Image Transfer in Sports Creative Sponsorship and Participation Sponsorship: What’s the Difference?

Participation sponsorship and creative sponsorship of sports events can have strategic communication effects. In this study of the car brand “Peugeot” in Tunisia, the impact of participation sponsorship on image transfer was compared with that of creative support, considering the moderating effect of the involvement of spectators on the image transfer process. Structural equation modeling analysis was used to validate the measurement of images and structural models of image transfer. The results largely favored sports-related creative sponsorship. Furthermore, when moving from “low involvement” to “high involvement” groups, observers of the creative backing were perceived as more intense, indicating that participation sponsorship has a decreasing impact. Research results have shown that it can be a more effective communication tool than multi-sponsoring. We conclude that creative support presents a real image transfer opportunity for sponsors that is less costly but more persuasive than participation. This assumption deserves to be verified for the indirect sports audience and in other cases of events with high and low media coverage.


Introduction
Events are at the heart of sports systems (Ferrand et al., 2006). This explains why the number of sports events and their frequencies have multiplied and diversified considerably (Nuseir, 2020). Indeed, the value of the sporting event lies in both the uncertain performance and the emotion shared between participants and spectators (Desbordes, 2004;Desbordes & Falgoux, 2007;Ford, 2018). Indeed, the value of the sporting event lies in both the uncertain performance and the emotion shared between participants and spectators (Desbordes, 2004;Desbordes & Falgoux, 2007;Ford, 2018). In contrast, events also produce experiences for consumers. These provide a variety of symbolic and aesthetic experiences as well as hedonistic responses (Filser, 2002). This consumer experience is a subjective state with positive and negative meanings based on interactions with other individuals who are not necessarily consumers (Cova & Cova, 2002). According to Benavent and Evrard (2002, p. 2), this dramatization means that the value of the experience no longer resides solely in the characteristics of the good but in its staging and the value that the consumer brings to it through his or her interpretation. Therefore, sports events are considered the main communication tools for companies and constitute a strategic element of their marketing mix (Koronios et al., 2022).

Conceptual Model of Brand Transfer
According to Schwebig (1988), for most sponsors, sports sponsorship actions are characterized by a focus on the image. It as a set of symbolic and functional attributes, perceived by consumers (Al-Nsour & Al-Otoum, 2019). The individual associates everything with the brand, allowing them to recognize it. Keller (1993) groups these associations into "attributes" linked to the descriptive features of the product, into functional, experiential, or symbolic "benefits" of the product, and into "attitudes" relating to an overall evaluation of the brand. This identification, combined with the objective and subjective character of the brand, has led researchers to study it using several approaches. The first is purely cognitive (Dobni & Zinkhan, 1990), and the second is exclusively affective (Baux, 1991;Walliser, 1994). However, because of its highly developed emotional aspect, several sports researchers believe that an image is a concentration of cognitive and affective aspects (Didellon, 1998;Ganassali & Didellon, 1996;Speed & Thompson, 2000). Other studies have also noted the effects of reverse causality in the transfer process (Derbaix & Pham, 1989), referred to in the literature as "the affective halo effect", which is a positive relationship between the affective component of the event and the cognitive component of the sponsor.
Another additional relationship is considered between the cognitive component of the event image and the affective dimension of the sponsor's image, constituting the classical halo effect. McKenzie et al. (1986) and McKenzie and Lutz (1989) think that there is a link between how people think about the event and how they feel about it. feel about it. Baux (1991) mentioned the possibility of an affective reaction to the sponsor involving a cognitive response. Considering these different relations, the conceptual model of retained image transfer is composed of six structural relationships. These relationships represent the six hypotheses that need to be verified for each action.

