The model identifies factors that are likely to affect the development of the destination visitor image, such as destination location, and other destination environment cues. In addition, factors that are likely to moderate and mediate the relationship between self-congruity and travel behavior are identified and explained.

Consumer research has shown that a consumer's attitude toward a product (and product purchase) is influenced by the matching of the product user image with the consumer's self-concept [1]. Indeed, consumer research has had a long tradition of addressing how selfconcept is used as a cognitive referent in evaluating symbolic cues. Product symbolic cues refer to stereotypic images of users of a product or store (example sexy, classy, fashionable, young). The same can be said in relation to a tourist's attitude toward a destination (and travel behavior). That is, tourists perceive destinations differently in terms of the destination's typical clientele or visitors. This stereotypic image of the kind of people who typically visit a given destination is referred to here as destination visitor image.

1. To explicate issues pertaining to Penang's image as tourist destination with a view to improving the contribution of tourism to economic development. a) To analyze the motivation of individuals to visit Penang;

b) To analyze the impact of self-concept on consumer satisfaction;

b) To analyze the impact of self-concept on consumer satisfaction;

d) To explain the implication of self congruity for tourism geography development.

2. To determine the relationship between self-congruity on satisfaction that will lead to destination loyalty, which subsequently has direct implication on tourism development.

The focus of this research was to determine and investigate the effects of self-congruity, the match between the visitors' destination image and tourists' self-concept (actual, ideal) on satisfaction and destination loyalty of tourist. Penang was chosen as the case study to examine these issues; used earlier on Penang is a portent tourist destination and new knowledge would contribute to sustainable tourism in the state. The respondents were randomly selected and self-administered questionnaires were distributed to them at Penang airport.

This study aim was to examine the effect of self-congruity on satisfaction and destination loyalty of tourist particularly with impact to Penang.

Self-congruence is a significant area of research in consumer behavior/marketing, because it provides the marketing manager with strategic insights concerning positioning and advertising research and may also serve as a basis for market segmentation. It is important to note that because self-congruity involves a process of matching a tourist's self-concept to a destination visitor image, factors such as destination location and other destination environment cues are likely to affect the development of the destination visitor image [2]. Tourism managers who understand how self-congruity works may be in a position to develop effective positioning strategies that increase profitability.

Self concept

Self-concept first appeared in the marketing literature nearly thirty years ago [3]. Self-concept incorporates the 'cognitive and affective understanding of who and what we are' [4]. A number of self-concepts have been proposed: actual, ideal, social, ideal social, expected, situational, extended and possible [5]. Most research has concentrated on actual and ideal self-concept.

The self-concept literature in consumer behavior is muddled with ambiguity and confusion in regards to the precise conceptualization of self-concept. The actual self-concept in this vein has been labeled as "actual self", "real self", "basic self", "extant self"' or simply "self".

Two factors link self-concept to marketing: product symbolism and self-concept/brand image congruity [6]. According to this view, the relationship between self-concept and product or brand image influences product evaluations, and hence affects purchase behavior. Situational self-concept is 'the meaning of self an individual wishes others to have of him/her' [7]. Self-image/brand-image is therefore moderated by situation conspicuousness.

In the multiple self-construction tradition, self-concept has been conceptualized as having two or more dimensions. Some investigators argued that self-concept has to be treated as having two components - that of the actual self-concept and the ideal self-concept (the ideal self-concept is defined as the image of oneself as he/she would like to be) [8-21]. The ideal self-concept has been referred to as the "ideal self", "idealized image", and "desired self".

The social self-concept (sometimes referred to as "looking-glass self" or "presenting self") was defined as that image that one believes others have of him/her; and the ideal social self-concept (sometimes referred to as "desired social self") denotes that image that one would like others to have about him/her [22]. Hughes and Guerrero (1971) [23] talked about the actual self-concept and the ideal social self-concept. French and Glaschner (1971) [24] used the actual self-concept, the ideal selfconcept, and the "perceived reference group image of self". Dornoff and Tatham (1972) [25] used the actual self-concept, ideal self-concept, and "image of best friend." Sanchez and Brien (1975) [26], on the other hand, employed the actual self -concept, ideal self-concept, and the "expected self" (which refers to that image somewhere between the actual and the ideal self-concept). Furthermore, Munson and Spivey (1980) [27] referred to the "expressive self" who pertains to either the ideal self-concept or social self-concept.

This construction is labeled differently in the sociology and social psychology literatures - "presenting self", "role self", "social self", etc. In the case where situational self-image can be expressed through product consumption, the individual compares product or brand images to that of his/her situational self-image. The brand whose image matches the situational self-image influences the individual's decision making regarding his preference and purchase of that brand.

