INTRODUCTION

Consumers sometimes benefit from special treatment, which is to say, they are granted privileges that others are not offered: access to a special cash desk, invitations to private sales, free alterations to clothing, etc. While some researchers have demonstrated that special treatment can have a positive impact on variables such as attachment to the firm, the desire to maintain the relationship or even word-of-mouth (Lacey, Suh and Morgan, 2007), others have underlined the dangers linked to these practices and, in certain cases, even question their compatibility with the principle of relationship marketing (Hennig-Thurau, Gwinner and Gremler, 2002). It is true that special treatment is a controversial practice that has fuelled heated debates on the issues of discrimination and ethics. It can cause feelings of unfairness, even neglect, among customers who do not enjoy special offers (Hennig-Thurau, Gwinner and Gremler, 2002). Furthermore, it can delight as well as embarrass customers. Some people do not like being openly favored over others or feeling indebted to the firm (de Wulf, Odekerken-Schröder and Iacobucci, 2001). In this article, we assert that these pitfalls can be avoided through an individual and differential approach to special treatment. Indeed, given the variability of reactions to special treatment from one person to another, only an individual approach allows firms to reserve special treatment for consumers who are receptive to this practice.

After a brief review of the literature on special treatment, we will present the results of seven studies which have enabled us to construct and validate a scale to identify consumers who are the most likely (and the least likely) to seek special treatment.

REVIEW OF THE LITERATURE

Definition

Special treatment has been studied in two broad research fields: personalization (Surprenant and Solomon, 1987; Forman and Sriram, 1991; Mittal and Lassar, 1996), and relationship marketing (Berry, 1983; Crosby, 1991; Gwinner, Gremler and Bitner, 1998; de Wulf, Odekerken-Schröder and Iacobucci, 2001, among others). The former have defined it as a means of personalization concerned exclusively with the exchange process (Salerno, 2001), while the latter view it as a practice designed to reward the most loyal customers. However, two approaches coexist in relationship marketing. The first is objective and assimilates special treatment to increased service ("core service upgrading", Berry, 1983; Crosby, 1991) that can be granted (or not) in the context of a loyalty program (Deighton, 2000). The second is subjective and oriented more toward the consumer and his or her perceptions. It does not concern special treatment in and of itself, but rather its perceived benefits. Thus, Gwinner, Gremler and Bitner (1998), in a study aimed at identifying perceived consumer benefits linked to a long-standing relationship with a firm, pinpointed three main types of relational benefits: confidence benefits, social benefits and special treatment benefits. These special treatment benefits have two facets: an economic facet (example of an item: I am entitled to discounts or special offers that most consumers do not get) and a personalization facet (e.g., They offer me a service that most customers do not get). De Wulf, Odekerken-Schröder and Iacobucci (2001) and Odekerken-Schröder, de Wulf and Schumacher (2003) have, on the other hand, made a point of differentiating special treatment from tangible benefits granted to loyal customers. They define special treatment as "a consumer's perception of the extent to which a retailer treats and serves its regular customers better than its non-regular customers" and they measure this with items that completely disregard the economic facet.

In keeping with Kumar and Shah (2004), who propose distinguishing rewards granted according to formalized procedures (level 1 rewards) from rewards granted more informally, at the firm's discretion (level 2 rewards), we propose to distinguish the expression "special treatment" from "preferential treatment" and to define them in the following manner:

- A consumer receives special treatment when he or she is selectively granted non-contractual advantages.

- A consumer receives preferential treatment when he or she is selectively granted contractual advantages.

Special treatment and preferential treatment therefore have content (privileges) that is the product of a selective granting process. Both create a difference between consumers. In the case of preferential treatment, privileges are contractually due. In the case of special treatment, they are non-contractual.

A controversial tool

Until now, special treatment has been systematically considered as a marketing tool used by firms in the context of their customer relationship strategies. Therefore, it has always been a determinant in the models tested. Most research published on the subject also considers that consumers play only a passive role in the granting of privileges, with the firm deciding upstream whether or not to grant certain advantages. Thus, de Wulf, Odekerken-Schröder and Iacobucci (2001) and Odekerken-Schröder, de Wulf and Schumacher (2003) have focused on the influence of special treatment on perceived investment of the firm in the relationship, while Hennig- Thurau, Gwinner and Gremler (2002) have examined its impact on satisfaction, commitment, loyalty and word-of-mouth. Lacey, Suh and Morgan (2007), on the other hand, have studied the link between special treatment and increased sales, customer share, wordof- mouth and consumer feedback.

This research has often produced mixed and even contradictory results. The cause: the variability of reactions to special treatment from one individual to another. Indeed, while some people love attention, others do not appreciate being openly favored (Odekerken-Schröder, de Wulf and Schumacher, 2003). Yet others feel indebted to the firm (de Wulf, Odekerken-Schröder and Iacobucci, 2001; Kang and Ridgway, 1996), which dampens, even cancels out, the benefits of special treatment. Due to the heterogeneity of individual reactions, it is impossible to offer a single answer to the question of whether special treatment is effective or not. This has led some researchers to introduce moderating individual variables in their models. For instance, Salerno (2005) demonstrated that special treatment only had a positive and significant impact among individuals with a strong need for uniqueness (Tian, Bearden and Hunter, 2001). Bahia and Perrien (2003) introduced the notion of the client's relationship proneness,1 which they present as a prerequisite for any customer relationship approach. Some consumers do indeed have a highly developed need for independence (Spears, 2001), which translates into a feeling of "entrapment" when confronted with customer relationship practices. As de Wulf, Odekerken-Schröder and Iacobucci (2001) have pointed out, corporate efforts directed at regular customers only have a positive effect on perceived quality of the relationship among those who are highly predisposed. Finally, Boyd and Helms (2005) have studied consumers who, invariably, consider themselves to be special customers and expect the firm to immediately adapt to their desires (the notion of consumer entitlement). Since such consumers feel they are entitled to more than others, we can expect they are particularly inclined to seek special treatment.

