Exploring the Key Attributes Influencing University Students’ Domestic Accommodation Choice: A RIDIT Analysis

Abstract

This study aims to provide an understanding of the preferences of student tourists regarding service attributes related to domestic accommodation. To clarify how lodging firms formulate policies for the student travel market, a “Relative to an Identified Distribution” (RIDIT) analysis in which responses are treated as ordinal data instead of equally spaced data is performed to examine the Likert-scale data obtained from 296 university students in Taiwan. In our RIDIT analytical results, respondents’ preferences are sorted in descending importance. Of the examined attributes, “Cleanliness of rooms” is the attribute most prioritized by the students, followed by “Located in a safe neighborhood/feeling safe in the room” and “Close to scenic area for meeting the trip requirement”. The attribute least prioritized by the students is “Interaction with other guests”. This analysis provides conclusions that differ from those obtained through commonly used techniques including the arithmetic average and exploratory factor analysis. The preference rankings indicate that managers should focus on the most prioritized service attributes and implement the necessary measures to enhance their service quality and maximize customer satisfaction.

1. Introduction

The onset of the COVID-19 pandemic, which led to the implementation of restrictions, regulations, and efforts toward avoiding infection, has fundamentally altered travel behavior [1]. Fernández et al. [2] found that the number of international tourists decreased by 72% over the first 10 months of 2020. The pandemic consumption habits changed [3], and tourists began to engage in domestic travel [4]. The travel behavior of university and college students changed after COVID-19 and is continuing to evolve [1,5,6]. Based on Cavagaro and Staffieri [7], the behavioral tendencies of younger generations lead to the development of new approaches to tourism. The student tourism market is crucial because students will someday become the main consumers and form the foundation of the tourist market [8]. However, research into their preferences regarding accommodation during domestic travel remains sparse and warrants further exploration. Considering the tremendous potential of this travel market segment, in addition to the current instability and dynamic competition, new strategies should be developed to ensure accommodations are available to meet young domestic travelers’ expectations.

The relevant literature has mainly focused on international travel by students [9,10,11]. Accommodation costs account for the largest portion of student tourists’ travel budgets [12]. Research on young travelers has often made assumptions regarding such travelers’ accommodation budgets or classified such travelers, particularly millennial travelers, as backpackers [13]. Few studies have investigated the market for domestic tourism among young adults, possibly because of the considerable differences in youth travel behaviors across cultures [11,14]. Researchers have reported that university students are a heterogeneous group and called for further investigations in a cross-country context [15,16]. The factors influencing young tourists’ domestic accommodation choices likely differ from those affecting their international accommodation choices. In addition, the increased travel costs in response to the COVID-19 pandemic have restricted overseas travel, which has increased the appeal of domestic travel [17]. Thus, an empirical case study investigating the accommodation preferences of student domestic travelers would enrich the tourism literature.

Another research issue worthy of attention is various analytical methods to gauge guests’ accommodation preferences. To examine the rationale underlying accommodation choice and the concerns regarding such choices, researchers typically collected tourist data by conducting Likert-scale surveys. Tourists differ in terms of their rating of accommodation options, and their evaluations are subjective; that is, two tourists may assign the same rating to an accommodation choice-related attribute even though the extent of their preference for the attribute may differ. In addition, researchers often evaluated such data by employing descriptive statistical techniques, ANOVA/t-test, or exploratory factor analysis to identify collective characteristics [18]. Further, complex processes such as multiple regression and structural equation modeling have been employed to develop preference functions [19].

For addressing the above-mentioned research gaps, RIDIT analysis may be considered as a straightforward tool though few studies have applied this approach to examine the tourism sector, and understanding of how prospective tourists think is limited. According to Bhattacharya and Kumar [20], RIDIT analysis can be used to perform service quality gap analysis or service quality benchmarking in the hospitality, tourism, and travel sectors. Thus, RIDIT analysis is suitable to evaluate the preferences of young tourists related to accommodation service attributes, and it may become a key analytical tool yielding empirical findings in this research domain.

Thus, the present study aims to identify and rank the attributes related to the domestic tourism accommodation choices of young tourists. University students were recruited as the sample in the subsequent analysis. Specifically, our research objective is to rank the service attributes related to domestic accommodation choices by priority, thereby facilitating the development of plans for maximizing the satisfaction of young tourists. Except the RIDIT analysis regarding the priority of the accommodation service attributes based on previous lodging experiences of young tourists, we also compare the analytical results with those obtained through other statistical methods. The research findings may help lodge managers, particularly in hostels, to identify key areas of improvement and maximize their service quality.

Our study has a two-fold contribution. Firstly, we illustrate RIDIT analysis as an alternative analytical methodology for the study of accommodation service quality. Secondly, we provide a tool so that practitioners can use priority rankings to decide their course of actions as to which one to focus on more and which one later. This can also improve organizational efficiency. The remainder of this article is organized as follows. Firstly, we review the literature and briefly introduce the concept of RIDIT analysis. Next, we present the results of our RIDIT analysis and other statistical methods. Finally, we conclude by discussing the implications and application of our findings as well as future research directions.

