Corporate image and a sport’s governing body

ABSTRACT This study provides a topical examination of corporate image and a sport’s governing body. Our study is based on 500 worldwide followers of football and has major implications because it serves as a basis for understanding how sports fans might develop and maintain trust in a governing body, and what the attendant antecedents and consequences are. The evidence from our research shows that the key attribute that enables a sport’s governing body to build trust is corporate image; corporate image is different to corporate reputation because it is built on mental imagery rather than value judgements. This finding, we argue, is also applicable to other sports’ governing bodies, particularly when there is a need to repair or re-build trust. The management of corporate image, as a mental image, is particularly significant if a sport’s governing body’s corporate image is called into question because income streams may be undermined.


Introduction
This article focuses on the role of corporate image and trust in the context of a sport's governing body. By its very nature, sport must be seen to be representing a fair contest embodying the true spirit of competition. Reactions to scandals involving doping, match fixing and other such unsavory events are testament to the crucial importance of stakeholders upholding a strong corporate image and trust in the governance and outcomes of sporting encounters (Maening, 2005). Equally, there is much at stake as the global sporting industry generates revenues of many billions of dollars through gate receipts, broadcast rights, merchandize sales, sponsorship and other revenue streams (Milano & Chelladurai, 2011). Should the corporate image be portrayed poorly and trust in a sport be called into question, then it is likely that revenue streams would be threatened or undermined (Glantz, 2010).
Despite the undoubted importance of a strong image and resultant trust and the role of the relevant governing body in creating and maintaining conditions that maximize potential benefits for all stakeholders, including commercial sponsors, our understanding of corporate image and its interplay with constructs such as trust is rudimentary (Tran, Nguyen, Melewar, & Bodoh, 2015). Given the importance of corporate image for a governing body, the precise objectives of the study are to: (a) present and validate a coherent set of antecedents of corporate image for a sport's governing body; and, (b) identify how these antecedents impact on corporate image and, in turn, how corporate image informs trust. Therefore, taking the objectives together we make an important contribution to our understanding of corporate image and trust in a sport's governing body.
The rest of this article is organized as follows. First we critically appraise the relevant literature and present our hypothesized relationships including the rationale for the relationships within our model. We then move on to presenting our methodology; at this stage we also address some domain-related issues. Further, given the nature of our study, gaining data on a global basis is important for overcoming the limitations of a mono-country study which might be unduly influenced by that country's characteristics. Following this, and based on data evaluation using partial least squares (PLS) path modeling, we present our findings. Lastly, we highlight our study's implications and draw attention to its limitations.

Conceptual framework
The conceptual framework that we propose and test (as shown in Figure 1) is founded on the notion that corporate image comprises various perceptions held by stakeholders in an organization (Nguyen & Leblanc, 2001). Corporate image can, therefore, be influenced by impacting on such perceptions and can, in turn, impact on various consequential factors. Such a point is illustrated by the centrality of corporate image in our conceptual framework. We add to prior literature by proposing that corporate image is a second-order construct reflecting the first-order factors of trustworthiness, transparency, reputation and responsibility. Furthermore, we model a number of antecedents of corporate image, while trust (cognitive and affective) and word-of-mouth are positioned as outcomes of corporate image.

Nature of corporate image
First, we consider the nature of corporate image and justify our contention that it is a second-order construct. Nguyen and Leblanc (2001) posit that corporate image underpins the positioning of an organization, while Worcester (2009) contends that corporate image is based on the experiences and knowledge that are gained about an organization. Rooted in an examination of the corporate-image-identity-reputation literature, Barnett, Jermier and Lafferty (2006) suggest that organizations with a favorable corporate image influence satisfaction, which elicits loyalty.
This study conceptualizes corporate image as the overall impression of a sport's governing body, and one that stems from experiences. As corporate image stems from experiences, we specify that as a second-order construct it reflects four key first-order indicators: transparency, trustworthiness, reputation and responsibility. Corporate image is related to transparency, with transparency reducing skepticism (Pomering & Johnson, 2009), while corporate image improves trustworthiness by reducing uncertainty and perceived risk (Gürhan-Canli & Batra, 2004). We predict congruence between corporate image and corporate reputation and it would be myopic not to view the two as strategic tools (Balmer, 2011). The final relationship we estimate is between corporate image and responsibility, which is a relationship centered on taking responsibility for actions (Arendt & Brettel, 2010).