Product image

Tucker (1967) [28] argued that consumer's personalities could be defined through product use.

Products, suppliers, and services are assumed to have a personality or image, just like people. Images of products, suppliers, or services are also formed by other associations such as stereotyped image of the generalized or typical user and other related factors [3,29,30].

It should be noted that product images could be classified as being functional or symbolic [31,32]. Symbolic product images refer to the stereotypic personality images consumers have of a specific product.

Self-image/Product-image congruity

The discussion of self-concept and product-image congruity was initiated by Gardner and Levy (1959), Levy (1959) [29,33]. The main attention was focused upon the image projected by various products. Consumers were thought to prefer products with images, which are congruent with their self-concept. Self-image/product-image congruity specifically refers to the match or mismatch of one or more actual selfimage, ideal self-image, social self-image, or ideal social self-image with the corresponding personality images of the designated product. The match between actual self-image and product image has been referred to as "self-congruity"; between ideal self-image and product image", "ideal congruity"; between social self-image and product image as "social congruity"; and between ideal social self-image and product image as "ideal social congruity" [6,34-36].

The issue whether product image interact with one or more selfperspectives such as, the actual self-image, ideal self-image, social self-image, and the ideal social self-image has been addressed by Sirgy (1981b, d) [31,37]. Many studies have treated self-image/productimage congruity only and exclusively in terms of a single congruity effect, namely self-congruity (that is, involving only the actual selfimage) [23,38-41]. Most of these studies provided support for the relationship between self-congruity and consumer choice.

Many other studies have been found to assess the independent effects of ideal self-image versus actual self-image (or ideal congruity versus self-congruity) on consumer choice [6,8,14,16,17,25,34].

Customer satisfaction

Customer satisfaction is a post-purchase evaluation of a service offering [42,43]. A traditional definition of customer satisfaction followed the disconfirmation paradigm of consumer satisfaction/ dissatisfaction (CS/D), suggesting that CS/D is the result of interaction between the consumer's pre-purchase expectations and post purchase evaluation [44]. Furthermore, satisfaction not only consists of cognitive element but have to include emotional element in determining customer satisfaction.

Creating customer satisfaction is a defensive strategy and the behavioral objective for defense is customer loyalty [45].

Fornell (1992) [45] discussed that customer satisfaction will be influenced if the demand and supply are different. To retain customer, switching barrier and customer satisfaction are the two basic forms which need to be fulfilled. Switching barriers make it costly for customer to switch to competitors and customer satisfaction makes it costly for competitor to take away another firm's customer. According to Fornell (1992) [45], switching barrier is less effective compared to customer satisfaction. This study concentrates on using defensive strategy, particularly customer satisfaction to win customer loyalty while attitudinal dimensions refer to favorable customer intention to repurchase and recommend, which are good indicators of a loyal customer [46,47].

Customer loyalty

A lot of studies on the topic of loyalty have been measured by behavioral aspect of brand loyalty, such as repeat purchase without considering cognitive aspects of brand loyalty [48]. For instance, Fader and Schmittlein (1993) [49] conducted a research investigating the advantage of high share brands in brand loyalty, suggesting that high share brands have significantly higher brand loyalty than low share brands. Based on the behavioral element of brand loyalty, Sheth (1968) [50] provides an operational definition of brand loyalty; that is brand loyalty is a function of a brand's relative frequency of purchase in both time-independent and time dependent situation. Dick and Basu (1994) [47] proposed that loyalty should be evaluated with both attitudinal and behavioral criteria (Composite brand loyalty). From the conceptual and operational definition of brand loyalty, we can get the two most important elements of brand loyalty: attitude and behavior.

Customer consists of brand loyalty [47], has been described as a behavioral response and as a function of psychological processes [51]. There are two dimensions to customer loyalty: behavioral and attitudinal [52]. The behavioral dimension refers to a customer's behavior on repeat purchase, indicating a preference for a brand or service consistently over time [53].

Satisfaction as antecedent of customer loyalty

When customers are able to have the opportunities to evaluate the quality of the delivered service, satisfaction is expected to have an effect on customer loyalty. Cronin and Taylor (1992) [54] proved that customer satisfaction exerts a stronger influence on future purchase intention. They added that loyalty increases with customer satisfaction at an increasing rate. However, satisfaction-loyalty relationship is not linear; customer satisfaction leads to customer loyalty. But to significantly prove that satisfaction predict loyalty remain uncertain.

Besides, there are still some inadequacies in regards to determining the elements of satisfaction on customer loyalty. Similarly, the study of Yu and Alison (2001) [55] found that emotional component of satisfaction is a strong predictor on customer loyalty.