2. Literature Review

2.1. Accommodation Choices of University Students

Research on university student travel accommodation has been sparse. Heung and Leong [21] found that university students in Hong Kong tend to stay in hostels while traveling rather than in hotels or motels. This may be because in hostels, students are able to prepare their own meals in the hostel kitchens and engage in social interactions that may lead to friendships. Cooking in these kitchens is generally less costly than dining out is, and therefore, it can reduce overall travel costs if consistently practiced throughout the trip. Moisă [12] reports that the accommodation budgets of young tourists accounted for 18% of their travel expenses and is ranked by them as the second-most influencing factor. One recent study on the topic by Alsawafi and Almuhrzi [22] identified reputation and quality as the primary factors influencing Arab college students’ accommodation choices during travel.

The student tourism segment is frequently included within the broader youth travel market in tourism research. Young tourists also form the core of the backpacker population. Based on an overlap in the characteristics of working holiday makers (WHMs), students, and backpackers [14], many backpackers possess working holiday visas, and some backpackers are international students with student visas. In addition, WHMs are often categorized as a subset of backpackers [23,24]. Thus, reviewing the literature on young tourists who prefer WHM, backpacker, or hostel accommodations can be valuable. Backpackers frequently opt for accommodations such as hostels or backpacker-specific lodging, and WHMs and backpackers have similar characteristics, such as a preference for budget accommodations [25]. A study indicates that the three factors with the most influence on accommodation choices are price, location, and ratings and reviews; the other factors are cleanliness, staff friendliness, facilities, atmosphere, and safety [26]. In another study evaluating the overall quality of a hostel, surveyed backpackers rate social atmosphere as the most influential service dimension out of six service quality dimensions (i.e., social atmosphere, location and city connections, staff, cleanliness, security, and facilities) [9].

Given the heterogeneity in the markets of backpackers and students [27], Pearce and Son [28] have conducted a further investigation of their accommodation preferences. They discovered that, compared to backpackers, international English-language student travelers are more likely to travel in large groups, opt for hotels or motels, and prefer city-based travel activities. Furthermore, Bicikova [15] provided empirical evidence regarding the heterogeneity of students’ travel preferences, demonstrating that age and gender, among other variables, significantly influence their tourist behavior. Considerable differences have been observed in the travel behavior between international and domestic students, and Bicikova [15] argued that nationality and cultural background should be considered during segmentation of the student travel market. Xiao, So, and Wang [16] advocated for differentiated marketing strategies tailored to the distinct preferences of domestic and international student travelers. Thus, a study focusing on the accommodation preferences of domestic student travelers may shed insights into the related research issue.

Because an increased amount of attention has been given to exploring peer-to-peer (P2P) accommodation and its attraction for young tourists [29], this study reviewed the literature on this topic. In addition, this study addressed the changes in accommodation choices and the factors influencing tourists’ preferences and selections after the COVID-19 pandemic.

2.2. Peer-to-Peer (P2P) Accommodation Choices

With the increasing popularity of P2 accommodation platforms, understanding the perspectives and preferences of young tourists is crucial because of their varied tastes and preferences in the product market and the tourism sector [29]. Millennials have considerable purchasing power and are, thus, a key target segment of P2P accommodation platforms [30]. Research has demonstrated that Airbnb appeals more to younger demographics than to older ones [29,31]. Travelers seek various benefits including value for money and authentic accommodation experiences. Factors that influence the choice of Airbnb accommodations include cost, unique local experiences, cultural exchange, social interaction with hosts and local residents, and insider tips on local attractions [29,31,32,33]. Other highly influencing factors include household amenities, home coziness, and large spacious areas [34]. Sustainability-related dimensions, such as reducing the environmental, social, and economic effects of accommodation consumption, may be a factor influencing travelers’ selection of Airbnb accommodation [29]. Notably, the location of an accommodation is a critical factor. Tussyadiah and Zach [35] reveal that travelers prefer quiet neighborhoods that are located within walking distance of local restaurants and are a short bus ride away from the city center. P2P platforms tend to focus on the discussions of neighborhoods and local businesses, whereas hotels tend to stress their proximity to attractions [36].

Several salient P2P accommodation features, such as opportunities for meaningful social encounters, contribute to the main characteristics that differentiate Airbnb from traditional accommodation services [32]. When hosts are available at accommodations, guests can experience a community-focused and social atmosphere and form authentic connections with locals [37]. Mody et al. [38] identify several dimensions in which Airbnb accommodations are superior to hotels; they include localness, communities, personalization, escapism, aesthetics, and entertainment. Tussyadiah and Zach [35] report that P2P accommodations create more conditions for building relationships with hosts and local areas, whereas hotels provide better functional services and accessibility, such as airport shuttle services, in-room services, breakfast options, and free parking. Similarly, Belarmino et al. [36] discover that conversations with hosts is a key service attribute of P2P accommodations, and that hotel room amenities (e.g., food, beverages, and scent) are the predominant themes discussed in guest reviews. Varma et al. [39] conclude that Airbnb guests regard P2P accommodations as an alternative to traditional accommodations.