All these studies suggest that traditional segmentation criteria used to select special customers (e.g., loyalty, lifetime value) are not necessarily relevant (Bendapudi and Berry, 1997; Christy, Oliver and Penn, 1996; Day, 2000). Among the most loyal customers, some will never seek special treatment, nor attach particular value to any of the privileges they are granted. To deploy an effective customer relationship strategy, one that optimizes costs by granting special treatment only to those customers who are positively influenced by such practices, it is necessary to adopt an approach that integrates individual differences in the propensity to seek such treatment. A measurement tool is therefore required to develop and confirm the concept. This is the goal of our study.

CONSTRUCTION OF A MEASUREMENT SCALE

Exploratory study

To explore the concept of the consumer's need for special treatment and identify its underlying psychological mechanisms, twelve semi-structured interviews were conducted with individuals identified as particularly inclined or not inclined to seek special treatment in the context of consumption. These individuals, selected by three research professors in marketing on the basis of observed consumer behavior a posteriori, were among the family or friends of the judges solicited. Therefore, they formed an ad hoc sample that we checked for diversity in age, gender and socio-professional categories (See Appendix A1). The sample size was dictated by the semantic saturation criterion (Strauss and Corbin, 1998), which ensures exhaustiveness of the facets of the phenomenon studied.

Before starting each interview, the interviewer rapidly presented the research field (consumer behavior) in order to put the respondent in the consumer's frame of mind. Then, so as not to influence the answers, the theme of the interview was introduced in very broad terms by the following question: In your opinion, what is special treatment? (the complete interview format is presented in Appendix A2).

Thematic content analysis of the material collected (Point and Voynnet-Fourboul, 2006) enabled us to identify three motivations that prompt certain consumers to seek special treatment: the desire to stand out (theme of distinction), the desire to negotiate, for the fun of it, and obtain something other consumers do not get (theme of play) and the desire to take advantage of the tangible benefits that come with special treatment (theme of utility). Table 1 presents the three themes extracted from the content and relates them to the nature of the benefits sought by the consumer.

Each of these themes has revealed a wide range of variability between individuals confirming the need for a differential approach to special treatment. For example, while some respondents appeared to be particularly motivated by the need for distinction, others confessed they were embarrassed, rather than flattered, by being singled out from other consumers (e.g., "I don't like being treated better than other customers"; "I hate openly receiving privileges, even if I am entitled to them").2 Similarly, while some respondents say they are inclined to negotiate, others admit they "can't stand" asking for anything. According to these respondents, special treatment must never be requested, but spontaneously granted at the firm's initiative. Finally, the need for utility has also produced individual differences. Some respondents are completely indifferent to the content of the privileges they could be granted.

Furthermore, three elements confirmed the dispositional character of the need for special treatment: 1) the ease with which the judges called on to define the sample could identify, among their acquaintances, subjects with a very strong or very weak need for special treatment; 2) the capacity of respondents in the sample to identify themselves as having a strong or weak tendency to seek special treatment; 3) the convergence between the way these people perceived themselves and were perceived by others.

Defining the construct

The results of the exploratory study suggest that certain consumers are more or less inclined to seek special treatment and there are three main types of privileges that interest them: symbolic privileges of distinction, fun privileges associated with negotiation games and utilitarian privileges associated with the content of special treatment. Among these utilitarian privileges, saving money has been extensively studied in the marketing literature (Shimp and Kavas, 1984; Blattberg and Neslin, 1990; Schneider and Currim, 1991; Lichtenstein, Netemeyer and Burton, 1995). Furthermore, it has already been linked to an individual variable, deal proneness, for which Lichtenstein, Netemeyer and Burton (1995) and d'Astous and Jacob (2000) have published scales in French and English. Since there are already tools that can be used to identify bargain hunters, we decided to exclude this trait from our study. The consumer's need for special treatment (NST) in the context of this study is therefore defined as:

- a stable and recurring tendency for certain consumers to seek special treatment for reasons other than saving money;

- fed by three motivations: 1) the desire for distinction, i.e., symbolic privileges associated with differences created between consumers through special treatment; 2) the desire for play, i.e., fun privileges associated with the process through which special treatment is granted; 3) the desire for utility, i.e., utilitarian benefits associated with special treatment.

The nature of the construct

In keeping with the recommendations of Crié (2005), we should examine the reflective or normative nature of the NST construct. This depends on the type of relationship that exists between NST and its three dimensions. According to Jarvis, MacKenzie and Podsakoff (2003), a second-order construct is formative when: 1) its first-order dimensions are considered as its defining characteristics and not its manifestations; 2) its first-order dimensions do not necessarily share a common theme; 3) a change in measurement of a first-order dimension does not necessarily entail a change in measurement of all the other dimensions and 4) the first-order dimensions are not assumed to have the same antecedents and consequences. A second-order construct, on the other hand, is reflective if the opposite of these assertions is true.

