The Relationship Between the Hotel Rating System, Service Quality Improvement, and Hotel Performance Changes: A Canonical Analysis of Hotels in Thailand

Abstract

This study investigates the relationship between the hotel rating system, service quality improvement, and hotel performance changes in Thailand. A total of 306 hotel managers participated in the study and shared their perceptions of the impacts of the Thailand Hotel Standard on hotel service quality improvement. Four dimensions in service quality improvement were identified: service delivery, hotel employees, guest facilities and surroundings, and prestige. The four dimensions of service quality improvement were not significantly associated with hotels' star levels. The canonical correlation analysis found two significant relationships between service quality improvement and changes in hotel performance.

Keywords:

• Hotel rating systems
• service quality
• hotel performance
• Thailand Hotel Standard
• hotels in Thailand

INTRODUCTION

Over the decades, tourism has been one of the major sectors contributing to economic prosperity in Thailand. With the exception of a sharp decline in 2003 due to fears over the Severe Acute Respiratory Syndrome (SARS) and the war in Iraq (Tourism Authority of Thailand [TAT], 2004a; TAT, 2004b), the number of international tourist arrivals to Thailand has continuously increased since the Gulf War in 1991. According to the World Tourism Organization, in 2003 Thailand had the third largest accommodation capacity in Asia with a total of 321,000 guestrooms in hotels and similar establishments (World Tourism Organization [WTO], n.d.). The growth of the Thai tourism industry has attracted investments not only from international chain hotels, but also from local entrepreneurs.

Despite the rapid expansion of the hotel industry, Thailand has been one of the few major travel destination countries without a national rating program to reflect tangible and intangible characteristics of a hotel property, and to offer a clear indicator of its marketing value till 2004. The development of the Thailand Hotel Standard was expected to raise Thai hotels' overall service quality by benchmarking and aligning local standards with international standards and practices (Thai Hotel Association [THA], Tourism Authority of Thailand [TAT], & Association of Thai Travel Agents [ATTA], n.d). It should also provide a common ground to grant equal recognition to both Thai‐owned and ‐managed hotels and to international chain hotels. Customers and tour operators can identify the levels of hotel service and evaluate the service performance of a hotel against their expectations (TAT, n.d.).

The Foundation of Standard and Human Resources Development in the Service and Tourism Industry, a nonprofit organization, was founded to administer all aspects of implementing the Thailand Hotel Standard (THA, TAT, & ATTA, n.d.). The assessment committee—consisting of representatives from the Thai Hotel Association, the Tourism Authority of Thailand, the Association of Thai Travel Agents, and hotel management programs in academic institutions—conducts assessment and certification annually. Participation in the program is voluntary. A qualified hotel property must have a valid hotel operation license and have been in operation in Thailand for at least a year. Each hotel is inspected and scored between one and five stars, based on three criteria: construction and facilities, maintenance, and service. If the total score of the three criteria exceeds the minimum requirement, a hotel will be certified with the corresponding star level. If failed, a hotel will be offered three alternatives: cancellation of the application, acceptance of a lower star level, or improvement within 180 days for re‐inspection. Since the official launch of the Thailand Hotel Standard program in 2004, a total of 142 hotels were certified by 2006 (THA, n.d.).

As the Thailand Hotel Standard program is still in its infancy, little is known about its impacts on the hotel industry in Thailand. The purpose of this study is to investigate the perceived impacts of the hotel rating system on hotels' service quality improvement and performance changes. As hotel managers are usually the key decision‐makers to participate in the program, it is important to examine their perceptions of the hotel rating system and its relationship with service quality improvement and hotel performance changes. The specific objectives of the study are as follows:

1. To identify the dimensions of service quality improvement as a result of the introduction of the Thailand Hotel Standard;

2. To examine the relationship between service quality improvement and hotel star rating levels; and

3. To examine the relationship between service quality improvement and hotel performance changes before and after the introduction of the Thailand Hotel Standard.