Destination choice

Most models on tourism destination choice include a reference to the importance of previous experience on the destination choice process, often symbolized by a feedback loop after the actual destination experience into the evaluative stages of future destination decisions [56,57,58]. Woodside and Lysonski's (1989) [58] model of traveler destination choice included previous destination experience in the traveler's variables that influence destination awareness as well as traveler destination preferences. They also hypothesized, "Previous travel to a destination relates positively to inclusion of the destination in a consumer's consideration set versus other mental categories [inert, inept, and unavailable sets] of vacation destinations."

Measures and methods of self-congruity

There are two primary methods of measuring self-image congruence [59]. The traditional method of measuring self-congruity is based on tapping participants' perceptions of destination visit image and participants' perceptions of their self-image in relation to the destination patron image. It should be noted that self and destination visitor image dimensions can be measured using semantic differential scales or Likert-type scales. These measures are mostly tailor-made, that is, the image dimensions are elicited as a function of the destination in question [60].

Despite prevalence of the traditional method of measuring selfimage congruence, Sirgy et al. (1997) [59] identified methodological problems inherent in the discrepancy score measure of self-congruity. To address these limitations, Sirgy et al. (1997) [59] designed a new method to measure self-image congruence; however, their research focused on self-image congruence only in consumer goods and services settings.

The new method of measuring self-congruity applied in a tourism context can be exemplified as follows: Participants are instructed to take a moment to think about [destination x].

"This [destination x] is consistent with how I see myself" (actual self-image).

"This [destination x] is consistent with how I like to see myself" (ideal self-image).

"This [destination x] is consistent with how I believe others see me" (social self-image).

"This [destination x] is consistent with how I would like others to see me" (ideal social self-image).

Participants would then mark their response on a Likert-type scale in relation to the last four self-image statements. Sirgy et al. (1997) [59] argued that this method does not cue participants to a specific image category or dimension. The method cues participants to conjure up their own image of the destination visitor. Then, the method guides them to indicate their global perception of degree of match or mismatch between how they see themselves (self-image) and the destination patron image. In other words, this measurement procedure captures self-congruity directly (not by measuring destination visitor image and self-image separately) and globally (not by asking participants to indicate their perception of congruity with predetermined images). As such, the new method assumes that self-image congruence is processed holistically or globally, not analytically or piecemeal as the traditional method assumed.

Measures and methods of destination loyalty

Jacoby and Chestnut (1978) [51] suggested that the behavioral approaches could be further subdivided into five types: brand purchase sequence, brand purchase proportion, brand purchase probability, synthesis measures, and miscellaneous measures.

As regards proportion of purchase, other authors such as Cunningham (1956) [61] extended the concept of one-brand loyalty to dual-brand or triple-brand loyalty, with loyalty defined as the percentage of total purchases devoted to the top two or three brands.

Finally, the loyalty indices involved a composite measure of several behavioral components, such as brand expenditure share, switching behavior, and number of brands available.

Summary

This research draws from the self-concept literature in consumer behavior to help explain the psychological underpinnings of travel behavior. Consumer research has shown that a consumer's attitude toward a product (product purchase) is influenced by the matching of the product user image with the consumer's self-concept [1]. The greater the match between the destination visitor image and the tourist's self-concept, the more likely it seems that the tourist has a favorable attitude toward that destination (and the more likely it seems that the tourist would visit that destination). This matching is then referred to as self-congruity [2].

According to Sirgy (1986) [62], the self-congruity theory proposes that consumer behavior is determined, in part, by the congruence resulting from a psychological comparison involving the product-user image and the consumer's self-concept (actual self-image, ideal selfimage, social self-image and ideal social self-image). Self-congruity affects consumer behavior through self-concept motives such as the needs for self-consistency and self-esteem.

Self-image congruence is a significant area of research in consumer behavior because it provides the marketing manager with strategic insights concerning positioning and advertising research and may also serve as a basis for market segmentation.

The intent of this study is to analyze spontaneous self-congruity on satisfaction and destination loyalty among the tourists in Penang. Since their choice of destination have been traced from this Island City, information regarding the area will be relevant. Based on the discussion on the preceding chapters, origin of the data source, sample selection, data collection procedure, operationalization of variables and statistical procedures employed for the analyses will be discussed in this chapter.

Theoretical framework

The theoretical framework is to discuss the interrelationships among the variables that are deemed to be integral to the dynamics of the situation being investigated. The primary interest of this study is to examine the effects of self-congruity on satisfaction and destination loyalty of tourist in the case of Penang.