Although utilitarian, hedonic, and social values are key concerns that influence the selection of P2P accommodations [40,41], perceived risk has a strong negative influence on customer satisfaction and even customer purchase intention [41,42]. Relative to users of P2P accommodations, non-users perceive higher risks associated with using P2P accommodations and exhibit lower purchase intentions with respect to booking P2P accommodations [42]. Young tourists from different generation cohorts may also differ in terms of P2P accommodation choices. Fan et al. [40] indicate that the Generation Z cohort place more emphasis on sustainability and the value of social interactions relative to millennials. These findings provide insights into the accommodation preferences of young university students.

In summary, P2P and traditional accommodations share several fundamental attributes, even though most studies have indicated that P2P accommodations are attractive because of their low pricing and diverse options [43]. Except for host profiles, few differences have been identified in the factors influencing tourists to select Airbnb or traditional accommodations such as B&Bs [44]. Nevertheless, the present study aims to identify the factors that influence young tourists choosing domestic accommodations.

2.3. Post-COVID-19 Pandemic Students’ Travel Behavior and Accommodation

The tourism market changed considerably after the pandemic. According to Longwoods International [4], tourists shifted from a preference for international travel to that of domestic tourism, that is, tourism in which individuals travel within their own country. Domestic tourism offers a sense of increased safety and comfort because tourists are closer to home throughout their trip. According to the report of the World Tourism Organization [45], domestic tourism opportunities have expanded: by 2020, the market reached six times that of the international market in terms of tourist numbers. Several studies have focused on post-pandemic domestic travel behaviors [46]. Sohn et al. [47] discovered that the pandemic led to Korean tourists’ developing a stronger preference for short-haul destinations because of social distancing requirements in the country. Balińska and Olejniczak [48] found that Polish tourists altered their tourism behavior by avoiding crowded spots and abstaining from international travel to comply with hygiene requirements in the area. Yu [6] observed that many Chinese citizens have integrated travel into their daily lives, that is, have developed a preference for engaging in local leisure activities, which may have reduced the frequency with which they engage in long-distance travel. In summary, tourist behaviors and motivations have changed because of the pandemic in terms of, for example, destination choice, use of tourist services, travel frequency, measures of preserving health, and sense of travel safety [5].

Studies have indicated that student tourists’ behaviors have also changed because of the COVID-19 pandemic. Kusumaningrum and Wachyuni [49] found that after the pandemic, the most common type of millennial Indonesian tourist became the explorer type, followed by the drifter type. In addition, they determined that young tourists mainly had physical or physiological motives for traveling and that interpersonal motives were also crucial to them. Furthermore, the researchers found that the millennial Indonesian tourists primarily chose natural destinations. Uysal [8] reported that travelers from Generation Z differed from those from Generation X and Generation Y and revealed a transformation from traditional tourist behavior to modern tourist behavior in the post-COVID 19 pandemic. Olszewski-Strzyzowski, Pasek, and Lipowski [5] indicated that the travel motives of university students and their engagement in tourism activities changed after the pandemic. A survey among Romanian university students highlighted that vacation expenses were a critical concern for such travelers because the students generally had low monthly incomes [1]. Yu [6] found that college students were more cautious and more likely to prioritize safety and affordability after the pandemic, indicating that security and affordability should be highlighted as the main attributes of travel products.

Various characteristics were found to cease to be advantages of P2P accommodations, particularly in the post-COVID-19 era. Gassmann et al. [50] identify price as a predominant factor for selecting P2P accommodation. However, price is not a significant contributing factor for consumer loyalty toward Airbnb [51]. In addition, social interactions do not play a key role in influencing the P2P accommodation choices of consumers [19,52]. This may be related to health and safety concerns and the reduction in interpersonal interactions. Similarly, authentic experiences and social benefits are crucial for enhancing consumers’ commitment to Airbnb [51]. However, Santos et al. [53] describe that Airbnb users are the least focused on experiencing local culture, interacting with other guests, and living like a local.

2.4. Statistical Methods Used in Related Studies

Spoerr [18] summarized the analytical methods used to categorize the hotel attributes. To face various accommodation attributes, exploratory factor analysis (EFA) has been used frequently to examine the structure of the interrelationships among a large set of variables to define some common underlying dimensions. Indeed, many researchers applied EFA to reduce the number of accommodation attributes and identify the underlying constructs [9,18,54,55]. However, it is inevitable to lose some important information by grouping variables. Some parametric tests, for example, t-test or ANOVA may be conducted further for each factor to investigate the differences based on the respondents’ characteristics. Furthermore, the test procedure should meet the requirement that is under a compulsory distribution.