NST is fed by the desire for distinction, play and utility. Consequently, a change or the elimination of one of these dimensions would alter the construct. For example, integrating financial benefits in the desire for utility would broaden its conceptual scope; removing the allure of fun during the negotiation process would considerably reduce the dimension of play. Assertion 1) therefore applies to NST. Moreover, the desire for distinction, play and utility are distinct concepts. We can therefore expect that some individuals seek special treatment exclusively as a way to stand out, while others are only interested in the utilitarian aspects of the privileges granted. Due to this conceptual independence, the antecedents behind seeking distinction, play or utility could be quite different, as well as the types of special treatment sought after. Assertions 2), 3) and 4) also apply to NST. Consequently, NST is a second-order construct formed by three first-order constructs: the desire for distinction, the desire for play and the desire for utility. To measure it, we have adopted a type II model from the typology defined by Jarvis, Mackenzie and Podsakoff (2003): reflective first-order, formative second-order (See Figure 1).

To construct the reflective part of the scale (firstorder), traditional procedures taken from classic test theory (Nunnally, 1978) and recommended by Churchill (1979) were followed: generating a set of statements, factor analysis, testing the internal consistency of indicators and intra-construct correlations (Studies 1 and 2). The formative part of the scale (second-order), on the other hand, was constructed according to the recommendations of Diamantopoulos and Winklholfer (2001) by ensuring the absence of multicollinearity between first-order dimensions and examining external validity (Study 3).

Study 1: Developing and purifying the first-order scales

Starting from a qualitative study and the literature, 34 items3 assumed to measure each of NST's three dimensions were formulated and submitted for evaluation by three experts, all research professors in marketing. After studying the concept of need for special treatment, the experts were asked to attribute each item to one of four categories: 1) representative of the Distinction dimension, 2) representative of the Play dimension, 3) representative of the Utility dimension and 4) representative of none of these three dimensions. They were also asked to cite any redundant items. This classification, which produced an agreement rate between judges of 0.70 (or a satisfactory PRL index of 0.82, Rust and Cooil, 1994), led to the elimination of seven items (four non-representative and three redundant).

The 27 remaining items were administered with 254 individuals solicited during train journeys. Thirty questionnaires were discarded because they were incorrectly filled in or incomplete, reducing the final sample size to 224 (60% male with an average age of 34). After eliminating items with a Pearson's correlation coefficient greater than 0.8 with one or more items on the list, principal components analyses were conducted in order to purify the list even further. Indeed, with more than three to five items per dimension, not only did the scale risk discouraging both practitioners and respondents (Rossiter, 2002), but also artificially increasing the correlations between the construct measured and other constructs, creating what Feldman and Lynch (1988) refer to as selfgenerated validity. It was therefore important to produce a concise scale.

PCA allowed us to eliminate, through an iterative process, poorly represented items (communality lower than 0.5), those that contributed little to factor definition (saturation lower than 0.5) and those that contributed to define several factors (difference of saturation between two factors lower than 0.3). At the end of this process, 17 items were discarded and a three-dimensional structure, corresponding to the three themes identified during the exploratory study, was created using the Kaiser criterion. Table 2 shows the factor loadings of the ten items in these dimensions. 4

The Distinction dimension explains 41.84% of the variance, the Play dimension 16.6% and the Utility dimension 14.37%, for a total of 72.37%. Furthermore, these dimensions are weakly correlated (r = 0.314 between Distinction and Utility, r = 0.252 between Distinction and Play, and r = 0.309 between Play and Utility), which confirms the choice of a formative second-order model (Jarvis, MacKenzie and Podsakoff, 2003). Finally, the values of their Cronbach alphas indicate good internal consistency (αDistinction = 0.865; αPlay = 0.822; αUtility = 0.784).

Study 2: Confirmation and validation of the scale's first-order structure

To confirm the scale's first-order structure and ensure its validity, a second study was performed. In an effort to achieve concision, only ten items were selected after the purification phase. The questionnaire was administered with 172 people in conditions similar to those in Study 1 (during a long train journey). Four questionnaires were discarded because they were incomplete, reducing the final sample size to 168 (52% male with an average age of 32).

After reproducing, thanks to PCA, the factorial structure obtained in Study 1, a confirmatory factor analysis was conducted to confirm the psychometric qualities of the scale (Kline, 2005). This analysis was performed with AMOS software on a concatenation of samples 1 and 2 (N = 392). Since the conditions of use had been filled and the distributions were considered normal (|symmetry| < 1 and |kurtosis| < 1.5), the model shown in Figure 2 was estimated using the maximum likelihood method. Its adjustment indicators and the commonly admitted empirical thresholds used to confirm its goodness of fit with the data are shown in Table 3.

All the items have standardized weights significant at p < 0.001 and squared multiple correlations greater than 0.5. They are therefore valid and reliable (MacKenzie, Podsakoff and Jarvis, 2005). Furthermore, since the model was well-adjusted to the data, the convergent validity, discriminant validity and reliability of its three sub-scales could be evaluated.

The convergent validity was established according to the recommendations of Fornell and Larcker (1981), i.e., by verifying that the rho of convergent validities were greater than 0.5 (ρCV Distinction = 0.588; ρCV Play = 0.612; ρCV Utility = 0.582). Discriminant validity was studied by comparing the rho of convergent validity with percentages of variance shared by dimensions (Fornell and Larcker, 1981).5 Finally, the reliability of the sub-scales was established by ensuring the values of Jöreskog's rho (1971) were all greater than 0.7 (ρJöreskog Distinction = 0.851; ρJöreskog Play = 0.825; ρJöreskog Utility = 0.805). In the end, the second study allowed us: 1) to validate the first-order structure of the measurement model and 2) ensure the intra-construct validity and reliability of the three sub-scales.