HOTEL RATING SYSTEMS

According to the World Tourism Organization and the International Hotel and Restaurant Association, hotel rating is the classification of accommodation establishments denoting a system, duly published, “in which accommodation establishments of the same type (e.g., hotels, motels, and inns) have been conventionally broken down into classes, categories, or grades according to their common physical and service characteristics and established at government, industry or other private levels” (World Tourism Organization [WTO] & International Hotel and Restaurant Association [IH&RA], 2004, p.9). The European Standardization Committee (CEN, the Comité Européen de Normalisation) clarifies the term “accommodation rating or classification scheme” as “a system providing an assessment of the quality standards and provision of facility and/or service of tourist accommodation, typically within five categories, often indicated by one to five symbols” (WTO & IH&RA, 2004, p.68).

Hotel rating systems offer benefits to various sectors—travel agencies, tour operators, hotels, governments, and consumers (WTO & IH&RA, 2004). The systems facilitate travel agents' selection of hotels for their customers. Major tour operators, such as First Choice, Thomson Holidays, Airtours, and Thomas Cook, have their own hotel classification systems to assist product packaging and marketing. Hotel companies use classification schemes as “branding” (WTO & IH&RA, 2004), which convey both qualitative and quantitative grading of properties to their customers. The rating system provides consumers with an easy way to compare hotels. Governments use hotel ratings or classification systems to regulate the hotel industry with tariffs and taxes, and to meet basic requirements of safety and hygiene. Research found that, when choosing hotels, customers staying at three‐ to five‐star (or other equivalent symbols) hotels use rating systems more often than do those staying at one‐ and two‐star hotels (Callan, 1995).

A hotel rating system embraces two parts: a basic registration standard and a grading standard (Callan, 1993). The basic registration standard is the physical requirement that a hotel property must meet; it is the minimum quality requirement. The grading standard is an extension of basic requirements of qualitative and intangible services, allowing a hotel to be compared with other properties. To communicate the quality level a hotel achieves, a variety of grading symbols is used—such as stars, crowns, diamonds, suns, or letters. The symbol most universally recognized is stars, as most countries have at least one rating system using stars to represent quality grades (Callan, 1993; WTO & IH&RA, 2004).

The 100‐plus hotel rating systems worldwide can be categorized into two groups: official and nonofficial (Qing & Liu, 1993). Official hotel rating systems are established and conducted by government agencies and followed on a compulsory and regulatory basis, while nonofficial hotel rating systems are developed and implemented by private organizations, such as hotel or tourism associations and national or regional automobile associations, on a voluntary participation basis by hotels. The purpose of official systems is mainly to control lodging tariff and taxes, whereas nonofficial systems impose no social obligations (WTO & IH&RA, 2004). Callan (1994) specified that some hotels tend to oppose the compulsory grading scheme due to concerns over bureaucratic interference, with customers expecting guaranteed service quality out of these official systems.

China is one of the countries with an official hotel rating system, while some countries—such as the U.S. and Britain—have only nonofficial rating systems. In the U.S., the three most popular hotel rating systems are AAA (diamond rating), Mobil (star rating), and Utell (Official Hotel Guide‐OHG). In Britain, there are a number of hotel classification schemes offered by private organizations and regional tourist boards. For example, the English Tourist Board (ETB) awards crowns to hotels in England, whereas the Automobile Association (AA) and the Royal Automobile Club (RAC) rate tourist accommodations by star systems with different criteria and judgments (Conway, 2004). As a result, the same property could have multiple ratings.

In 2004, the World Tourism Organization (WTO) and the International Hotel and Restaurant Association (IH&RA) conducted an international study to survey hotel classification schemes in both public sectors, such as the National Tourism Organizations (NTOs), and private sectors, such as the National Hotel Association. The study revealed the complexity of hotel rating systems by reviewing their evaluation criteria and administration in different countries. To reduce the complexity of numerous hotel rating systems in existence, the WTO and the IH&RA have been developing a universal hotel grading scheme that can be used in countries throughout the world to benefit both customers and tourism service providers (Daily, 2004). A joint promotional campaign among AA, RAC, and ETB was also undertaken recently to develop a harmonized hotel‐classification scheme (Conway, 2004).