There are three sets of variables used for this purpose, namely: dependent, intervening and independent variables. The dependent variable chosen for the purpose of this study is destination loyalty. The intervening variable is satisfaction leading to the destination loyalty transfigured by the respondents' self-image congruity. Self-congruity factors categorized by actual self-image, ideal self-image, social selfimage and ideal social self-image were selected as the independent variables.

Figure 1 depicts the relationship between these sets of variables:

1) Dependent variable – destination loyalty.

2) Intervening variable – satisfaction.

3) Independent variables – actual self-image, ideal self-image.

Figure 1: Relationship between these sets of variables.

Research approach

Before we start using the theory we must finalize Penang image as a tourist destination (Product Image) for that we have chosen 40 tourists of different ethnic group e.g. Asian, European and American both male and female selected to discover the performance perception of Penang using a series of questions designed to measure the respondent's expectations and performance perceptions of 15 functional attributes of the destination area (Table 1). The 15-item attributes were derived from a previous study which was conducted to identify the image of Norfolk as a travel destination [63]. The 15-item functional attributes were:

1) Places for water activities such as swimming, sail boating and cruise ships.

2) Places for golfing and other sports in Penang

3) Places of historical interest.

4) Places of cultural interest.

5) Festivals in Norfolk.

6) Places of scenic beauty.

7) Nice treatment from local people.

8) Restful and relaxing atmosphere.

9) Good shopping places.

10) Variety and quality of restaurants

11) Availability of entertainment or night life

12) Availability of suitable accommodations such as hotels or motels.

13) Tours of naval base and naval ships.

14) Easy accessibility to the area.

15) Variety and quality of attractions.

Table 1: Distribution of items of the questionnaire.

For each of the 15-item functional attributes of Penang tourism features, the respondent was requested to rank seven attractions from the previous list.

The image of Penang as a tourist destination is finalized as follow: Penang is generally a vacation destination and I would visit it for fascinating sheep shopping, exotic food friendliness of local people, and variety of greet tourist attractions including historical site tropical beaches, unspoiled country site, and a beautiful scenery, relaxing and calm.

It seems that, self-congruity is a holistic-like perception. Traditional research methods, which assume a piecemeal process, may not adequately capture the holistic nature of self congruence and, therefore, may have limited predictive validity. For instance, we follow up the new method of measuring to alleviate the problems inherent with the traditional method [59]. We hypothesize that the new method will be more predictive of a variety of tourist behaviors such as satisfaction or dissatisfaction. This study, will contribute through extending the self-image, product image, congruity theory to product loyalty. Due to holistic nature of self-congruity, the study expects to determine that self image and product image congruity will lead to product loyalty.

Sampling design

The sample of population involved individuals who had actually visited the city of Penang from the various parts of the world starting from April 2001 till May 2002. Random samples of 400 individuals were secured for the study and 145 were returned.

Content and purpose of the questionnaire

The traditional methods of measuring the self-image congruence are besieged with three key problems: (1) the use of discrepancy scores (2) the possible use of irrelevant image and (3) the use of compensatory decision rule. The problems relate, in part, to the assumption that image congruence is multidimensional. In nature most of self-congruence studies have been focused mostly on pre-purchase types of consumer behavior rather than post-purchase behavior. The pre-purchase types of consumer behavior study do not mitigate the limited predictive validity as well as unable to meet the holistic nature of self-image congruence. For instance, the study was pursued under the combination of new and traditional method to mitigate the fallback of traditional method. The five dimensions of self-image had measure spontaneously to the destination loyalty of the typical tourist in Penang. The post-purchase self-image dimensions were able to attribute for individual tourist loyalty that reflected the image question such as "who I am"; "similar to me" and "very much like me" according to the way of designed questionnaire.

Operationalizing the variables

Various factors influence self-image congruity to the destination loyalty among a given number of the population sample. Self-congruity factor is categorized by actual self-image, ideal self-image, under the independent variable. Within a given community, different factors may be differentially associated with self-congruity because the situational atmosphere within which the community exists may impinge on these factors. While few have a highly significant association with perceived self-congruity in the population sample, other factors may either have insignificant or negatively significant associations. Possible event such as motivation factor may result in some effect having several levels of significance to destination loyalty.

Destination loyalty was selected as dependent variable. Satisfaction that leads to the destination loyalty transfigured by the respondents' self-image congruity is an intervening variable. The four independent variables were selected for the study as previously mentioned.

Hypotheses formulation

H1: Self-image will have a positive influence on loyalty.

H1a: Ideal self-image will have a positive influence on loyalty.

H1b: Actual self-image will have a positive influence on loyalty.

H2: Self-image will have a positive influence on satisfaction.

H2a: Ideal self-image will have a positive influence on satisfaction.