Methodology
A quantitative methodological approach was used, based on the case study. Although it is specific and poses a problem of external validity for some researchers (Eisenhardt, 1989), others consider it an advantage, allowing us to focus on a precise phenomenon without concern for generalization. The main strength of a case study is its internal validity and real representation of the phenomenon studied (Woodside & Wilson, 2003). Eisenhardt (1991) suggested that the cases chosen must be comparable in terms of resources, relational reputations, and organizational capacities. According to Rouse and Daellenbach (2002), this choice is a function of three criteria: the logistical link to the mastery of the geographical and cultural area, the economic link corresponding to the considerable financial spin-offs, and the temporal link associated with the proximity to the field and the data.
Using these factors, the decision was made to award the contract to the Tunisian branch of the French automaker Peugeot, which hosts the annual "Les Rencontres Peugeot Roland Garros" tennis tournament and takes part in the worldwide "Tunis Open." The first was a gathering that the corporation suggested its dealers in a number of African nations hold. Tennis fans in pairs competed in a mixed doubles tournament (a man and a woman). The matches are staged in a joyful and pleasant atmosphere, attracting not only tennis enthusiasts but also public and political personalities from the country. Two weeks later, the same venue hosted the second event, the "Tunis Open" international tennis tournament. As the name suggests, this is a professional tennis tournament held on red clay courts in an outdoor setting. Since 2005, it has been contested annually at the Tennis Club de Tunis in Tunis, Tunisia, as part of the ATP Challenger Tour.

Method and Structure of the Two Samples
We used a questionnaire to obtain the data necessary for the empirical study. We chose two samples of reasonable convenience corresponding to each sporting event (Ardilly, 1994). Moreover, considering that the two samples are different, the approach would be to have two homogeneous samples concerning variables that we can control (sex, age, socio-professional category, and interest in sport). This approach does not allow us to attribute any differences between the effects of the two sponsorship operations on these observable variables. We excluded persons who had interaction with the message broadcast on the site from our sampling and instead elected to conduct in-person interviews with fans at both athletic events. All respondents gave their informed consent and participated voluntarily, therefore all ethical standards were met. Our research relies on two samples of varying sizes due to financial, manpower, and time restraints. This is why we only kept 299 respondents: 160 from the Rencontres Peugeot Roland Garros and 139 from the Tunis Open.

Measurement Tool
Brand image is an abstract concept that is evaluated by items and beliefs (Kapferer, 1988;Keller, 1993). Each brand image has its own identity and values. The construction and validation of brand measurement scales are necessary to study image transfer. Thus, the process is verified using structural equation models. The analysis makes it possible to introduce the notion of latent and abstract concepts of the image and to identify possible relations with the observable variables that measure them (Evrard et al., 2003). The nature of the relationships between latent concepts and their measurement variables facilitates understanding of this causal effect (Anderson & Gerbing, 1998;Roussel et al., 2002). The method of analysis chosen was Maximum Likelihood (ML), the most used in marketing, and it is appropriate for exploratory studies (Baumgartner & Homburg, 1996 followed are those of Churchill's (1979) paradigm. As a first step, we contacted the company "Peugeot" in Tunisia and the organizer of the event "Tunis Open." At the end of the interviews, 21 items representing the brand image were announced ("Peugeot-event pair"). This list was then distributed to a sample of 50 randomly selected tennis spectators at a regional tennis tournament, to retain only the common values of the "sponsor-event pair" having a high score (more than 50%, using a five-point Likert scale: from disagree to strongly agree). At the end of this data-collection stage, 11 items were retained for the sponsor/sponsored entity pair. In the second step of the exploratory phase, we performed a principal component analysis to refine the scales and test their reliability. Four measurement scales were subjected to this analysis: the scale measuring the brand image of Peugeot in the case of the creative event Les rencontres Peugeot Rolland Garros; the scale measuring the brand image of Peugeot in the case of the participation event Tunis Open; the scale measuring the brand image of the sporting event Les Rencontres Peugeot Rolland Garros; and the scale measuring the brand image of the sporting event Tunis Open. The results revealed the bidimensionality of these scales and a grouping of the cognitive and affective components (Ganassali & Didellon, 1996;Speed & Thompson, 2000). The reliability tests were good for both axes and each scale. These scales also have good representational quality (Didellon & Valette-Florence, 1995). Following the constitution of these measurement scales, we proceeded to conduct confirmatory analysis by referring to the theoretical indices of Roussel et al. (2002). Although some values were lower than the theoretical norms, probably owing to the small size of the samples, the measurement models were accepted. We then analyzed and compared the image transfer process for each action (CS/PS).  The measurement scale developed by Strazzieri (1994) was utilized in this study so that the moderating influence of involvement in the sporting event on the image transfer process in each sponsored action could be verified. The length of this scale is less than that of other scales. It is a scale with only one dimension, and its three categories are interest, relevance, and attractiveness. In the participants of the study, our objective was to establish its dependability.