SERVICE QUALITY

While technology makes products similar, it is service quality that differentiates them in the market (Denburg & Kleiner, 1993). The importance of service quality is well recognized in the hospitality industry, since hotels cannot survive intense competition without satisfying their customers with quality service. Garvin (1988) defined perceived quality as the subjective perception of quality through indirect measures of quality comparison. Gronroos (1993) stated that service quality was developed based on the confirmation/disconfirmation concept in the perceived service quality model introduced in 1982. The notion of the model explains that perceived service quality is the result of comparing a consumer's real experience with his/her expectation of the service.

To achieve high service quality, a company needs to understand its customers' expectations. Service improvement programs can be developed, which include issues of customer segmentation, service culture, communication to customers, recruitment and training of service personnel, empowerment of employees, and appraisal systems. The essentials of service quality are rationalized as enhancement in customer loyalty, increase in market share, higher returns to investors, reduction in costs, lowering vulnerability to price competition, and establishment of a competitive edge (Knutson, 1988; Haksever, Render, Russell, & Murdick, 2000; Wuest, 2001).

Based on the theory of perceived service quality, Parasuraman, Zeithaml, and Berry (1985) proposed their “conceptual gap” model of service quality and explained that perception of service quality is a comparison of consumers' expectations with actual service performance and quality evaluation involves evaluation of both process and outcome of service delivery. In their succeeding studies, SERVQUAL, a scale for measuring consumer perceptions of service quality, was proposed (Parasuraman, Zeithaml & Berry, 1985, 1988, 1991). The 7‐point Likert scale, which is regarded as a generic service quality measurement tool for service industries (Lovelock, 2001), measures customer expectations and perceptions with 22 items in five dimensions: reliability, responsiveness, assurance, empathy, and tangibles (Parasuraman et al., 1988). Reliability is defined as the ability to perform the promised service dependably and accurately. Responsiveness is defined as willingness to help customers and provision of prompt service. Assurance is defined as the knowledge and courtesy of employees and their ability to inspire trust and confidence. Empathy is defined as caring and individualized attention to customers. Tangibles are physical facilities, equipment, and appearance of personnel. Despite the critiques on its validity (Carmen, 1990; Teas, 1993, 1994), SERVQUAL has been widely applied and well accepted in various service industries (Brady, Cronin, & Brand, 2002; Cronin & Taylor, 1992 & 1994; Hoffman & Bateson, 1997).

In the context of hospitality management, the measurement of service quality was derived from the concept and studies of service quality in general business. A number of studies applied or modified the SERVQUAL instrument to measure service quality in the hospitality industry. LODGSERV, a modified instrument of SERVQUAL, was developed to measure hotel guests' expectations and experience of service quality (Knutson, Stevens, Wulaert, Patton, & Yokoyama, 1991). HOLSERV, an extension of the SERVQUAL scale, was introduced to the hospitality industry in 1999 (Mei, Dean, & White, 1999). Ekinci and Riley (1999) proposed the application of the Q‐sort technique in the context of hotel service quality evaluation to validate the dimensions of the established models of service quality. Using the same approach of SERVQUAL, the Lodging Quality Index (LQI) was developed to measure customers' perception of service performance (Getty & Getty, 2003).

HOTEL RATING SYSTEMS, SERVICE QUALITY, AND HOTEL PERFORMANCE

Competitive marketing demands local and international hotels to seek standards and tools to reflect their service quality. One of the answers to this need is a reliable hotel rating system, which ranks, certifies, and reflects the quality of hotels' facilities and service. Hotel rating systems are vital to hotel marketing. A survey conducted by Callan (1989) found that hotel operators valued the classification awards and grading as promotional assets, and that the classification schemes provided significant amount of business to the hotels. Furthermore, Vallen and Vallen (2005) noted an approximately 20% increase in sales among top‐rated hotels in the Mobile Guide and a 40% increase in business among small hotels rated in AAA after hotels were awarded with classifications. Therefore, hotel ratings functioning as a promotional tool might be associated with growth in hotel business measured by volume of sales and occupancy rate.

Hotels also perceive ratings as a pricing tool. Israeli and Uriely (2000) examined the impact of star ratings and corporate affiliations on hotel room prices in Israel. It was found that price variations could be partially explained by star ratings and the star rating system was a significant predictor of hotels' decision in setting prices. Lollar's (1990) exploratory study further supported this conclusion. It was found that a hotel charged more once a higher level of classification was awarded and in some countries four‐ or five‐star hotels could charge any price at their will. This research proposes that changes in hotel performance are associated with service quality improvement as a result of participating in the hotel rating system.