Other complex analytical approaches have also been applied to identify specific accommodation attributes such as a hierarchical regression model [56], conjoint analysis [57], confirmatory factor analysis [58], and structural equation modeling [59]. Ye et al. [19] employ multiple linear regression based on the ordinary least square method to determine whether tourists preferred hotel or P2P accommodations. They construct a relative preference model that considers multiple value attributes and personal traits. Although the complexities of statistical techniques have shed insights into the research issue, using a powerful but easy analytical method makes it challenging to draw conclusions about the importance of individual accommodation attributes.

3. Research Methods

3.1. RIDIT Analysis

Scales are commonly adapted to collect data in consumer studies. Clason and Dormody [60] report that data collected through Likert-scale surveys usually involve the use of ordinal form or interval scales; this strategy assumes that the latent variables reflect the attitudes and opinions of respondents. Consumers usually fill out a questionnaire with scale items based on their experiences, thoughts, and preferences. Each scale item is rated and assigned a numerical value that is represented by a verbal category. Two consumers may rate an item as 5 on a 7-point scale; however, their perception of the meaning of the rating may differ. Relatedly, Pouplard, Qannari, and Simon [61] indicate that variations exist between consumers, and that such biases influence data quality because of the subjectivity of Likert-scale results. Nonetheless, this method of data collection is frequently employed in consumer and social science research. Clason and Dormody [60] report that descriptive statistics (frequency, means, and standard deviations), exploratory factor analysis, and multivariate analysis are commonly employed to analyze Likert-scale data; however, they argued that these statistical techniques cannot adequately rank scale item results in an ascending or descending order.

In contrast to other statistical methods that assume a parametric distribution, Bross [62] reports on the RIDIT technique for analyzing and ranking categorical data. In addition to no requirement for the assumption of the sampling distributions, Fleiss, Levin, and Paik [63] claim that the RIDIT technique produces scores that are relative to an identified distribution. Wu [64] indicates that RIDIT analyses can be used to rank Likert-scale item results in an ascending or descending order by priority and to investigate the relationships among scale items in terms of significance level.

RIDIT scores can be derived through mathematical methods. Bhattacharya and Kumar [20] demonstrate that RIDIT analysis can be performed using a straightforward format in Microsoft EXCEL software (Microsoft Office 2019) without the need for complex computation. On the basis of the descriptions provided by Jansen [65], the basic concept of RIDIT analysis is as follows. Ordinal data are converted to a probability scale through a procedure similar to that for a Wilcoxon rank-sum test. The mean statistics obtained through RIDIT analyses and a Wilcoxon rank-sum test are directly connected. The hypothesis that no significant differences exist between mean RIDIT statistics across all scales of responses by using the χ² statistic can be tested [63]. A similar test between any group and the reference group was defined by them.

According to Pouplard, Qannari, and Simon [61], with a reference distribution function, it is easy to perform the calculations of the RIDIT values that are related to the cumulative distribution function associated with the probability function. Statistically, the mean RIDIT value is 0.5. On average, the items rated by respondents as being more important (i.e., scores of 4 or 5) tend to exhibit a RIDIT value > 0.5, whereas those rated as being less important (i.e., scores of 1 or 2) tend to exhibit a RIDIT value < 0.5. Accordingly, a higher RIDIT value for a given item generally indicates that the item is more prioritized by a given sample. Priority rankings can be generated because the items with the highest RIDIT values are the most prioritized items. Jansen [65] states that RIDIT analysis can be applied to samples of all sizes; furthermore, it can be easily applied to identify high levels of variance or homogeneity in a given response data set.

For the applications of RIDIT analysis in the field of travel and tourism, Bhattacharya and Kumar [20] suggest that RIDIT analysis can be applied across numerous fields (e.g., tourism management) to statistically analyze tourist preferences. In this study, it is well suited for ranking the preference of young tourists for accommodation attributes without making any assumption regarding the sample distribution.

3.2. Questionnaire Design

For achieving our research objective and for data collection, we developed a questionnaire comprising two parts. The first part included questions for evaluating travel characteristics, and the second part comprised 18 questions for assessing accommodation attributes. The measurement items of the questionnaire were mainly adapted from the instruments used in other studies (e.g., [10,26]). Respondents rated attributes (including facilities and specific features) that influenced their accommodation choices on a Likert scale (see Figure 1 for a description of the attributes), with endpoints rating from 1 (not at all important) to 5 (extremely important), thereby indicating how influential each accommodation attribute was from their perspective. To ensure clarity and readability, the 18 items of the second part of the questionnaire were pre-tested in a sample of 30 people. A criterion was established to eliminate question items with average scores of 3.0 and standard deviation values > 1.5. On the basis of the pre-test result, all 18 items were retained.