Study 3: Intra-construct validation of the scale's second-order structure

To confirm the formative part of the model (second-order), it was important to check two points: 1) the absence of multicollinearity between firstorder dimensions and 2) the external validity of the measurement (Diamantopoulos and Winklholfer, 2001). To this end, Diamantopoulos and Winklholfer (2001, p. 272) and Jarvis, MacKenzie and Podsakoff (2003, p. 213) suggest estimating a structural equations model linking the second-order construct to at least two theoretically appropriate reflective indicators.7 A satisfactory overall fit of the model is therefore a validation criterion for all the indicators making up the index (Crié, 2005, p. 12).

Five items reflecting the consumer's need for special treatment were generated (Table 4) and then subject to a face validity test similar to the one used for the NST items. The test was conclusive: all the items were unanimously deemed representative of effects of NST.

The five items, as well as the three sub-scales of NST, were then administered with a sample of 315 individuals (48% male with an average age of 24), including 209 students doing a Master's degree in management and 106 non-students solicited during a long journey by train. To limit demand effects (Herbert, 2005), the items measuring NST effects were mixed with those of the NST scale.

In an initial phase, the absence of multicollinearity between the three dimensions of NST was checked by examining VIF values (VIFDistinction = 1.268; VIFPlay = 1.154; VIFUtility = 1.152 < 10). Then, the model presented in Figure 3 was estimated with AMOS using the maximum likelihood method (the standardized weights are all significant). It is important to note that two of the five items reflecting the effects of NST were eliminated due to their weak averages and standard deviations.8

Since the model fits the data well (Table 5), we can confirm the intra-construct validity of the global NST scale.

Study 4: Reliability

To assess the reliability of the NST scale, we decided to use the test-retest technique. A questionnaire with the ten NST items was administered with a sample of students enrolled in a management degree program at 2-week intervals. The first session allowed us to collect 132 questionnaires and the second 114. Due to student absences and errors in filling out the forms, the correlation between scores at t and t + 2 weeks could only be calculated for 87 observations. These correlations reached 0.72 for the Distinction and Play dimensions, 0.71 for Utility and 0.71 for the global scale, which is satisfactory. We can therefore confirm the scale's temporal stability.

VALIDATING THE SCALE

Studies 1 to 4 enabled us to construct a scale for measuring NST and establish its intra-construct validity and reliability. Then, it was necessary to examine its inter-construct validity (Study 5) and its nomological validity (Study 6).

Study 5: Convergent and discriminant validity

There is no measurement of need for special treatment in the marketing literature. To evaluate the convergent validity of the NST scale, we therefore examined its link with a similar conceptual construct: consumer entitlement. This trait designates "the extent to which a customer expects special treatment in a retail environment" (Boyd and Helms, 2005). Two points distinguish this concept from the need for special treatment: 1) the importance given to the perceived legitimacy of privileges and 2) the role of the consumer in the privilege granting process. While a consumer with a high score on the consumer entitlement scale really feels he has the right to special treatment, he does not necessarily actively seek it.

To evaluate the convergent validity of the NST scale, we therefore examined its link with a distinct construct: smart shopping (Lichtenstein, Netemeyer and Burton, 1995; Mano and Elliot, 1997). This trait designates the "tendency for consumers to invest time and effort in seeking and utilizing promotional information to achieve price savings" (Mano and Elliot, 1997, p. 504). Its highly material overtones (overtones we decided to intentionally omit from the NST field) make it a conceptually distinct construct from NST.

Procedure

To avoid making the questionnaire too heavy, and ensure the quality of the data collected, two versions were created. The first contained the NST and smart shopping scales10 (Lichtenstein, Netemeyer and Burton, 1995). It was administered with 101 students enrolled in a management degree program. The second contained the NST and consumer entitlement scales (Boyd and Helms, 2005).11 This second version was administered with 301 students (one half enrolled in a Master's degree in management in Paris, and the other half in a medical school in Nice).

Results

After testing the dimensionality of the scales and their reliability (αConsumer entitlement = 0.67 and αSmart shopping = 0.82), correlational and variance analyses were conducted. For each variable, two groups were constructed: the first group was made up of individuals whose scores were in the lower quartile of the distribution and the second group included individuals whose scores were in the upper quartile. Then, the means of the two groups in the NST scale and its three dimensions were compared using t-tests. Table 6 shows the results of these tests as well as those of the correlational analyses (Levene's tests are indicated to check the homogeneity of intra-group variances).

Consumer entitlement is moderately and significantly correlated with NST, while smart shopping is not. Furthermore, t-tests indicate that consumers with the highest scores on the consumer entitlement scale are much more likely to seek preferential treatment than consumers with the lowest scores (MNST = 4.18 vs. 3.04; t = -7.957; p < 0.001), which is not the case for smart shoppers (MNST = 3.85 vs. 3.51; t = -1.356; p = 0.18). The NST scale therefore presents good convergent and discriminant validities.

Study 6: Nomological validity

To establish the nomological validity of the NST scale, we used Mowen's personality model (Mowen, 2000), according to which personality traits can be classified according to their omnipresence in individual behaviors. A distinction is made between elementary traits, which are the deepest predispositions, composite traits, which stem from a combination of elementary traits and environmental factors, situational traits, produced by combining higher-level traits and the characteristics of certain situations, and surface traits, which depend the most on context. Since the need for special treatment refers to consumption situations, it is a situational trait. Thus, we decided that it was relevant to link it to two types of traits: higher-level traits, which could be considered antecedents, and traits on the same level. In addition to their contribution to nomological validation, these variables allow us to refine the psychological profile of individuals with a need for special treatment.