Fernandez and Bedia (2004) examined whether a hotel rating system was a good indicator of its service quality and found that, based on values of expectations and perceptions and their differences, the ranking of the hotel groups did not correspond exactly to the ranking of stars. They also found that customers from highly rated hotels had negative differences between perceptions and expectations indicating a demanding clientele.

In summary, the hotel rating system is one of many instruments that the hotel industry uses as a guideline to reach an expected level of service quality and to convey this message to the public. A conceptual framework to examine the relationship among hotel rating systems, service quality, and hotel performance is proposed in Figure 1. Hotel rating systems are related to service quality improvement, while service quality improvement is associated with changes in hotel performance. It is suggested that hotel rating systems encourage hotel operators to improve their service quality, which may lead to changes in hotel performance.

Figure 1 Proposed Conceptual Framework

METHODOLOGY

This exploratory study attempted to examine the relationship between the hotel rating system, service quality improvement, and hotel performance changes. Data were obtained by conducting a cross‐sectional survey of hotel managers in Thailand between April and May, 2006.

Sampling Plan

The target population of the study was general managers of hotel properties with at least 1 year of operation in Thailand. According to the hotel directory of the Tourism Authority of Thailand (2004), there were approximately 5000 hotels and similar establishments in Thailand. The sampling frame was derived from three sources: the directory of hotels with the Thailand Hotel Standard Certification, the hotel membership directory of the Thai Hotel Association, and the accommodation directory of the Tourism Authority of Thailand. The sampling method exercised a census and a simple random sampling. The sample size was 1500, including 119 hotels certified by the Thailand Hotel Standard till 2004, 337 THA hotel members, and a random sample of 1044 non‐THA members listed in the TAT hotel directory. The three lists were cross‐examined and purged to avoid redundancy.

Data collection was administered through mail surveys and telephone follow‐ups. After the distribution of 1500 questionnaires, 354 completed surveys were received, yielding a response rate of 23.6%. After an exclusion of questionnaires with missing data, a total of 306 valid surveys were used. The instrument used in this study was self‐administered questionnaires with two sections—hotel managers' perceived impacts of the Thailand Hotel Standard and general information about the hotel property.

Content Validity and Reliability

To ensure content validity, an extensive review of the hospitality literature was conducted to find the important variables related to the constructs of both service quality and hotel rating systems, which were then reviewed by professors and professionals in hotel management. Corrections and adjustments were made according to their suggestions. A pilot study of 75 hotel managers was conducted. Cronbach's alpha values for all scales were greater than 0.7, indicating that measurements were sufficiently reliable and internally consistent.

Data Analysis

Data analysis consisted of four stages. In the first stage, the sample was profiled using descriptive statistics. Second, the underlying dimensions of service quality improvement were identified using exploratory factor analysis. Principal component analysis was applied, and latent root criteria or eigenvalues greater than 1 were used. Third, differences regarding service quality improvement were examined based on hotels' application and certification status. Independent sample t‐tests and analysis of variances were used. Finally, the relationship between service quality improvement and performance changes was explored by using canonical correlation analysis, a generalized multivariate collective analysis to examine the relationship between multiple dependent variables and multiple independent variables (Hair, Anderson, Tatham, & Black, 1998). Whereas multiple regression predicts a single dependent variable from a set of multiple independent variables, canonical correlation allows simultaneous prediction of multiple dependent variables of hotel performance changes from multiple independent variables of service quality improvement.

FINDINGS AND RESULTS

Respondents' Profile

As shown in Table 1, among the 306 respondents, over half (55%) were downtown business hotels primarily serving business travelers, and the remainder (45%) were resorts close to tourist attractions targeting mainly pleasure travelers. About 80% of the hotels were independent properties, while the remainder (20%) were hotels operated under chain brands. Among the 64 chain‐affiliated hotels, 21 properties were under international chain brands and the others were associated with Thai brands, accounting for 6.9% and 14.1% of the total respondents, respectively. The size of the hotel properties ranged from 3 to 1200 guest rooms. More than half had fewer than 100 guest rooms. The hotels were in business for at least 1 year as required by this study, and the longest period of operation was 46 years.