3.3. Data Collection

In the present study, a traditional paper-based survey was conducted to collect data. Because of practical considerations and budget constraints, all participants were recruited using a convenience sampling method from a university in central Taiwan with which the authors were affiliated as an employee or student. Approval for and assistance with data collection were obtained from faculty colleagues, and data collection was conducted in the classroom setting between 23 February and 15 March 2021. After consenting to participate, the respondents were informed about the research objectives, procedures, and the related notifications. The respondents were requested to answer questions about their most recent domestic travel and accommodation arrangements and the factors that they considered important when making their accommodation choices. The sample was either the students who had accommodation experiences during the period of the COVID-19 pandemic or were willing to consider domestic accommodation in the near future. After incomplete questionnaire responses were excluded, a total of 296 valid responses were included in the subsequent analyses.

Table 1. Respondent profiles for travel and accommodation characteristics (N = 296).

Variables	Frequency (%)	Variables	Frequency (%)
Gender		Accommodation choice for the recent trip
Male	95 (32.1%)	Hotel	100 (33.8%)
Female	201 (67.9%)	B & B	93 (31.4%)
Most recent trip 1		Airbnb	9 (3%)
0.5 months or less	121 (40.9%)	Living in relative/friend’s home	18 (6.1%)
0.5–1 month	40 (13.5%)	No accommodation	76 (25.7%)
1–3 months	49 (16.6%)	Traveling companions for the recent trip
3–6 months	38 (12.8%)	Alone	5 (1.7%)
6 months or more	46 (15.5%)	Family members	102 (34.5%)
Number of days in the recent trip 2		Relatives/friends	159 (53.6%)
2-days-1-night	105 (35.5%)	Classmates	18 (6.1%)
3-days-2-night	81 (27.4%)	Others	12 (4.1%)
4-days-3-night	17 (5.7%)	Type of travel 3
5 days or more	9 (3%)	Self-independent	266 (89.9%)
1-day	83 (28%)	Package by travel agency	5 (1.7%)
Ever used accommodation type 4		Arrangement by travel agent	6 (2.0%)
Hotel	268 (90.5%)	Incentive travel	18 (6.1%)
B & B	266 (89.9%)
Airbnb	55 (11.8%)

Note: ¹, ², and ³ represent missing data and accounted for 0.7%, 0.3%, and 0.3%, respectively. The number ⁴ represents multiple choices.

presents the descriptions and distributions of all relevant respondent variables.

4. Results

In the present study, RIDIT analysis was performed as per the steps outlined by Wu [64]. Bhattacharya and Kumar [20] demonstrate how Microsoft Excel functions can be used to calculate RIDIT values. In the Kruskal–Wallis test, the calculated W value was 1273.56, which was significantly higher than the χ² value (18–1) of 27.59; this result indicated that the responses regarding the attributes varied statistically among the respondents. This test is nonparametric and based on rank data. However, it does not require the data to exhibit a normal distribution. In the present study, this test was used to identify significant differences in the results for accommodation attributes across five categories of scores. The overall ranking of the attribute results is presented in

Table 2. Calculation of the RIDIT values for the service attributes of accommodation choice with respect to their importance and prioritization.

	1	2	3	4	5	ρi	Lower Bound	Upper Bound	Priority Ranking
Cleanliness of rooms (1)	0	0	0.005	0.057	0.685	0.748	0.714	0.781	1
Located in a safe neighborhood/feeling safe in the room (2)	0	0	0.011	0.098	0.583	0.692	0.659	0.726	2
Close to scenic area for meeting the trip requirements (3)	0	0.001	0.014	0.183	0.413	0.611	0.577	0.644	3
Providing self-catering facilities with good levels of comfort & amenities (4)	0	0.001	0.019	0.180	0.399	0.598	0.565	0.632	4
Security of payment (5)	0	0.001	0.027	0.135	0.435	0.598	0.564	0.632	5
The value derived for money spent (6)	0	0.001	0.021	0.172	0.399	0.593	0.560	0.627	6
Convenient transportation (7)	0	0.002	0.021	0.158	0.391	0.572	0.538	0.605	7
Guests’ word-of-mouth & recommendation (8)	0	0.001	0.022	0.191	0.349	0.564	0.530	0.598	8
Service staff friendly & polite (9)	0	0.001	0.019	0.219	0.325	0.563	0.530	0.597	9
Having specific architecture or appealing decorative design (10)	0	0.002	0.045	0.197	0.231	0.475	0.442	0.509	10
Having single/double/twin accommodation available (11)	0	0.003	0.042	0.194	0.234	0.472	0.438	0.506	11
Provision of a comfortable ambiance (12)	0	0.002	0.047	0.186	0.231	0.467	0.433	0.500	12
Provision of meals (13)	0	0.002	0.050	0.212	0.176	0.440	0.406	0.473	13
Close to site or event attractions (14)	0	0.003	0.060	0.169	0.198	0.430	0.396	0.463	14
Unique accommodation experiences (15)	0	0.003	0.067	0.152	0.184	0.407	0.373	0.440	15
Interaction with B & B host or landlord (16)	0	0.007	0.074	0.136	0.107	0.325	0.291	0.359	16
Provision of a local trip (17)	0	0.010	0.081	0.109	0.063	0.263	0.230	0.297	17
Interaction with other guests (18)	0.001	0.015	0.068	0.064	0.036	0.183	0.150	0.217	18