Transparency
The compelling role of transparency is evident in cases of organizational failures, for example Enron and others during the early 2000s, when a perceived lack of transparency played a part, particularly as Kaptein (2008) observes when examining the nature of ethical behavior inside organizations.
In synthesizing the various research streams, 'transparency is the perceived quality of intentionally shared information from a sender' (Schnackenberg & Tomlinson, 2014, p. 5). When there are variances in the quality of transparency, they tend to be a corollary of the quality of the information that is provided (Schnackenberg & Tomlinson, 2014). Transparency is therefore important for an organization, and there is established evidence of an association between corporate image and transparency.

Trustworthiness
Trust is an abstract state of mind with a set of perceptions whereas trustworthiness can be influenced as a result of the actions of others (Sekhon, Ennew, Kharouf, & Devlin, 2014). Mcknight, Choudhury, and Kacmar (2002) refer to trustworthiness as a perception that trust will be retained, thus reducing uncertainty. Unlike trust, which is nebulous, trustworthiness is needed for long-term development of a relationship (Chang, Doll, Wout, Frank, & Sanfey, 2010). In drawing a link with trustworthiness, corporate image plays a part in helping understand what an organization is offering and by doing so it helps to reduce uncertainty about that organization (Lin & Lu, 2010).

Corporate reputation
Corporate reputation is related to, but distinct from, corporate image (Eberl & Schwaiger, 2005). Whereas corporate image is a concept based on a mental imagery, corporate reputation is based on value judgments and attributes and as such evolves over time as a consequence of consistent performance (Gray & Balmar, 1998). Economists have tended to equate corporate reputation with economic value; as an intangible resource this can have a significant influence on performance (Nguyen & Leblanc, 2001). Thus, there are a number of financial and non-financial benefits accrued from a strong corporate reputation (Eberl & Schwaiger, 2005). As such, there is consequent relationship between corporate image and corporate reputation with a strong corporate image influencing corporate reputation (Gotsi & Wilson, 2001).

Responsibility
Xie and Peng (2009) note how organizations may provide a manifold signal to customers through their willingness to assume relevant responsibility, which may lead to a favorable impression being formed among customers. Moreover, stakeholders' perceptions and attitudes about an organization's ethics play an important role in shaping their behavior with regard to responsibility (Balmer & Greyser, 2002). Adopting a meta-analytic approach, Peloza and Shang (2011) show a link between corporate image and responsibility. The preceding discussion means that the sport's governing body is not only responsible towards fans, but towards all stakeholders that have an interest in the game.