Table 1. Respondents' Profile

	Respondents, n	Min	Max	Mean
Number of rooms in the hotel	300	3	1,200	135.2
Length of hotel operation (in years)	287	1	46	13.7
	Frequency		Percent (%)
Hotel category	N = 306		100.0
Business	168		54.9
Resorts	138		45.1
Independent vs. chain affiliation	N = 306		100.0
International hotel chain	21		6.9
Thai hotel chain	43		14.1
Independent hotels	242		79.0
Awareness of the Thailand Hotel Standard	N = 307		100.0
Yes	268		87.3
No	39		12.7
Application for the Thailand Hotel Standard	N = 307		100.0
Yes	78		25.4
No	229		74.6
Proposed star level	N = 78		100.0
One star	1		1.3
Two star	7		9.0
Three star	21		26.9
Four star	38		48.7
Five star	11		14.1
Star rating certification for the proposed level	N = 78		100.0
Yes	62		79.5
No	16		20.5
Application plan of the hotels which did not apply for the hotel rating	N = 228		100.0
Never	85		37.3
Within 3 years	94		41.2
More than 3 years	44		19.3
Not sure	5		2.2

In terms of involvement with the hotel rating system, most respondents (87%) were aware of the Thailand Hotel Standard, but only a quarter had ever applied. The majority of these applicants requested three‐ to five‐star evaluation, and 80% were awarded with corresponding rating certifications. Among the 228 respondents that never applied for certification, over 40% intended to apply within three years and about 20% planned to apply in three years. Approximately 37% indicated no intention to apply, while 2% had not decided.

Factor Analysis of Service Quality Improvement

The study intended to identify the dimensions of service quality improvement as a result of implementing the Thailand Hotel Standard. The participants rated their perceived extent to which the service quality attributes have been changed (or may be changed) due to an application for certification of the hotel rating under the Thailand Hotel Standard. The attributes were measured on a 7‐point Likert scale with 0 indicating no change, 1 indicating minimum changes, and 7 indicating maximum changes. Besides the 78 certified hotels, 94 hotels in the application process—which indicated that changes were made due to the introduction of the Thailand Hotel Stardard—were included in this analysis. These selected cases were representative of hotels with service quality most likely to be affected by the Thailand Hotel Standard.

The assumptions of factor analysis were checked by the Bartlett test of sphericity and the measure of sampling adequacy (MSA). The consideration of practical and statistical significance resulted in retaining 42 variables and disregarding 12 variables due to low factor loadings and cross loadings. As shown in Table 2, four factors of service quality improvement with eigenvalues above one were identified, which accounted for 81.04% of the total variance. VARIMAX orthogonal rotation was used as it provides clear separation of factors. Based on the representative items, the four factors were named as “Service Delivery,” “Hotel Employees,” “Facilities and Surroundings,” and “Prestige.” The reliability coefficients of the four factors ranged from 0.85 to 0.99.