H2b: Actual self-image will have a positive influence on satisfaction.

H3: Satisfaction will have a positive influence on loyalty.

H4: Satisfaction mediates the relationship between self-image and loyalty.

H4a: Satisfaction mediates the relationship between ideal selfimage and loyalty.

H4b: Satisfaction mediates the relationship between actual selfimage and loyalty.

Findings

The characteristics of the respondents have been previously introduced in this study. Subsequently, the instrument (questionnaire) for this study was tested by running factor and reliability analyses to confirm the validity and reliability of the items (or questions). Correspondingly, multiple regression analyses were employed for the relationships between 1) self image (ideal and actual) and customer satisfaction, 2) self image (ideal and actual) and customer loyalty. Additionally, customer loyalty was tested to ascertain its mediating effect on the relationship between self image and customer loyalty.

Sample profile

Data for this study was collected from international tourists whom had visited Penang. A total of 400 questionnaires were sent out with 145 returned (response rate of 36.25%) out of which only 100 were usable (usable rate of 68.97%). Table 2 reveals majority of the respondents who had visited Penang either exhibited one of the following attributes, which are as follows: 1) male (55%), from the mature age group (40%), 3) married (53%) or 4) From the 20,000 to 30,000 annual income brackets. Surprisingly, majority of the respondents surveyed indicated that they had at least visited Penang for 5 times. The reasons for their visits include vacation (60%), meeting up with friends (10%), business (14%), education (6%) and others (4%).

Table 2: Profile of respondents.

Goodness of measures

Factor analysis: Factor analysis is a process of identifying the underlying factors that are present among the items used for the survey. It should be noted that the sample in this study meets the variable to sample ratio of 1:10. Kaiser-Meyer-Olkin (KMO) is used to measure the sampling adequacy of the items and to ascertain whether the factor analysis is fit for selection. All anti-image correlations for the following factors that is, the negative of partial correlation) generated are significantly high (>0.5) which implies that the components can be used for subsequent analyses (example multiple regression analyses). The total variances explained by all factors are above 50%. Additionally, the latent root criterion of Eigen value greater than 1 was used as a priori for the extraction method while Varimax rotation was used to ease the identification of the factors as it allows unique separation between components.

The factor loadings for all the three variables were satisfactory as they were loaded on the expected number of components. The cutoff point for this study uses the criteria of a minimum loading of 0.51 for a sample of 100 [64]. The minimum factor loading for this study is 0.68 (that is, third and fourth items of actual self image). There were two components to self-image that can be categorized as ideal self-image and actual self-image as shown in Table 3. On the other hand, all three items measured satisfaction as predicted with loadings ranging from 0.78 to 0.89 as shown in Table 4.

Table 3: Factor loadings for self image.

Table 4: Factor loadings for satisfaction.

Additionally, loyalty is being measured by 2 items, which exhibited a single component after rotation. This construct consists of 2 items, the first in which respondents' intention to revisit Penang were solicited, while the second is concerned with their willingness to refer or recommend their friends to Penang as a tourist spot in the future as shown in Table 5.

Table 5:Factor loadings for loyalty.

Reliability analysis: After having employed factor analyses, calculation of reliability is conducted (by using Cronbach Alpha) on each dimension that has been extracted. The general rule of thumb is 0.70 and the minimum value for this study is 0.68 for customer loyalty. However, according to Stern et al. (1977) [12], values of 0.70 and lower could be considered if the study is exploratory. Therefore, since this study is novel in its application in Malaysia, it is considered appropriate to accept alpha values as low as 0.50. The Cronbach alpha reliability coefficient is shown along the diagonal of the Pearson correlation matrix in Table 7.

Table 6: Descriptive statistics.

Table 7: Pearson correlation matrix.

Descriptive analyses: The average (or mean) is used as the measure of central tendency for the variables under study. Overall, all constructs registered higher than the middle of their scales and have broad ranges, which indicate high sample variability as shown in Table 6.

Pearson correlation analysis: Pearson's product moment correlation (r) is used as a measure of association to explore the patterns that exist in the relationship between the variables. The correlation coefficient (r) takes values ranging from –1.00 (perfect negative linear relationship) to +1.00 (perfect positive linear relationship). A value of 0, on the other hand, suggests that the variables are uncorrelated with each other. However, it does not tell us about one variable's influence on another. Hence, Pearson's coefficient only identifies the extent of one variable's position that occupies the same relative position on another variable.

All variables except actual self-image and satisfaction do not correlate with each other at p<0.01 or 0.05. This provides indication that the variances of actual self-image do not influence satisfaction.