The antecedents of NST

Three antecedents of NST were studied: ruleconsciousness (Cattell, 1946; 1957), self-monitoring (Snyder, 1974) and narcissism (Raskin and Hall, 1979; 1981)

Rule-consciousness reflects the degree to which social rules defining "right" and "wrong" are interiorized and dictate conduct (Cattell, 1946; 1957). Subjects who comply with social conventions are therefore those who insist on the importance of following rules. Conversely, those who have little respect for conventions tend to avoid the law and bend the rules, either because their moral values are poorly interiorized, or they believe in values that are not exclusively based on compliance with conventions. Since special treatment is informal, and sometimes granted despite rules, we can expect that it would be more highly prized by people less concerned by compliance with social conventions.

Self-monitoring (Snyder, 1974) is a multidimensional trait that designates an individual's awareness of social cues and ability to adapt expressive behaviors and representations (the "Acting" dimension of Briggs, Cheek and Buss, 1980). High self-monitors are particularly at ease in social situations and capable of carefully regulating their behavior to present themselves in the best possible light. Excellent actors, they can assert contradictory ideas without difficulty or reservations. We can therefore expect that the quest for preferential treatment, especially if it involves bargaining, would be particularly easy, and even entertaining, for these individuals. This supposes that there is a link between the Acting dimension of self-monitoring and NST.

Finally, narcissism characterizes individuals with an excessive need for admiration and a feeling of superiority that translates into a firm belief they are entitled to special treatment. To satisfy their need for attention, narcissists are constantly searching for situations that can improve their self-image (Bogart, Benotsch and Pavlovic, 2004). Since special treatment is inherently selective (it creates a distinction between consumers that reflects well on the privileged person), it should be particularly sought after by narcissists.

The correlates of NST

Parallel to the antecedents of NST, three situational traits were studied and associated with the dimensions of NST: the consumer's need for uniqueness (Tian, Bearden and Hunter, 2001), bargaining proneness (Harris and Mowen, 2001) and market mavenism (Feick and Price, 1987).

Individuals with a strong need to feel unique (Tian, Bearden and Hunter, 2001) seek differentiation as an ultimate goal through the ostentatious acquisition and use of goods. Distinction from other consumers, thanks to special treatment granted in a highly selective and visible way, can enable them to develop their individuality. Since the pleasure of bargaining is an intrinsic value in the search for special treatment (appreciated in and of itself), it is not the negotiation itself, but rather its outcome, that allows the consumer to differentiate himself. We can therefore expect that there is a positive link between the consumer's need for uniqueness and the Distinction and Utility dimensions of NST.

Bargaining proneness characterizes individuals strongly inclined to haggle with salesmen to get a discount (Harris and Mowen, 2001). This stems from two motivations: a utilitarian motivation concerning the content of the negotiation (the price difference) and a more playful motivation associated with the negotiation game (Rubin and Brown, 1975). Since we have excluded the quest for purely financial benefits from the concept of NST in our definition, only these entertaining benefits can be used to establish a link between bargaining proneness and the Play dimension of NST.

Finally, market mavenism (Feick and Price, 1987) characterizes consumers who are the most knowledgeable about every aspect of the market (products, innovations, outlets), but also the most inclined to share information with friends and acquaintances. Since these consumers know the market better than anyone else, they are likely to see opportunities for special treatment that most other consumers miss, and therefore they could be particularly inclined to detect ways of improving their consumer experience. They are not interested in negotiation games or standing out from other consumers, but the actual content of the special treatment. This is why we expect a positive link between the Utility dimension of NST and market mavenism.

Procedure

Whenever possible, existing scales in French were used. Rule-consciousness was measured using a well-known personality assessment inventory used in psychology: the 16PF (Sixteen Personality Factor Questionnaire, Cattell, 1946, 1957), the self-monitoring scale (the French version translated and validated by Gana and Brechenmacher, 2001, of Snyder and Gangestad's scale, 1986), the consumer's need for uniqueness (the French version, validated by Salerno, 2005, of Tian, Bearden and Hunter's uniqueness scale, 2001). Finally, the scales of Ames, Rose and Anderson (2006),12 Mowen and Harris (2001) and Feick and Price (1987) were used to measure, respectively, narcissism, bargaining proneness and market mavenism. These three scales were translated into French using the parallel-blind translation technique (Mayer, 1980).

Considering the number and length of the scales, several versions of the questionnaire were administered. In all, 510 college students in business or medical school were asked to fill out forms during their classes.13 This took an average of 15 minutes.

Results

After testing the dimensionality of the scales and their reliability (See Appendix A3), correlational and variance analyses similar to those in Study 6 were conducted. The results are presented in Table 7.

Except for rule-consciousness, whose link with NST is weaker than expected, all the expected relationships were confirmed empirically. This study therefore enables us to ensure the nomological validity of the NST scale and also to gain a better understanding of the psychological profiles of consumers interested in special treatment.

Table 8 presents a summary of all the phases used to construct and validate the NST scale.

CONCLUSION, LIMITATIONS AND AVENUES FOR FUTURE RESEARCH

Even though it is increasingly used by firms, special treatment has not been widely studied in marketing research (Lacey, Suh and Morgan, 2007). However, the interindividual variability it creates makes it a difficult tool to use and therefore a particularly interesting research topic. To help firms use special treatment effectively, this article proposes a new instrument: a scale for measuring the consumer's need for special treatment, whose validity and reliability have been confirmed in six studies. This tool has a dual interest. On the one hand, it can be used to identify consumers who are the most likely (and the least likely) to seek special treatment. Therefore, firms can select consumers who are the most receptive to special treatment and exclude those who are not. From this standpoint, it is an effective targeting and segmentation tool. On the other hand, it can be used to match each consumer profile with a type of special treatment, thus optimizing resources devoted to this practice. As suggested by the psychological profile depicted in Study 6, the motivations underlying the search for special treatment are linked to different determinants. This means that firms need to differentiate both the content of special treatment and processes by which it is granted.