Table 2. Dimensions of Service Quality Improvement

Attributes	Factor Loadings	Eigen‐ value	Variance Explained	Reliability (Cronbach α)
Factor 1: Service Delivery		15.7	37.4	0.99
Enabling guests to feel safe about their transactions	0.903
Showing willingness to help guests	0.870
Providing error‐free service	0.862
Consistent staff courtesy to guests	0.856
Providing services at agreed time without delay	0.846
Telling guests exactly when services will be performed	0.843
Having guests' best interests in mind	0.842
Delivering consistent service	0.824
Offering convenient service hours	0.823
Providing prompt service to guests	0.820
Understanding guests' specific needs	0.804
Exhibiting caring attitudes when solving guests' problems	0.785
Staff having sufficient knowledge to answer guests' questions	0.777
Never being too busy to respond to guests' requests	0.775
Completing arrangements at agreed time without delay	0.763
Giving guests individual attention	0.760
Proper sanitation and safety	0.753
Providing services correctly without repetition	0.727
Staff giving personal attentions to guests	0.705
Providing efficient service in general	0.697
Factor 2: Hotel Employees		9.6	22.8	0.98
Improving employees' foreign language competency	0.820
Providing quality service by receptionists	0.810
Providing quality service by cashiers	0.783
Providing quality service by bellboys	0.777
Improving employees' service attitudes	0.775
Improving employees' discipline	0.770
Improving restaurant manager's and staff's up‐sale skills	0.765
Improving waiters/waitresses' service manner, discipline, language competency, and service efficiency	0.755
Improving employees' service manner	0.731
Improving cleaning service in guestrooms	0.711
Improving employees' way of dress	0.693
Factor 3: Facilities and Surroundings		5.6	13.4	0.93
Improving surroundings	0.760
Improving location/physical structure	0.723
Improving lobby	0.707
Improving guest rooms	0.652
Improving restaurants	0.648
Improving business centers	0.616
Improving parking	0.607
Improving public, health, and entertainment facilities	0.581
Factor 4: Prestige		3.2	7.5	0.85
Receiving awards from international contests	0.843
Receiving awards from domestic contests	0.810
Providing services to VIPs	0.768
Note: Total variance explained: 81.04%

Factor 1, Service Delivery, covered most statements in SERVQUAL with modifications, and the overall service efficiency—an attribute in the Thailand Hotel Standard. The factor with 20 items included characteristics of proper and effective hotel services, such as guests' feeling safe about their transactions, employees' willingness to help guests, error‐free service, courtesy to guests, personal attention to guests, and the overall service efficiency. All factor loadings were 0.697 or above. The factor explained 37.4% of total variance in the original variables, and its reliability coefficient is 0.99, an indication of internal consistency. Factor 2, Hotel Employees, described the competencies of hotel employees to deliver quality service by 11 statements in the Thailand Hotel Standard. The factor loadings were from 0.69 to 0.82. It accounted for 22.8% of the total variance, with a reliability coefficient of 0.98. Factor 3, Facilities and Surroundings, represented improvement in hotels' physical environment. The factor of eight attributes explained 13.4% of the total variance, with a reliability coefficient of 0.93. Factor 4, Prestige, had three attributes related to recognitions of hotels, including receiving awards from international contests, receiving awards from domestic contests, and providing services to VIPs. It explained 7.5% of the total variance, with a coefficient of 0.85.

Service Quality Improvement and Star Rating Levels

The relationship of service quality improvement and hotels' star rating levels was examined by using analysis of variance (ANOVA). Since no one‐star hotels passed the assessment by the Thailand Hotel Standard, four levels of star ratings, ranging from two to five, were used in the independent variable. The dependent variables were the four factors derived from factor analysis of service quality improvement: service delivery, hotel employees, facilities and surroundings, and prestige.

As shown in Table 3, our analysis showed no significant differences in the four aspects of hotels' service quality improvement due to different star ratings. However, hotels at each star level all indicated improvements in service delivery, employee competency, guest facilities and surroundings, and prestige. Therefore, the full spectrum of service quality was reinforced, regardless the star level for which a hotel applied and was certified.

Table 3. ANOVA Results of Service Quality Improvement by Star Levels

Factor	2‐star Mean(SD)	3‐star Mean(SD)	4‐star Mean(SD)	5‐star Mean(SD)	F	Sign
Factor 1: Service delivery	5.12 (2.2)	4.85 (1.6)	4.55 (2.1)	4.26 (2.8)	0.307	0.820
Factor 2 *: Hotel employees	5.75 (1.2)	4.92 (1.5)	5.05 (1.9)	4.39 (2.7)	0.991	0.416
Factor 3 *: Facilities & surroundings	4.05 (1.0)	4.43 (1.6)	4.83 (1.8)	3.36 (2.4)	1.479	0.248
Factor 4: Prestige	3.47 (1.6)	3.71 (2.0)	3.81 (1.8)	3.48 (2.3)	0.117	0.950
Note: *Equal variances not assumed.