Multiple regression analysis for the relationship between self image and loyalty: The relationships between the variables that have been explored provide a basis for us to verify the influences among the variables in this study. Hence, multiple regression analyses were employed with the direct relationships investigated as shown in Table 8.

Table 8: Relationship between self-image (ideal and actual) and loyalty.

The model investigated was found to be significant at p<0.01 with 24% ability in explaining the variance in loyalty. Both independent variables were found to have a positive impact on loyalty. Primarily, ideal self-image (β=0.32) was found to have a positive influence at p<0.01, whereas actual self-image (β=0.23) was significant at p<0.05. This denotes the first step in Baron and Kenny's (1986) test of mediation, in which we have sufficient evidence to believe that this relationship might be mediated by a third variable (Figure 2).

Figure 2: Partial mediating effect of satisfaction on the relationship between ideal self-image and loyalty.
1. For full mediator effect: βb (0; (d (0; (f (0 and (j (0; (h = 0
2. For partial mediating effect: (b ( 0; (d ( 0; (f ( 0 and (j ( 0; (h ( 0 but (h < (f
3. Steps 1 to 3 are taken from a modified version of Kenny, Kashy and Bolger (1998) in Handbook of Social Psychology chapter, which is available at http://users.rcn.com/dakenny/mediate.htm
From the Figure: (b ( 0; (d ( 0; (f ( 0 and (j ( 0; (h ( 0 but (h < (f
Therefore, satisfaction acts as a partial mediator on the relationship between self-image and loyalty.

Multiple regression analysis for the relationship between self image and satisfaction: This second multiple regression analyses is to introduce the mediating variable (satisfaction) and to test whether the relationship between the independent variables and mediating effect are supported as shown in Table 9.

Table 9: Relationship between self-image (ideal and actual) and satisfaction.

In this model, ideal self-image (β=0.42 at p<0.01) managed to explain 17% of the variation in satisfaction with actual self-image (β= -0.02) not found to have a positive influence on satisfaction.

Multiple regression analysis for the relationship between ideal self-image, satisfaction and loyalty: In this third analysis, this study investigates both ideal self-image and satisfaction to loyalty. Since, actual self-image was found not to have a positive influence on satisfaction, only the mediating effect of satisfaction on the relationship between ideal self-image and loyalty is pursued. However, by correlating satisfaction to loyalty does not necessitate a mediating effect, as the influence might be carried through by the independent variable. Hence, it is required to control the effect of ideal self-image while investigating the relationship of satisfaction on loyalty. The result is presented in Tables 10,11.

Table 10: Impact of ideal self-image and satisfaction on loyalty.

Table 11: Summary of hypotheses testing.

Both independent variables [that is, ideal self-image (β=0.28) and satisfaction (β=0.35)] have a positive influence on the dependent variables at p<0.01. Hence, it can be observed that the effect of ideal self-image effect on loyalty has been reduced in the presence of the mediating variable, that is, satisfaction. Nonetheless, the effect of satisfaction on loyalty while controlling for ideal self-image is not zero. Therefore, it can only be concluded that satisfaction have a partial mediating effect on the relationship between ideal self-image and loyalty (Figure 2).

The prerogative here is to discuss the findings of the results and how the objectives are met for this research. For ease of reference, the objectives are reiterated as below:

1) To identify the influence of self -image (that is, actual and ideal on loyalty);

2) To understand the influence of self -image (that is, actual and ideal on loyalty) on tourist satisfaction;

3) To study the influence of satisfaction as a predictor of tourist loyalty; and

4) To examine the mediating effect of satisfaction on self-image on tourist loyalty.

Objectives of this research are subsequently answered together with theoretical support and justification from past researchers. The conflicting results are compared and contrasted with similar researchers,while non-significant results without a priori literature support are highlighted for comment due to the current context our study (that is, Malaysia). Furthermore, limitations and recommendations for future research are discussed prior to concluding with practical and theoretical implications.

Influence of self image (actual and ideal) on tourist loyalty

As a reiteration of the literature, this study developed actual and ideal self-image to reflect the higher order construct of destination image. Much attention has been devoted to understand a dynamic structure of destination image; particularly when current destinations have been more competitive with other destination markets. This is reflected in the current devaluing of the ringgit (that is, a reflection in the declined of the US currency) which saw many tourists from many countries around the region flocking into Malaysia. Hence it is our study's interest to understand the destination attractiveness as well as market competitiveness of the tourists through two important constructs of self image, that is, actual and ideal image.