In practice, the NST scale could be administered via a registration form or a customer satisfaction survey, when possible at the same time as a promotion awareness campaign. The goal is to clearly identify all facets of the need for special treatment and to have a very precise vision of the customers studied. Then, different types of privileges could be granted to the segments identified.

To best satisfy consumers with a strong need for special treatment for fun, privileges should be granted in an informal manner, through a process involving a negotiation game. Since these consumers are particularly at ease in social situations, contact persons should be trained in distinguishing "bluffing" from genuine demands for special treatment.

Consumers who seek special treatment as a means of distinction should be granted privileges in a more proactive way, if possible by accentuating the exclusivity of the special treatment offered. Once again, contact staff play a key role: they need to satisfy the consumer's often excessive need for attention while managing the typically aggressive streak of narcissistic personalities.

Finally, since consumers who seek special treatment for utilitarian reasons are market mavens, firms must make sure they fully appreciate the utility of the privileges granted. With these consumers, distinction or negotiation games are a waste of time.

Despite the care and efforts devoted to the studies conducted during this research, certain limitations are worth mentioning. The first concerns the conceptual territory of NST. Although the search for savings was mentioned by several of the respondents during the qualitative study, we voluntarily chose to exclude this facet from the construct because it had already been used to create scales (e.g., Lichtenstein, Netemeyer and Burton, 1995; d'Astous and Jacob, 2000). This decision limits the theoretical and managerial impact of the NST scale. This implies, as we have seen, the use of the NST scale alongside another measurement tool to assess the strictly economic facet of special treatment.

The second limitation concerns the approach used to validate the NST scale. To establish the scale's convergent and discriminant validities, we studied its link with a similar construct (consumer entitlement, Boyd and Helms, 2005) and a completely distinct one (the smart shopping trait, Lichtenstein, Netemeyer and Burton, 1995). These correlational analyses should be completed by the use of a MTMM (Multi-Traits Multi-Methods) matrix (Campbell and Fiske, 1959).

The third limitation concerns the NST scale's ecological validity (Brunswik, 1949), in particular its Utility dimension. To identify individuals motivated by the utilitarian benefits of special treatment, we formulated items applicable in the retail sector, in particular to firms that organize private events. This contextualization, which we preferred in order to preserve the construct's content (the privileges associated with special treatment are very specific, but also vary a great deal, from one sector to another), limits the scale's scope. A promising research topic would consist in reproducing the factor structure of the NST scale by adapting its items to the specific characteristics of the business sector in which they are used. For example, items such as In general, I try to wait in the airline's private lounge because it's quieter or I like to receive special offers because they allow me to learn more about the destinations served could be applied to the air travel sector.

A fourth limitation concerns the interrelations that could exist between the variables used to establish the nomological validity of the NST scale. Since several versions of the questionnaire were administered, only univariate relationships could be established between NST and the determinants studied. It would therefore be interesting to organize a global data collection in order to establish a consumer typology. Linking the segments identified to easily observable features (e.g., age, gender, socioprofessional category, customer history, etc.) could provide firms with segmentation criteria that are particularly simple to use.

Finally, given the multidimensional character of the need for special treatment, a particularly interesting research theme would consist in linking the NST construct and its three dimensions to different forms of special treatment. Such an approach would not only demonstrate the scale's predictive validity, but also generate tangible managerial recommendations.

1. The client's relationship proneness designates, according to Bahia and Perrien (2003), "a client's enduring tendency to a) expect a relationship approach from a given company, and b) appreciate this approach".

2. These comments are from an interview with C.B. (female, age 51) and A.B.L. (male, age 30), both considered by their friends and family as having a weak need for special treatment.

3. In a study of 200 scales published in the reference work Handbook of marketing scales (Bearden and Netemeyer, 1999), Hardesty and Bearden (2004) indicate that the initial number of items generated in the development phase varies between 10 and 200: 34 items therefore seemed sufficient.

4. These dimensions were extracted thanks to an oblique Oblimin rotation. Indeed, given the exploratory nature of this initial study, we wanted to have an estimate of the correlations between factors in order to examine their independence. It should be noted that an orthogonal rotation (Varimax) produced an identical structure.

5. r(Distinction, Play)2 = 0.362 < 0.588 and 0.612; r(Play, Utility)2 = 0.352 < 0.612 and 0.582; r(Distinction, Utility)2 = 0.332 < 0.588 and 0.582.

6. The standardized weights are indicated in normal typeface and the bivariate correlations in italics.

7. When a model only includes a second-order construct, the first-order constructs and their indicators, it is indeed not identified because the composite variable does not have a direct effect on the variable observed (Crié, 2005, p. 14). To eliminate this identification problem, the secondorder construct (in this case NST) must emit at least two different paths toward reflective variables or be linked to reflective indicators (McCallum and Browne, 1993).

8. The items Effect_3 and Effect_4 presented very weak standard deviations compared to the three other items (1.21 and 1.32 compared to 1.79 minimum for Effect_1), which is the sign of weak discriminatory power. To restrict the number of parameters estimated in the model, we therefore decided to eliminate them.

9. The standardized weights are indicated in normal typeface and the bivariate correlations in italics.

10. The consumer entitlement and smart shopper scales were translated into French using the parallel-blind translation technique (Mayer, 1980). They are presented in Appendix A3.

11. These questionnaires also included part of the variables examined in Study 6 in order to assess the nomological validity of the NST scale.

12. Ames, Rose and Anderson's scale (2006) is a short version of Raskin and Hall's 16-item Narcissistic Personality Inventory (1979).

13. Among these 510 individuals, 402 were part of the sample in Study 5. The additional 108 individuals were all business students.