Service Quality Improvement and Hotel Performance Changes

The relationship between hotel performance changes and service quality improvement due to the introduction of the Thailand Hotel Standard was explored by employing canonical correlation analysis. The independent variable set contained the four factors of service quality improvement derived from factor analysis. The dependent variable set included hotel performance changes measured by the differences of volume of sales, average daily room rate, and occupancy before and after a hotel being certified by the Thailand Hotel Standards, or the differences between the expected performance after certification and its current performance.

The canonical correlation analysis of changes in three hotel performance measures, and the four factors of service quality improvement, generated three canonical functions. Three criteria are recommended to consider interpretations of canonical functions (Hair, Anderson, Tatham, & Black, 1998). Generally, a significance level of p<0.1 is used in exploratory studies to indicate significant associations of two sets of variables in a canonical correlation. Multivariate tests, such as Wilks' lambda, Hotelling's trace, Pillai's trace, and Roy's gcr, are used to evaluate all functions simultaneously. As shown in Table 4, the combined canonical functions were statistically significant at 0.05, while individually only the first two canonical functions were statistically significant at 0.10.

Table 4. Overall Model Fit of Canonical Correlation Analysis

Canonical Correlation Analysis
Canonical Function	Canonical Correlation (Rc)	Canonical R^2	F‐Statistic	Prob	Redundancy Index
					Dependent Set	Independent Set
1	0.32	0.10	1.99	0.03	0.017	0.038
2	0.30	0.07	1.80	0.10	0.054	0.023
3	0.13	0.02	0.95	0.39	0.002	0.001
Multivariate Tests of Significance
Statistic	Value	Approximate F Statistic	Prob.
Wilks' Lambda	0.82	1.99	0.03
Pillai's Trace	0.19	1.99	0.03
Hotelling's Trace	0.21	1.97	0.03
Roy's gcr	0.10

The second criterion is the practical significance measured by the magnitude of a canonical relationship. As an arbitrary rule of thumb, a canonical function should be interpreted if its canonical correlation, a measurement of the strength of relationship between two variates, is equal to or greater than 0.30 (Garson, 2006). As shown in Table 4, the canonical correlation of the first function was 0.32 and the second was approximately 0.30, which were sufficient to be of further interest. The third canonical function had a correlation much lower than 0.30, which was excluded from further analysis.

The third criterion of redundancy indexes measures the shared variance in a canonical variate explained by the variance of the other canonical variate in each canonical function (Hair, Anderson, Tatham, & Black, 1998). High redundancy indexes indicate strong predictability of one canonical variate to explain the other set in a given canonical function. The redundancy indexes of the dependent variate were 1.7%, 5.4 %, and 0.2% respectively in the three functions. Redundancy indexes of 1.5% or 2% were often used as cut‐offs (Baloglu & Uysal, 1996; Oh et al., 1995 quoted in Baloglu, Weaver, & McCleary, 1998), therefore the first two functions were considered for further analysis. The first two functions also explained most of the total redundancy with the first accounting for 23.29% and the second accounting for 73.97%, as shown in Table 5.

Table 5. Canonical Loadings in the Canonical Functions

Variable	Canonical Loadings
	1	2	3
Correlations between the dependent variables and their canonical variates (Hotel Performance Changes)
Change in Sales	− 0.69	0.72
Change in Average Daily Room Rate		0.91
Change in Occupancy		0.93
Correlations between the independent variables and their canonical variates (Service Quality Improvement)
Service Delivery	− 0.62
Hotel Employees		0.80
Facilities & Surroundings	− 0.55	0.67
Prestige	− 0.82
Proportion of total redundancy in the dependent set	23.29%	73.97%

The combined consideration of significance levels, magnitude of canonical correlations, and redundancy indexes suggested further analysis of the first two canonical functions. As shown in Table 5, the first two canonical functions accounted for almost 90% of the variation in the variable set of hotel performance changes and about 70% variation in the set of service quality improvement variables.

Table 5 contained the canonical loadings or structure correlations of variables in the dependent and independent canonical variates. To describe the relationship between the independent and dependent measures, canonical loadings greater than absolute value of 0.50 are accepted for further interpretation (Hair, Anderson, Tatham, & Black, 1998). The results revealed different structures of the variable sets in the dependence relationship. The first dependent variate containing only one dependent variable—change in sales—has a variance of 48% with canonical loading of − 0.69. Canonical loadings of the independent variables ranged from − 0.82 to − 0.55. The first canonical function suggested that change in sales had a positive and significant relationship with changes in prestige, service delivery, and facilities and surroundings. The second dependent variate had a high shared variance of 74% among the three dependent variables. In the second canonical function, a positive relationship was also found between the independent and dependent variates and indicated that changes in occupancy, average daily room rates, and sales were strongly associated with improvements in hotel employees and facilities and surroundings.