In general, image can be referred to as the sum of beliefs, ideas, and impression that people have of a place or destination [65,66]. Ideal selfimage reflects how the tourists want to be perceived by others, while actual image reflects the tourist perception on him-/her self. Although we are aware that Sirgy (1982) [1] had developed a multidimensional model of self concept, which consists of actual self, ideal self, social self, ideal social self, expected self, situational self and global selfattitude, there is still a lack of literature establishing how these dimensions generate customer loyalty. They remain at a stage where an understanding of customer satisfaction through the said dimensions remains in void of how loyalty can be reinforced at the tourist level. Secondly, we do not want to presuppose the existence of established links between the remaining dimensions and loyalty. But, this study took a step further and relied on the more acceptable dimensions of actual and ideal self, subsequently rendering an explanation to their relationships with customer loyalty. Further links should be tested by future researchers with regards to customer loyalty. This study strongly holds true that the additional dimensions of social self, ideal social self, expected self, situational self and global self-attitude are just forms of interaction outcomes between the ideal and actual self. Due to the need of maximizing our explanatory power, this study departs from the positivist method of scrutinizing and "splitting hairs", resulting in overlapping in the interpretation of the same dimensions. Therefore, the aforementioned reasons propel this research to categorize self-concept into ideal and actual self-image for a parsimonious understanding of customer satisfaction/loyalty.

Our study discovered that there is a positive influence of selfimage (ideal and actual) on customer loyalty. This is of no surprise, judging by the fact that all typologies of self-image are based on the perceptions towards the attributes and attractions available within a destination area [67]. In addition, "loyalty packages" from various tourist operators and the government's efforts in various campaigns (example Visit Malaysia Year) are influential in appealing towards the ideal and self image of the international tourists. One respondent noted that "the friendly people… very much as I imagine…" is influential in his intention to revisit and recommend Penang as a tourist attraction to his friend. Words and phrases such as more "relax", "free", "I hope my friends can see me now" and "part of me" shows that their self image (actual and ideal) has been tapped to address loyalty.

Influence of self image (actual and ideal) on customer satisfaction

Given that tourists consume a whole range of services in their holiday and vacation experience, they base their judgments on the quality of the said services through the satisfaction that they derived from the service value chain. Taking from the perspective that self image encompasses a variety of things which range from personal attributes to fantasies, they seek a destination to express themselves. In the consumable goods industry, individuals tend to buy brands whose personalities are close to their image (either ideal or actual is immaterial). In this study, the same concept is applied but in a different context, that is, in a tourist destination (Penang Island, Malaysia). Tourists choose a destination for various reasons; which among others includes living in their fantasies, to relax, or engage in adventurous activities. These motivations surround the self-image. In this study, we chose two types of self image (ideal and actual) to understand customer satisfaction. This is due to the fact that individuals will portray different "selves" in different settings (at home or away visiting another country).

From our findings, the ideal self-image has a positive influence on the guest satisfaction, whereas actual self-image did not play a part. This is due to the fact that ideal self captures the emotional and feelings of assimilation towards the culture of its surrounding environment. Most retired tourists agreed that they enjoyed their stay very much due to Malaysia's multi-ethnic population. They hope that they could be "surrounded by the diverse culture and friendly faces (as noted by one respondent)" in their home country. This portrays that the destination satisfaction has played a role in the ideal self. On the other hand, the actual self did not influence satisfaction, which could be because of their short duration in discovering Penang as a whole. Hence, the short duration could not change their belief systems, which are much ingrained in their culture, values and norms. This study asserts that "ideally" they are satisfied with the entire destination value chain; however, their innate beliefs and embedded actual self-image still struggles to attain the satisfaction. Additionally, the ideal self (that is, the way in which they want others to perceive themselves) influences satisfaction which plays a more important part than the actual self suggest that the tourists tend to be individuals who wants to be the first to travel to new places which their friends/relatives have not been. As reported by one respondent, "I would very much like to inform others about Penang when I get back…but I am irritated with the traffic jam and all". Another respondent agreed that he cannot wait to tell his friend about his escapades in Penang. These statements reflect the ideal self which has taken more importance over the actual self. Perhaps, destinations which are rich in culture and are far apart from the tourists will have slightly more impact on the ideal self image and satisfaction relationship. For those destinations which are closer to the tourists' home, a more important role of the actual self-image might be exhibited due to the adaptability of the tourists in identifying itself to the destination's national culture.