BIBLIOGRAPHICAL REFERENCES

Ames D.R., Rose P. and Anderson C.P. (2006), The NPI-16 as a short measure of narcissism, Journal of Research in Personality, 40, 4, 440-450.

Anderson J.C. and Gerbing D.W. (1988), Structural equation modelling in practice: a review and recommended twostep approach, Psychological Bulletin, 103, 3, 411-423.

d'Astous A. and Jacob I. (2000), Une étude des réactions des consommateurs aux offres promotionnelles comportant une prime, in R. Michon, J.-C. Chebat and F. Colbert (coord.), Proceedings of the 16th Congrès international de l'Association Française du Marketing, Montréal, École des Hautes Études Commerciales, 429-440.

Bahia K. and Perrien J. (2003), Les conséquences de la prédisposition relationnelle du client, Proceedings of the 19th Congrès de l'Association Française du Marketing, Gammarth, Tunis.

Bearden W.O. and Netemeyer R.G. (1999), Handbook of marketing scales: multi-item measures for marketing and consumer behaviour research, Thousand Oaks, Sage Publications.

Bendapudi N. and Berry L.L. (1997), Customers' motivations for maintaining relationships with service providers, Journal of Retailing, 73, 1, 15-37.

Berry L.L. (1983), Relationship marketing, in L.L Berry, G.L. Shostack and G.D. Upah (eds.), Emerging perspectives on services marketing, Chicago, American Marketing Association, 25-28.

Blattberg R.C. and Neslin S.A. (1990), Sales promotion: concepts, methods, and strategies, Englewood Cliffs, Prentice Hall.

Bogart L.M., Benotsch E.G. and Pavlovic J.D. (2004), Feeling superior but threatened: the relation of narcissism to social comparison, Basic and Applied Social Psychology, 26, 1, 35-44.

Boyd H.C. and Helms J.E. (2005), Consumer entitlement: theory and measurement, Psychology & Marketing, 22, 3, 271-286.

Briggs S.R., Cheek J.M. and Buss A.H. (1980), An analysis of the self-monitoring scale, Journal of Personality and Social Psychology, 38, 4, 679-686.

Brunswik E. (1949), Perception and the representative design of psychological experiments, Berkeley, University of California Press.

Campbell D.T. and Fiske D.W. (1959), Convergent and discriminant validation by the multitrait-multimethod matrix, Psychological Bulletin, 56, 2, 81-105.

Cattell R.B. (1946), The description and measurement of personality, New York, Harcourt, Brace & World.

Cattell R.B. (1957), Personality and motivation structure and measurement, New York,World Book.

Christy R., Oliver G. and Penn J. (1996), Relationship marketing in consumer markets, Journal of Marketing Management, 12, 1-3, 175-187.

Churchill G.A. (1979), A paradigm for developing better measures of marketing constructs, Journal of Marketing Research, 16, 1, 64-73.

Crié D. (2005), De l'usage des modèles de mesure réflectifs ou formatifs dans les modèles d'équations structurelles, Recherche et Applications en Marketing, 20, 2, 5-25.

Crosby L.A. (1991), Building and maintaining quality in the service relationship, in S.W. Brown, E. Gummesson, B. Edvardsson and B. Gustavsson (eds.), Service quality: multidisciplinary and multinational perspectives, Lexington, Lexington Books, 269-287.

Day G.S. (2000), Managing market relationships, Journal of the Academy of Marketing Science, 28, 1, 24-30.

Deighton J. (2000), Frequency programs in service industries, in T.A. Swartz and D. Iacobucci (eds.), Handbook of services and management, Thousand Oaks, Sage Publications, 401-407.

Diamantopoulos A. and Winklholfer H.M. (2001), Index construction with formative indicators: an alternative to scale development, Journal of Marketing Research, 38, 2, 269-277.

Feick L.F. and Price L.L. (1987), The market maven: a diffuser of marketplace information, Journal of Marketing, 51, 1, 83-97.

Feldman J.M. and Lynch Jr. J.G. (1988), Self-generated validity and other effects of validity on belief, attitude, intention, and behaviour, Journal of Applied Psychology, 73, 3, 421-435.

Forman A.M. and Sriram V. (1991), The depersonalization of retailing: its impact on the "lonely" consumer, Journal of Retailing, 67, 2, 226-243.

Fornell C. and Larcker D.F. (1981), Structural equation models with unobservable variables and measurement error: algebra and statistics, Journal of Marketing Research, 18, 3, 382-388.

Gana C. and Brechenmacher N. (2001), Structure latente et validité de la version française du self-monitoring scale : échelle de monitorage de soi, L'Année Psychologique, 101, 3, 393-420.

Gwinner K.P., Gremler D.D. and Bitner M.J. (1998), Relational benefits in services industries: the customer's perspective, Journal of the Academy of Marketing Science, 26, 2, 101-114.

Hardesty D.M. and Bearden W.O. (2004), The use of experts judges in scale development: implications for improving face validity of measures of unobservable constructs, Journal of Business Research, 57, 2, 98-107.

Harris E.G. and Mowen J.C. (2001), The influence of cardinal- central and surface-level personality traits on consumers' bargaining and complaint intentions, Psychology & Marketing, 18, 11, 1155-1185.

Hennig-Thurau T., Gwinner K.P. and Gremler D.G. (2002), Understanding relationship marketing outcomes, Journal of Service Research, 4, 3, 230-248.

Herbert M. (2005), Comportement de réponse de l'individu en situation de questionnement : le biais du répondant en comportement du consommateur, Doctoral dissertation, University of Paris-Dauphine, Paris.