CONCLUSIONS

The study was undertaken to investigate the perceived impacts of an official hotel rating system—the Thailand Hotel Standard—on hotel service quality. The demographic profile of hotels sampled showed that most hotel respondents were independently operated business hotels or resorts, ranging from small to medium in size, and mostly in business less than 20 years. The Foundation for Standard and Human Resource Development in the Service and Tourism Industry or other relevant organizations need to encourage these hotels to participate in the rating system. The major concern of these small and medium‐sized independent hotels is constraint in financial resources (Vallen & Vallen, 2005); therefore, guidance to service quality management under revenue and cost control can be very helpful.

Most respondents were aware of the Thailand Hotel Standard as the only national hotel rating system in Thailand and showed strong interest, even though a significant number of hotels never applied. The Foundation for Standard and Human Resource Development in the Service and Tourism Industry and its regional representatives can take a more proactive role to promoting and encouraging hotels and similar establishments to be part of the hotel rating system. For example, the Foundation may offer service quality improvement programs or workshops to hotels with an intention to apply. This can generate stronger interest in the hotel standard and signifies a commitment in action, yet does not require hotels to seek ratings immediately.

The findings showed that most hotels applied for three‐ to five‐star ratings, while hotels at one‐ and two‐star levels lacked interest in being involved in the system. As one‐ and two‐star hotels target an important travel market of budget travelers, such as international backpackers and family travelers, the Foundation should consider promoting benefits and incentives that are specifically designed for these hotels, due to their unique characteristics and constraints.

An exploration of the underlying constructs of hotel service quality improvement revealed four dimensions: service delivery, hotel employees, facilities and surroundings, and prestige, representing the key areas that Thai hotels improved when applying for the Thailand Hotel Standard. The canonical correlation analysis indicated that there was a significant relationship between improvement in service quality and changes in hotel performance. In the first relationship, hotels' performance change in sales corresponded with improvement in service delivery, facility and surroundings, and prestige. Average daily rate and occupancies, the two parameters in room revenue measurements, were not significant. This relationship examined the association between changes in non‐room revenue and service quality improvement. Hoteliers can increase non‐room‐related revenue through improving service delivery with better service production and error‐free delivery, upgrading guest facilities and surroundings, and enhancing hotels' prestige.

In the second canonical analysis, changes in sales, average daily rates, and occupancy were strongly related to improvement in employees' service competencies and hotels' facilities and their surroundings. This relationship examined changes in hotel performance with an emphasis on room revenue. Hotel managers who intend to achieve higher room revenue need to focus on improvements in their properties' tangible assets—facilities and surroundings—and intangible assets—employees' service competencies. Therefore, maintaining and upgrading guest facilities and surroundings was important for generating both room and non‐room‐related revenue. As employees were found strongly associated with room sales, hotel managers should recognize the importance of recruiting and retaining qualified employees and offering competency and skill training programs.

Limitations

Cautions should be used when applying the research findings, due to several limitations. Sample prejudice may exist. The target population of the study included all lodging establishments in Thailand. Among approximately 5000 hotels and similar establishments, only about 400 hotels registered with the Thai Hotel Association. The majority of the accommodation consists of small hotels and guest houses. In addition, the number of star‐awarded hotels is small. Because the Thailand Hotel Standard has only been in operation for a few years, its promotional campaign throughout the country did not start till recently. Therefore, only a limited number of hotels participated in this study. Nonresponse bias can also affect the response rate. Questionnaires were distributed during the peak travel season in Thailand and hotel managers might be too busy to reply. Lastly, measurement bias may occur. Since measurement in this questionnaire is attribute‐based, it might overlook some respondents' quality perceptions, as suggested by Stauss (1993).

REFEREED ANONYMOUSLY