. Among the examined attributes (Figure 1), “Cleanliness of rooms” was the most prioritized attribute by the respondents, followed by “Located in a safe neighborhood/feeling safe in the room” and “Close to scenic area for meeting the trip requirement”. The least prioritized attribute was “Interaction with other guests”. These RIDIT results highlight the relative importance of each attribute to the respondents.

The most-used ranking technique involves the application of simple mean ranking, which allows for statistics such as mean and variance to be computed and for service attribute-related results to be compared.

Table 3. Comparison of traditional means and RIDITs for service attributes.

Attributes	Mean	SD	RIDITs
(1)	3.828	0.864	0.748
(2)	4.247	0.748	0.692
(3)	3.696	0.929	0.611
(4)	4.392	0.719	0.598
(5)	3.784	0.902	0.598
(6)	4.334	0.759	0.593
(7)	4.236	0.806	0.572
(8)	3.372	0.959	0.564
(9)	3.922	0.904	0.563
(10)	4.797	0.527	0.475
(11)	4.351	0.744	0.472
(12)	3.916	0.877	0.467
(13)	4.628	0.672	0.440
(14)	4.334	0.827	0.430
(15)	3.956	0.833	0.407
(16)	3.105	0.956	0.325
(17)	2.642	1.022	0.263
(18)	4.226	0.931	0.183

Note: Please refer to Table 2 for the abbreviations of the service attributes represented.

provides a comparison of the means and RIDIT results of all examined attributes. Notably, the two methods differ in terms of the rankings of the key service attributes. For instance, when simple mean ranking was employed, “Cleanliness of rooms”, “Close to scenic area for meeting the trip requirement”, “Security of payment”, and “Guests’ word-of-mouth & recommendation” were perceived to be less important by the respondents. In addition, when simple mean ranking was employed, it appears that “Having specific architecture or appealing decorative design”, “Provision of meals”, “Having single/double/twin accommodation available”, and “Close to site or event attractions” were perceived as being important by the respondents; however, the corresponding RIDIT values were low. The discrepancy between the rankings of service attributes conducted through these two methods can be explained by the high sensitivity of the attribute scoring process to small changes in data (e.g., the presence of outliers), which was due to the assignment of equally spaced scores [20].

The RIDIT analysis results (including ranking results) were also compared with those obtained through exploratory factor analysis (EFA), which is a commonly used technique. To identify the mechanisms underlying the variance in the service attribute results, we performed principal component analysis with orthogonal varimax rotation.

Table 4. Comparison of factor analysis, traditional means, and RIDITs for service attributes.

Attributes	Factors and the Related Loadings					Mean	SD	RIDITs
	F1	F2	F3	F4	F5
(7)	0.796					4.236	0.806	0.572
(10)	0.752					4.797	0.527	0.475
(12)	0.738					3.916	0.877	0.467
(4)	0.697					4.392	0.719	0.598
(18)		0.810				4.226	0.931	0.183
(16)		0.761				3.105	0.956	0.325
(17)		0.709				2.642	1.022	0.263
(15)		0.518				3.956	0.833	0.407
(14)			0.738			4.334	0.827	0.430
(9)			0.662			3.922	0.904	0.563
(3)				0.782		3.696	0.929	0.611
(13)				0.768		4.628	0.627	0.440
(11)				0.617		4.351	0.744	0.472
(8)					0.711	2.372	0.959	0.564
(2)					0.652	4.247	0.748	0.692
(5)					0.631	3.784	0.902	0.598
(1)					0.620	3.828	0.864	0.748
(6)					0.600	4.334	0.759	0.593
Eigen values	5.169	2.043	1.172	1.123	1.025
Cronbach’s α	0.754	0.720	0.514	0.613	0.697
Variance explained (%)	28.719	11.348	6.512	6.239	5.693
Accumulated variance explained (%)	28.719	40.067	46.579	52.818	58.511
KMO	0.845

Note: Please refer to Table 2 for the abbreviations of the service attributes represented. F1 = accommodation environment; F2 = interactive experience; F3 = hospitality services; F4 = geographic location; F5 = cost-effectiveness and safety.

lists the results of the EFA. Regarding adequacy and global significance, the results of the Bartlett test of sphericity and the Kaiser–Meyer–Olkin test were significant (0.845). Five factors corresponding to eighteen items were extracted, and these factors accounted for 58.5% of the explained variance. The five factors retrieved were as follows: accommodation environment, interactive experience, hospitality service, geographic location, and cost-effectiveness and safety. However, providing appropriate labels for each factor was challenging because of the varying nature of these attributes (e.g., geographic location). Indeed, although the EFA technique could group the number of scales into a smaller subset/dimension, the reduced scales are difficult to implement due to its hypothetical nature [20].