Influence of customer satisfaction on destination loyalty

Practitioners and researchers have not explicitly detailed the factors that could lead to loyalty. Nevertheless, there is a consensus that satisfaction could lead to loyalty [54,68]. Satisfied customers exhibit loyalty through promoting repeat visitation and referrals regarding the destination. In the lodging industry, Getty and Thompson (1994) [69] found that repeat business is generated through the satisfaction towards the service quality. However, results of a completed survey of 364 participants revealed only a weak connection between the two [70]. Further, in examining such factors as purpose of travel and demographics, the study found business travelers to be the least loyal guests. Nonetheless, in our study, satisfaction indeed has a positive influence in explaining destination loyalty. Although we did not explicitly test the various tourist demographics (example business or leisure travelers) and how it impacts on the destination loyalty, yet through interviews we discovered that loyalty can be attributed mostly to their staying experience. The interviewees tend to explain satisfaction in terms of the service that they receive in the hotels. Their views on the satisfaction of the destination (that is, Penang) are somewhat limited to the service they received while purchasing souvenirs, dining at restaurants and staying at hotels. However, they were generally satisfied with their stay in Penang and this has had a significant influence on them to revisit and recommend Penang to their family and friends.

Mediating effect of satisfaction on the relationship between ideal self image and loyalty

Results in this study assert that the key to building long-term loyalty relationships with customers lies in exceeding their expectations, through their ideal self image. It requires a shift of focus on destination product/service to one centered on customers. Merely meeting customers' satisfaction does not ascertain loyalty, tourists expect the destination to meet their expectations and appeal to their ideal self; and holiday escapades are such motivations that promotes the ideal self. Failing to meet the desires and wants of the tourists will result in failed facilitation of loyalty. The service quality in the entire interaction and activities that the tourists undergo in Penang, should be viewed with importance. Succinctly stated, the satisfaction that is met at every node which results in identifying the ideal self-image would in turn, lead to future business repeats.

From the theoretical point of view, these findings add a significant meaning to both consumer behavior literature and tourism marketing literature. The findings in this study clearly indicate the varying importance of the tourist's self-concept (ideal and actual self image) and his/her satisfaction/dissatisfaction with tourism. Its most important theoretical contribution is that the consumer's satisfaction/ dissatisfaction in the destination involves the evaluation of not only its actual but its ideal self image. More importantly, the effects of ideal self image to satisfaction could be carried to generate destination loyalty.

Additionally, most significant strategic destination marketing implications would be that Penang as a unique destination for many tourists should focus on both functional and symbolic attributes of the destination in tourism product development and promotion. This would be particularly important in designing promotional messages aimed at creating a desirable image of the destination in relation to specific market segments, taking into account their self-image. In this regard, it would be highly important for the destination area to identify the symbolic image of the area as perceived by the target market segment and adjust the product development efforts and promotional activities accordingly. Ultimately, tourism managers can conduct a promotion campaign showing a typical tourist reflective of the destination visitor image. If the right destination visitor image is shown in the promotion materials, then a majority of the potential visitors are likely to experience actual self-congruity. To accentuate the experience of actual self-congruity, the promotion could show the actor(s) expressing thoughts and feelings of actual self-congruity. An example would be to show a couple looking at a brochure describing the destination site and saying, "This is the kind of place that people like us go to on a vacation." Such expressions are likely to facilitate the experience of actual self-congruity. Ideal self-congruity refers to the degree of match between a tourist ideal self-image (how tourists like to see themselves). Tourism managers can conduct a promotion campaign showing a typical tourist reflective of the destination visitor image consistent with the majority of past tourists' ideal self-image. To facilitate the experience of ideal self-congruity, the promotion could show the actor(s) expressing thoughts and feelings of ideal selfcongruity. An example would be to show a person dreaming about being the typical person associated with the destination. Showing the dream in action is likely to accentuate the ideal self-congruity. The objective of this research provides an underlying foundation that destination loyalty will be addressed if promotional activities are in tandem with the self-image of the tourists.

This study seeks to note a few limitations. Firstly, the self-image congruence we explored is limited to two types (that is, ideal and actual self-image). Further, research should explore the other possibilities of self-image (example social self) and how it generates satisfaction, with an attempt to further explain destination loyalty. Future researchers could begin with identifying different activities and interactions that tourists face with locals in the whole entire chain of their stay. This can provide a "bird's eye view" of what activities that appeal to their ideal self as opposed to the other components of self image. This deduction logic might lead to the uncovering of many facets of self-image and how destination could be attained.

In general, this study provided support for a strong relationship between self-image and satisfaction, which in turn would lead to destination loyalty. More importantly, it shed light on the role played by satisfaction as a mediator for the ideal self image and not actual in generating future destination loyalty. More importantly, this research sends out the message that program, activities and promotional campaigns, should focus and appeal towards the tourists ideal self, while placing less importance on the actual self. Perhaps, loyalty programs could be drawn to generate repeat businesses and referrals with the intent to capture the actual self-image in generating satisfaction. By concentrating on both aspects of self-image, it is believed that short and long term goals could be achieved in generating loyalty, while still not neglecting the satisfaction which is derived out of their self-images.