Results and Analysis
According to the research's conceptual model, there are six different relationships that need to be confirmed for the tournaments CS (La Rencontre Peugeot Rolland Garros) and PS (Tunis Open). The activity both presented and interpreted the outcomes of the investigation.

Image Transfer Model in the Case of CS
The image transfer model in CS converged, and most of the adjustment indices had values close to the recommended thresholds, which allowed us to accept this model and analyze the results. This exclusive action's links are all statistically significant ( < 05). The model was confirmed for all the cognitive and affective components. Thus, six hypotheses were validated. The sports event "Les Rencontres Peugeot Roland Garros" allowed a total image transfer toward its exclusive sponsor, Peugeot. The interpretation of this structural model is based on the standardized and non-standardized parameter estimates, the multiple squares of correlation, and the regression coefficient. The transfer process was good for all structural relationships, according to these results. We saw a double transfer from the cognitive part of the image of the event to the emotional and cognitive parts of the image of the sponsor. These results also show that there is a positive link between the emotional part of the event and the cognitive and emotional parts of the sponsor. The "dual mediation effect" is what happens because of this double transfer. An affective halo effect is also confirmed, which means that the emotional aspect of the event is transferred to the sponsor. The same is true for moving the emotional part of the sponsor's image to the mental part of the sponsor.

Image Transfer Model in the Case of PS
The global model of image transmission from the sponsored entity to the sponsor in the multi-sponsor event "Tunis Open" has adjustment indices within limitations, allowing us to accept the tested model. The standardized regression coefficients are non-significant except for two. The cognitive-affective relationship was the first major link (H1). The event's cognitive component and "Peugeot" sponsor's cognitive dimension were the second significant transfer (H2). Thus, the remaining estimated structural equation parameters are insignificant. Hypotheses 3−6 were rejected.  The findings of the two sponsorship drives allowed us to determine that six transfers were successful in CS and two were successful in PS. When compared to having multiple sponsors, the exclusive sponsorship activity significantly improved the image transfer. Both the event and the sponsor profited via a single, entirely mental exchange. Only the sporting event benefited from the second vertical transfer from the cognitive to the emotive component. As a result, it appears that in the case of multi-partner sponsorship, the sponsor receives no special treatment in terms of image transfer.

Moderating Effect of Involvement on image Transfer
Our study is intended to be comparative; therefore, the analysis of the moderating effect of spectator involvement in the event on image transfer will only concern the significant paths that have been validated in each action. The respondents must be divided beforehand into homogeneous groups regarding the involvement variable, to verify the variety of the intensity of image transfer from one group to another (Holmbeck, 1997). However, since the size of our two samples was relatively small (160 in CS and 139 in PS), we avoided excessive splitting of the numbers and contented ourselves with a classification into two groups according to the median (Dabholkar & Bagozzi, 2002). The first group represents individuals with low involvement, whereas the second group is composed of individuals with high involvement. This was true for each sponsorship action. However, for a multi-group analysis, we put forward two hypotheses-M0 and M1-to confirm or deny the moderating effect of involvement on the image transfer process in each sponsorship action. According to this approach, the possible differences between image transfers can be attributed to the degree of involvement. Therefore, these two hypotheses are as follows: Model M0: The factor coefficients relating to proven transfers are different between the two groups. This model was referred to as the "free model." Model M1: All factor coefficients were identical between the two groups. This model was referred to as the "constrained model."

Moderating Effect of Involvement on image Transfer in CS
In CS, constrained and free multi-group analyses lead to a significant result. Thus, Hypothesis M0 is verified. The intensity of the transfer in CS varies from one group to another. The indices for evaluating the goodness of fit of the two free and contra models were only acceptable through the parsimony indices. According to Laborde and Boris (2009, p. 7), the model can be retained "all the more so when one considers the small sample size of each subgroup (less than 200 individuals)." The results of this moderating effect were positive and stronger when moving from the low to the high involvement group. We recorded three significant transfer relationships in the low involvement group and six in the high involvement group. The more individuals involved in the creation of the event "Les Rencontres Peugeot Roland Garros", the stronger the image transfer.