Antecedents and consequences of corporate image
Expertise is an important attribute for developing a corporate image because it demonstrates the ability to deliver what is expected (Sirdeshmukh, Singh, & Sabol, 2002). Gürhan-Canli and Batra (2004) support the idea of a link between expertise and corporate image. But, this expertise is not just at the point of engagement; judgments can be made from a distance that influence corporate image, for example online reviews (Vermeulen & Seeger, 2009). For a sport's governing body, expertise is fundamental in terms of how it manages the sport at a global level. It also means the sport's governing body demonstrating to fans that it has the ability to discharge its duties in an efficient and a proficient manner.
Integrity is about being honest (Mayer, Davis, & Schoorman, 1995) so as to ensure an organization can deliver a promised outcome. Elevated levels of integrity reduce the perceived risk associated with interactions. In an organizational setting it is about corporate citizenship. The manner in which integrity is demonstrated will have an impact on corporate image (Jones, Temperley, & Lima, 2009). For a sport's governing body, upholding integrity through clear processes is central to its operations and it must be seen to have high integrity; this is about consistency and honesty. How decisions are reached should be beyond question.
Effective communications engender positive behavior in recipients and effective communications have been linked with a strong positive image (Nandan, 2005). Specifically, three aspects of communication are important: accuracy, explanation and openness (Whitener, Brodt, Korsgaard, & Werner, 1998). Communication is important primarily because it acts to manage the information flow from one to the other. Thus, communication is about the governing body understanding the needs of fans so that they too have knowledge in terms of how decisions are reached and fans' voices are heard. This centers on regular and open communications from the sport's governing body as well as informing fans of new developments. It also means the governing body is open to the views of fans as without clear communication there could be a relationship breakdown.
We posit benevolence as the fourth important driver of corporate image. Benevolent behavior is about not acting in an opportunistic manner (Morgan & Hunt, 1994). As a construct, benevolent behavior has three main dimensions: (a) consideration and sensitivity; (b) protecting the interests of others; and, (c) refraining from exploiting others (Sekhon et al., 2014). The role of benevolent behavior is important where there is a greater degree of relationship intangibility. It is about the concern that is shown towards someone else by taking into consideration their interests (Nguyen & Leclerc, 2011). For us, benevolent behavior means Fédération Internationale de Football Association (FIFA) clearly demonstrating it has the best interests of fans at the core of what it does.
We hypothesize that shared values is the final key antecedent of corporate image with Simöes, Dibb, and Fisk (2005) noting that management plays a key role in promoting shared values and belonging, the emphasis of which should be on the internal controllable aspects of such a process. Culture reflects the management beliefs and/or practices that are internal to the organization but these take time to evolve (Galford & Drapeau, 2003). One of the ways in which behavior is likely to be influenced is by having the same values as another person (Kelman, 1961). In an organizational setting these values are part of the corporate culture that leads to the generation of corporate image (Melewar & Karaosmanoglu, 2006). Given the various debates surrounding the importance of shared values, for our domain they can be considered as being the extent to which a sport's governing body engages communities and has the same values as those communities and the fans.
Taken together, corporate image is central to the way that an organization presents itself, and given that the image can be influenced by a host of factors we therefore propose the following hypotheses: H1: For a sport's governing body, expertise positively influences corporate image. H2: For a sport's governing body, integrity positively influences corporate image. H3: For a sport's governing body, communication positively influences corporate image. H4: For a sport's governing body, benevolent behaviors positively influence corporate image.
H5: For a sport's governing body, shared values positively influence corporate image.

Relationship between corporate image and trust
There is a positive relationship between corporate image and customers' trust, an argument that finds support in Flavián, Guinaliu, and Torres (2005) also draws attention to a possible link between corporate image and trust. While there is some empirical work examining trust and image together, a deeper examination of the relationship between corporate image and trust is absent, particularly in the service sector.
Customers' trust has been treated empirically as comprising cognitive and affective elements (Sekhon et al., 2014). Cognitive trust is about meeting expectations in terms of the trust outcome. In effect, it can be equated with the trustee delivering on explicit or implied promises, and thus it is about rational choice rooted in the knowledge about the trustee (Sekhon, Roy, Pritchard, & Shergill, 2013). Cognitive trust tends to be present when the trustor has sufficient evidence that the trustee can deliver on their promises (Bachmann & Inkpen, 2011).
Unlike cognitive trust, affective trust involves a deeper, more profound connection based primarily on emotional bonds (Chua, Ingram, & Morris, 2008). Given this emotional connection, affective trust is particularly pertinent to situations within which emotional investments are made (Sekhon et al., 2013). Affective trust is built on the care and concern shown by the trustee towards the trustor (Swift & Hwang, 2013). In essence, rather than just delivering on promises, it is about having the trustor's true interests at heart.
We incorporate the important distinction between cognitive and affective trust into our model to emphasize the conceptual boundary between the two; together, cognitive and affective trust allow for accurate and multi-faceted assessments of overall trust (Dirks & Ferrin, 2002).
We conceptualize that for a sport's governing body, cognitive trust relates to it following through on what it says it will do, and being dependable and meeting basic expectations. From an affective trust perspective, it means the governing body demonstrating that it is concerned about sport's fans, as well being honest with them and making every effort to address their needs in addition to having their best interests at heart. Informed by the relationship between cognitive and affective trust we present the following hypotheses: H6: For a sport's governing body, corporate image influences cognitive trust. H7: For a sport's governing body, corporate image influences affective trust.