When the attributes were grouped under a given factor (e.g., interactive experience and cost-effectiveness and safety), a priority ranking was allotted to each individual item based on the RIDIT analysis results (see the furthest right column in Table 4). We attempted to find out the extent to which the items grouped into each particular factor tended to have low or high priority rankings. However, an analysis interpretation was difficult to carry out. The most prioritized attributes in the RIDIT analysis results, namely “Cleanliness of rooms” and “Located in a safe neighborhood/feeling safe in the room”, corresponded to Factor 5, which explained only 6.5% of the variance. On the contrary, the less important attributes, namely “Having specific architecture or appealing decorative design” and “Provision of a comfortable ambiance”, corresponded to Factor 1, which explained 28.7% of the variance. Compared with EFA, RIDIT analysis leads to more straightforward answers. Thus, the present study illustrates that RIDIT analysis may be an effective method for ranking items in an ascending or descending order.

EFA is criticized as being highly sensitive to variable selection, and different samples may generate different results [20]. We further employed RIDIT analysis to test the hypotheses as to whether the female and male respondents place the same importance on the accommodation service attributes. In the Kruskal–Wallis test, the calculated W values were 912.12 and 377.82, respectively, which were significantly higher than the χ² value (18-1) of 27.59 at the level of significance α = 0.05; these results indicated that the responses regarding the service attributes varied statistically among the two groups of the respondents. The overall ranking of the attribute results is summarized in

Table 5. Comparison of the RIDIT values for service attributes with respect to the female and male correspondence.

Priority	Female (N = 201)		Male (N = 95)
Ranking	Attributes	RIDITs	Attributes	RIDITs
1	(1)	0.745	(1)	0.753
2	(2)	0.683	(2)	0.711
3	(3)	0.624	(4)	0.626
4	(5)	0.606	(3)	0.582
5	(6)	0.599	(5)	0.581
6	(8)	0.595	(6)	0.580
7	(7)	0.589	(9)	0.579
8	(4)	0.585	(7)	0.534
9	(9)	0.556	(8)	0.498
10	(11)	0.447	(10)	0.487
11	(10)	0.470	(12)	0.470
12	(12)	0.465	(14)	0.470
13	(13)	0.437	(11)	0.461
14	(14)	0.410	(13)	0.447
15	(15)	0.406	(15)	0.409
16	(16)	0.323	(16)	0.330
17	(17)	0.257	(17)	0.227
18	(18)	0.173	(18)	0.206

Note: Please refer to Table 2 for the abbreviations of the service attributes represented.

. However, there was not much difference in the ranking of the attributes between two sub-samples. Among the examined attributes, “Cleanliness of rooms” was the most prioritized attribute by the respondents, followed by “Located in a safe neighborhood/feeling safe in the room”. The four least prioritized attributes were “Interaction with other guests”, followed by “Provision of a local trip”, then “Interaction with B&B host or landlord”, and “Unique accommodation experiences”. These results were consistent with the analysis results of the overall sample. Thus, RIDIT analysis is less sensitive to variable selection, and different samples could generate similar results.

5. Discussions and Implications

Our research results reveal that RIDIT analysis is a straightforward method for ranking service attributes measured on a Likert scale. Our research findings are similar to those of Nash, Thyne, and Davies [10], who report that a sample of backpackers gave high rankings to the attributes “cleanliness of rooms”, “value for money”, and “the presence of self-catering facilities”. However, the rankings of the factors by priority are different between the two studies. The service attributes identified in the present study reflect the characteristics, expectations, and needs of post-millennial tourists when choosing domestic accommodations. On the other hand, despite a different sample, similar results were found between Spoerr’s [18] and our studies. For Airbnb accommodations, Sthapit, Björk, and Jiménez Barreto [66] indicate that young travelers emphasize the promised quality of accommodations and the actual conditions. In view of the study of Chiang and Huang [67], bathroom quality was the top hotel attribute that tourists were found to rate negatively among online reviews of Taipei’s economy hotels, reflecting the importance of room cleanliness to Taiwanese travelers.

Likewise, Del Chiappa, Pung, and Atzeni’s study [54] indicates that the first key factor for accommodation is sanitization and the use of personal protective equipment (PPE) after COVID-19. The findings of Ye et al. [19] are similar to our findings on the preferences of domestic tourists for four accommodation attributes (i.e., economy, convenience, familiarity, and coziness) after the COVID-19 pandemic. In addition, the respondents in the present study indicated that they preferred accommodations that were located close to scenic areas and had convenient transportation options because of their limited travel budget and lack of a car. Our results also indicated that the respondents strongly preferred self-catering facilities, and that service staff did not significantly influence their choice of domestic accommodation.