Moderating Effect of Involvement on image Transfer in PS
In the case of PS, the multi-group analysis produced significant results (p < .05), allowing Hypothesis M1 to be accepted. This suggests that the intensity of the image transfer varies from one group to another, and the chi-squares between the constrained and free models is different. The goodness of fit of the two free and constrained models is only good for the Akaike information criterion index of parsimony. According to Sauer and Dick (1993), the main reasons for this are the small sample size (139 individuals) and the small number of subgroups formed. Moreover, this co-sponsoring action has reduced this moderating effect. Two transfer relationships are significant for the low -involvement group, but against only one purely cognitive relationship, linking the event to the sponsor. The latter is stronger when moving from the low to the high group (0.549 and 1.363), while from the cognitive component of the event to its affective component, the involvement of the spectators in Tunis Open moderated the strength of the image transfer (1.126 and 0.382). Compared with CS, whose effect is linear and positive, actions involving multiple sponsors seem to moderate the strength of image transfer, moving from the group with low involvement to that with high involvement.

Discussion
Evaluating the actions of sports event sponsorship remains a current problem, which has been enriched by many original theories and data but has not yet been completely clarified (Boronczyk & Breuer, 2020;Morgan et al., 2020;Zhang et al., 2020). These results are consistent with studies that question co-sponsorship (Didellon, 1998;Grohs et al., 2020;Stipp & Schiavone, 1996). Hutchinson and Aba (1991) believe that multi-sponsor actions involve additional stimuli that require added cognitive effort that the exposed individuals cannot always provide. These observations are confirmed by Gwinner (1997) and- Gwinner and Eaton (1999), who, assume that a high number of sponsors associated with the same event may moderate the transfer process. However, Carrillat et al. (2005) found the opposite, noting the absence of weakening and/or moderation of attitude transfers in the case of co-sponsorship. These researchers believe that a decrease in event-sponsorr transfers is offset by transfers from other sponsors. Ruth and Simonin (2003) state that controversial brands can improve their evaluation of PS action by benefiting from the indirect positive effects of other sponsors. Regarding the linear moderating effect of spectator involvement on image transfer found in CS, Pham (1992) mentions that spectators trigger more sympathy toward the sponsor when the sports event favors the entertainment aspect. Petty et al. (1983) and Scheinbaum et al. (2019) noted that consumers with a low level of involvement are more receptive to emotional messages, creating an environment favorable to a passive perception of information (Petty & Cacioppo, 1986;Valette-Florence, 1989). The presence of several sponsors seems to disperse spectators' attention and, therefore, weakens the identification of the sponsoring images. Given their often highly developed competitive nature, these sporting events do not appear to elicit more sympathy and closeness to the sponsor images. Indeed, these spectators are more focused on sports results than on other aspects of the event (Maanda et al., 2020). Nevertheless, other studies have found that the strength of the transfer does not change, regardless of the degree of involvement with the event and sponsored activity.

Conclusion & Recommendations
This study examines the comparative effectiveness of CS and PS in terms of image transfer and the moderating effect of spectators' involvement in the event on this process. The objective was to justify the choice to participate in and/or create a sporting event. This choice becomes a rational one in the face of the professionalization of the sector. The conceptual model constructed grouped the six relationships recommended in the process of image transfer in sports sponsorship together. The scales for measuring the image of each entity were also constructed and validated for CS and PS. The results were analyzed using structural equation models. The results are unanimously favorable for exclusive CS. The impact of "La Rencontre Peugeot Rolland Garros" is much better than that of the international tennis event, the "Tunis Open." All six structural links are significant in the case of CS, compared to only two links in the case of PS (a purely cognitive horizontal link, event/sponsor, and an event-specific "cognitive/affective" vertical link). The intensity of this image transfer process is also perceived as stronger and more linear in CS than in PS. This multi-sponsor action moderated the intensity of the image transfer process, moving from a group with low involvement in the event to one with high involvement.
This comparative study, in terms of image transfer from a sporting event to a sponsor, favors CS. This process is also better perceived by spectators of creative events. This exclusive action, which has yet to be fully exploited by sponsoring companies, could constitute a less expensive and more persuasive opportunity than participation. This assumption needs to be verified for other sports events with different amounts of media coverage. Thus, research focusing on the indirect audience of sports is recommended. It would also be relevant to increase the sample size for a better quality of fit in the structural equation models (Chin et al., 2003). Furthermore, it would be interesting to conduct a longitudinal study to verify the stability of the results obtained.