for tourism automation. Journal of Tourism Futures, 6(3), 263-270. Demographic change as a driver for tourism automation | Emerald Insight