Jarvis C.B., MacKenzie S.B. and Podsakoff P.M. (2003), A critical review of construct indicators and measurement model misspecification in marketing and consumer research, Journal of Consumer Research, 30, 2, 199-218.

Jöreskog K.G. (1971), Statistical analysis of sets of congeneric tests, Psychometrika, 36, 109-133.

Kang Y.S. and Ridgway N.M. (1996), The importance of consumer market interactions as a form of social support for elderly consumers, Journal of Public Policy & Marketing, 15, 1, 108-117.

Kline R.B. (2005), Principles and practice of structural equation modeling, New York, Guilford.

Kumar V. and Shah D. (2004), Building and sustaining profitable customer loyalty for the 21st century, Journal of Retailing, 80, 4, 317-330.

Lacey R., Suh J. and Morgan R.M. (2007), Differential effects of preferential treatment levels on relational outcomes, Journal of Service Research, 9, 3, 241-256.

Lichtenstein D.R., Netemeyer R.G. and Burton S. (1995), Assessing the domain specificity of deal proneness: a field study, Journal of Consumer Research, 22, 3, 314- 326.

MacKenzie S.B., Podsakoff P.M. and Jarvis C.B. (2005), The problem of measurement model misspecification in behavioral and organizational research and some recommended solutions, Journal of Applied Psychology, 90, 4, 710-730.

Mano H. and Elliott M.T. (1997), Smart shopping: the origins and consequences of price savings, in M. Brucks and D.J. MacInnis (eds.), Advances in Consumer Research, Provo, Association for Consumer Research.

Mayer C.S. (1980), Multinational marketing research: methodological problems, in H. Tborelli and H. Becker (eds.), International marketing strategy, New York, Pergamon Press, 162-171.

McCallum R.C. and Browne M.W. (1993), The use of causal indicators in covariance structure models: some practical issues, Psychological Bulletin, 114, 3, 533-541.

Mittal B. and Lassar W.M. (1996), The role of personalization in service encounter, Journal of Retailing, 72, 1, 95-109.

Mowen J.C. (2000), The 3M model of motivation and personality: theory and empirical applications to consumer behaviour, Norwell, MA, Kluwer Academic Publishing.

Nunally J.C. (1978), Psychometric theory, 2nd edition, New York, McGraw-Hill.

Odekerken-Schröder G., de Wulf K. and Schumacher P. (2003), Strengthening outcomes of retailer-consumer relationships: the dual impact of relationship marketing tactics and consumer personality, Journal of Business Research, 56, 1, 177-190.

Point S. and Voynnet-Fourboul C. (2006), Le codage à visée théorique, Recherche et Applications en Marketing, 21, 4, 61-78.

Raskin R.N. and Hall C.S. (1979), A narcissistic personality inventory, Psychological Reports, 45, 2, 590.

Raskin R.N. and Hall C.S. (1981), The narcissistic personality inventory: alternate form reliability and further evidence of construct validity, Journal of Personality Assessment, 45, 2, 590-621.

Rossiter J.R. (2002), The C-OAR-SE procedure for scale development in marketing, International Journal of Research in Marketing, 19, 4, 305-335.

Rubin J.Z. and Brown R.R. (1975), The social psychology of bargaining and negotiation, New York, Academic Press.

Rust R.T. and Cooil B. (1994), Reliability measures for qualitative data: theory and implications, Journal of Marketing Research, 31, 1, 1-14.

Salerno A. (2001), Une étude empirique des relations entre personnalisation, proximité dyadique et identité de clientèle, Recherche et Applications en Marketing, 16, 4, 25-46.

Salerno A. (2005), L'individualité du consommateur : développement du construit et analyse de son rôle dans l'efficacité des pratiques de singularisation, Proceedings of the 21st Congrès de l'Association Française du Marketing, Nancy.

Schneider L.G. and Currim I.S. (1991), Consumer purchase behaviors associated with active and passive deal-proneness, International Journal of Research in Marketing, 8, 3, 205-222.

Shimp T.A. and Kavas A. (1984), The theory of reasoned action applied to coupon usage, Journal of Consumer Research, 11, 3, 795-809.

Snyder M. (1974), Self-monitoring of expressive behaviour, Journal of Personality and Social Psychology, 30, 4, 526-537.

Snyder M. and Gangestad S. (1986), On the nature of selfmonitoring: matters of assessment, matters of validity, Journal of Personality and Social Psychology, 51, 1, 125-139.

Spears R. (2001), The interaction between the individual self and the collective self, in C. Sedikides and M.B. Brewer (eds.), Individual self, relational self, collective self, Philadelphia, Taylor and Francis, 171-198.

Strauss A. and Corbin J. (1998), Basics of qualitative research: techniques and procedures for developing grounded theory, Thousand Oaks, Sage Publications.

Surprenant C.F. and Solomon M.R. (1987), Predictability and personalization in the service encounter, Journal of Marketing, 51, 2, 86-96.

Tian K.T., Bearden W.O. and Hunter G.L. (2001), Consumers' need for uniqueness: scale development and validation, Journal of Consumer Research, 28, 1, 50-66.

de Wulf K., Odekerken-Schröder G. and Iacobucci D. (2001), Investments in consumer relationships: a crosscountry and cross-industry exploration, Journal of Marketing, 65, 4, 33-50.

Author affiliation:

Raphaëlle Butori

Assistant Professor

ESSEC

The author would like to thank Jean-Louis Chandon, editor in chief of the journal, and the four anonymous reviewers for their constructive comments. She would also like to thank Christian Pinson for his valuable advice and support.

The author can be contacted at the following e-mail address: butori@essec.fr

(ProQuest: Appendix omitted.)