A notable finding of the present study is that social interactions with hosts or other guests was the least important attribute for domestic young tourists when selecting accommodations. These findings contradict those of other studies on Airbnb accommodations or on young international travelers [29,31]. However, our finding is similar to that of Ye et al. [19]; that is, social interactions are not a key factor influencing the P2P accommodation choices of consumers. Moreover, our findings echo those of Santos et al. [53], who report that Airbnb users are the least concerned about experiencing local culture, interacting with other guests, and living like a local. These findings suggest that domestic travelers are not focused on making new friends because most of them have travel companions and do not feel the need to meet new people. In summary, the considerations for selecting domestic accommodations may differ from those for selecting overseas accommodations. Because youth are highly familiar with their local environment, they may place more emphasis on the quality of basic lodging services when they are traveling domestically. Therefore, lodging managers should improve service efficiency, quality, and performance. To this end, understanding the service attributes prioritized by travelers may help these managers to develop plans for improvements and to identify areas requiring improvement. These adjustments can also improve their overall organizational efficiency.

With young travelers turning to domestic tourism during the post-COVID 19 era, precise techniques and effective positioning strategies can be applied to adapt to the dynamic and competitive accommodation market. Various scholars have discussed the application of artificial intelligence (AI) to hotel operation and management [68]. This research trend aligns with the considerable development of AI technology, which has led to the development of robot-run hotels that provide the advantages of high efficiency, low labor cost, and novelty. Zhang and Qi [69] discover that the factors associated with education, attitude, and income level strongly influenced the intention of travelers to stay in robot-run hotels. Young travelers may derive the most enjoyment from automated services in hotels and may acquire positive experiences during the consumption process.

6. Conclusions

The present study differs from previous studies in that it employed RIDIT analysis to identify the attributes prioritized by young tourists with respect to their accommodation choices, what might be important for guests, and what certain guests perceive as important. In our RIDIT analytical results, respondents’ preferences are sorted in descending importance. Of the examined attributes, “Cleanliness of rooms” is the attribute most prioritized by the students, followed by “Located in a safe neighborhood/feeling safe in the room” and “Close to scenic area for meeting the trip requirement”. The attribute least prioritized by the students is “Interaction with other guests”. This analysis provides conclusions that differ from those obtained through commonly used techniques including the arithmetic average and exploratory factor analysis.

Our findings clarify how RIDIT analysis can be applied to study the opinions and preferences of prospective tourists, which are useful to hoteliers, as providing clear indications on how to improve certain quality aspects to meet guests’ requirements. A major advantage of RIDIT analysis is that it can measure the importance of attribute items on a Likert scale without the need for assuming a specific distribution (usually a normal distribution) and can test whether data points conform to a distribution assumption. The simple mean ranks could be affected by the presence of skewed data [20]. Another advantage of RIDIT analysis is that it allows researchers to conduct hypothesis testing based on the Kruskal–Wallis test. RIDIT is a very useful data analytical technique for processing a large number of tourists’ data, which provides researchers/practitioners with a simplified means of ranking the items measured on Likert-scale items. If the EFA results may not easily lead to decision making in relation to the allocation of resources to improve quality, the RIDIT values enable a manager to determine which quality attributes are the most important in customers’ evaluations and how the relative weight of each could be calculated. Therefore, the application of RIDIT analysis can identify the rank-wise positions of the preferences of tourists whenever required. It is expected that there will be greater use of this technique in service quality gap analysis or in service quality benchmarking in the tourism and travel industry. We also hope there will be more discussions on statistical ranking or ordering techniques to obtain better insight into tourists’ minds.

A major limitation of the present study is that the findings cannot be generalized because the study sample pertains to a specific geographical area. Additionally, the outputs of RIDIT analysis may vary with the different sample sizes. Future studies should confirm our findings through other samples and expand the geographical scope of their research to achieve more comprehensive conclusions. A second limitation is that the factors identified in the present study were not classified by accommodation type, and that several crucial factors may have been neglected because of infrastructural restrictions. Future studies should conduct more extensive literature reviews to identify other key factors of service attributes that significantly influence accommodation choices. Another limitation pertains to the sample domain selected for the present study. Future studies should explore other significant factors and identify their influence among other youth segments with varying socio-economic and demographic characteristics. Although the unequal size of male and female samples generated the similar RIDIT analysis, what we did not know was the extent to which the other variables may cause us to doubt the results. This suggests a need for further analysis to hold other factors constant, which will be the subject of future research on this topic. In addition, it may be beneficial to address the potential influence of the health factor among students. Finally, additional work may expand further studies using RIDIT and Grey Relational Analysis (GRA) approaches. Based on Kumar and Bhattacharyya [70] and Sharma et al. [71], these two independent methodologies could be applied to prioritize the scale items to verify the high correlation between the rankings.