Perceptions of Chinese hotels.

The evaluations of 941 respondents who had recently stayed in a Chinese hotel revealed that guests' satisfaction with the hotels were influenced by hotels' star ratings. By comparing the importance-evaluation matrices for each category of hotel, a clear trend emerges in which higher levels of satisfaction are discernable for five-star and deluxe hotels than for one- and two-star hotels. At all levels, however, respondents gave high ratings to the key attributes of cleanliness and bed comfort. The findings raise a question of how to ensure satisfaction for China's one- and two-star hotels.

Keywords: Chinese hotels; importance-performance analysis; hotel star ratings; hotel guest satisfaction

The primary purpose of this article is to compare matrices derived from importance-evaluation scales completed by Chinese guests staying at hotels in mainland China. These matrices are based on hotel categories that range from five-star deluxe to one star. The findings are based on data derived from a sample of 941 Chinese respondents who had stayed in a hotel six months prior to the completion of the questionnaire. They form part of a larger sample of 4,210 mainland Chinese respondents, thereby permitting comparison with less frequent users and nonusers of hotels.

Literature Review

Although we have seen few studies relating to China, the issues of hotel choice and determinants of satisfaction have been examined in many other studies. In this context, Dolnicar and Otter (2003) reviewed twenty-one studies undertaken between 1984 and 2000, looked at factors that determined occupancy rates, and generated a list of 173 items. These include, in no particular order, past experience and word-of-mouth recommendation, perceptions of corporate image, external rating of a hotel, location, the existence of loyalty schemes, pricing, promotional policies including special deals, and presence on the internet. Other studies have recognized the importance of market segments and their differing needs. For example, Weaver and Oh (1993) studied the needs of business travelers in the United States, while Lockyer (2002) looked at motel users in New Zealand. Studies have identified the importance of certain variables in determining levels of patronage, satisfaction, and repeat purchase, among them, price and cleanliness (Saleh and Ryan 1992; Lockyer 2002). Lockyer (2005) reported findings derived from focus groups that identify three key issues in the selection of and satisfaction with a hotel room (see also Lockyer 2002, 2003). These were (in order) the availability of a room, the price (which was a trigger point for initial hotel selection), and cleanliness, which is both a trigger point and a direct influence on choice. Many countries, including China, operate star rating systems that are based on criteria such as room size, facilities, existence of recreational facilities such as swimming pools, and range and choice of restaurants. Some countries, like New Zealand in its Qualmark scheme, also attempt to incorporate qualitative, perceptual measures from guests, as well as objective measures.

Much of the literature that pertains to perceptions of quality and client satisfaction is predicated on the SERVQUAL concept initiated by Parasuraman, Zeithaml, and Berry (1985, 1988, 1994a, 1994b, 1994c). However, we saw difficulties with this framework that caused us to use a different approach, based on ratings of importance and evaluation. With SERVQUAL, service quality is measured by gaps, one of which is the gap between guests' expectation and their evaluation of specific service performance and tangible attributes such as physical appearance. The model suggests five key dimensions to service quality, namely, reliability, assurance, tangibles, empathy, and responsiveness (often referred to as RATER). The stability of these dimensions is a subject of debate. For example, Carman (1990) and Saleh and Ryan (1991) were unable to reproduce the five separate dimensions found by Parasuraman's group. Saleh and Ryan (1991, 1992) introduced the discussion of trigger points in evaluations. Ryan (1999) questioned the independence of expectation and evaluation as separate variables, arguing on the basis of work previously undertaken by Cronin and Taylor and Teas that apart from difficulties in interpreting given gap values, the relationship between the two variables was interactive through feedback mechanisms. He also questioned the degree to which the model was wholly applicable to tourism and hospitality circumstances, where (1) the client is a key actor in the service process; (2) the client is able to engage in cognitive dissonance whereby poor performing components may subsequently be reconsidered as unimportant; (3) the duration of service encounters and experiences may often be far longer than those considered by Parasuraman, Zeithaml, and Berry; and (4) a tension exists between the degree to which the tourism experience is a series of sequential experiences or a holistic experience. Additionally, the concept of good quality service within the SERVQUAL model is based on the extent to which an evaluation exceeds expectations. This is challenging within the hospitality industry, where so-called "seven-star" hotels, such as the Emirates Palace (Abu Dhabi) and Burj al Arab (Dubai) are emerging, rendering client expectations high. We are certain that these properties are capable of matching expectations, but it seems unreasonable to argue that matching expectations without exceeding them fails to equate to quality service. Finally, the nature of the gap between expectations and evaluations has been further queried as to whether the expectations being measured are ideal or tolerable, and studies of correlations between expectations and evaluations have indicated that most variance lies in the evaluation scale, which can correlate highly with measures of aggregate satisfaction (Ryan 1999; Mansfeld 1995; Ryan and Cliff 1996, 1997). In 1994, Parasuraman, Zeithaml, and Berry proposed changes to their scale, including relocating items to different dimensions, and introduced a third questionnaire to better identify levels of tolerance (Parasuraman, Zeithaml, and Berry 1994c).

Today, it is possible to conclude that "quality is difficult to measure, so on certain occasions its determinants are studied rather than the concept itself" (Bigne et al. 2005, 1025). Indeed, one can maintain that the following issues raised by Buttle (1996) have not been answered (see also Aijzen 1991; Aijzen and Fishbein 1980). He wrote, "There are still doubts about whether customers routinely assess SQ [service quality] in terms of expectations and perceptions; there are doubts about the reliability and appropriateness of the disconfirmation paradigm; there are doubts about the dimensionality of SQ; there are doubts about the universality of the five RATER dimensions. These are serious doubts which are not only significant for uses of SERVQUAL but for all who wish to understand the concept of service quality" (Buttle 1996, 25).

Comparing the importance of attributes with an evaluation of performance on those attributes circumvents some of the difficulties cited by Buttle (1996). Respondents are asked how important an attribute is to them rather than asking what they expected. Conceptually, the approach of multiattribute theory is based on cognitive and affective distinctions. Extended models involve an evaluation of the consequence of the action, termed conative aspects (Aijzen 1991; Aijzen and Fishbein 1980; Oh 2001; Ryan 1995). This approach overcomes the objection regarding SERVQUAL that its survey respondents are being asked to recall expectations after actually experiencing the service. Importance is arguably a more general concept than expectation, and respondents' possible amendment of their estimate of levels of importance in the light of current experience matters less than do revisions in expectations when one is trying to assess future managerial action (Oh 2001; Ryan 1991, 1995). In applying the importance-evaluation matrix, we did not test gaps, as some studies have done, because it is possible to question what the gap actually represents. We draw a series of conventional importance-evaluation matrices on the premise that they possess the advantage of visual impact.

Structure of the Chinese Hotel Industry

According to China National Tourism Administration (CNTA) data, China is home to 11,828 hotels. Its stock of 1,332,083 rooms represents a tenfold increase in the number of hotels in the past twelve years (CNTA 2006, 2003). In 2005, 1,146 of these hotels were four-star properties, and 281 held five-star ratings. The average four-star hotel had 210 rooms, and the average for five-star hotels was 379 rooms. These were much larger on average than one-star properties (52 rooms), two-star hotels (75 rooms), and three-star properties (127 rooms).