Relationship between cognitive trust and affective trust
Affective trust is more idealized and deeper than cognitive trust (Sekhon et al., 2013). It is generally considered to be the case that the trustor moves from the former to the latter as the relationship becomes more established meaning that it is process oriented (Jap & Anderson, 2007). Thus, we present the following hypothesis: H8: For a sport's governing body, cognitive trust positively influences affective trust.

Relationship between trust and word-of-mouth behavior
We position word-of-mouth as the final outcome variable because advocacy behavior, in terms of positive word-of-mouth, is considered to be an important consequence of trust (Roy, 2013). Gremler, Gwinner, and Brown (2001) posit that customers who feel trusting and comfortable in their relationships with firms act as advocates for the firm. Podsakoff, MacKenzie, Paine, and Bachrach (2000) argue that positive word-of-mouth can be considered to be a type of customer citizenship behavior. In turn, increasing positive word-of-mouth is considered to be important in affecting future behavioral intentions (Roy, Lassar, & Butaney, 2014), and in positively influencing a firm's performance (Eisingerich, Auh & Merlo, 2014). For a sport's governing body this means that if there is positive word-of-mouth then there is likely to be greater fan engagement that goes beyond participation to also include becoming advocates through promotion and recommending the sport to others. We estimate that positive word-of-mouth would lead to greater involvement from sponsors given the publicity.
We conceptualize word-of-mouth as an outcome of trust, providing a compelling tool for high credibility (Garnefeld, Helm, & Eggert, 2011) and desirable citizenship behavior. This means that the trustor positively spreads the name of the trustee and thus acts as a powerful tool for the marketer (Martin & Lueg, 2013). Moreover, word-of-mouth is of greatest significance where the product is intangible (Harrison-Walker, 2001).
Similarly to Sweeney, Souter, and Mazzarol (2012) our study conceptualizes word-ofmouth to be distinct from loyalty and other associated constructs. Echoing Martin and Lueg (2013), we thus hypothesize that there is a direct link between trust and positive word-of-mouth behavior, with trust acting as an antecedent of positive word-of-mouth. Given this, we present the following hypotheses: H9: Cognitive trust in a sport's governing body will positively influence word-of-mouth. H10: Affective trust in a sport's governing body will positively influence word-of-mouth.

Research domain
In meeting our research objectives we focus our attention on the global governing body of football, FIFA, which we used as the base for our data. Football (or soccer, depending on where you are based) is routinely referred to as the 'global game', attracting huge numbers of spectators at matches across the globe with many times more watching games on television.
It should be noted that recent times have resulted in FIFA having to deal with a range of problems including issues of ethics and governance (Pielke, 2013); alleged and proven cases of bribery (Becker, 2013); cronyism and vote-rigging (de Almeida & Wanderley, 2013); and a failure to appropriately account for the interests of football fans (Cornelissen, 2004). It may be that these perceptions of FIFA are bound up with the perception of individuals representing FIFA, similarly to how perceptions of political parties tend to be related to the party leader in the way of personalization (Garzia, 2011).

Research instrument development
Taking our conceptual framework in its entirety, some of the relationships have been shown in the literature but our approach is novel in developing a single, more comprehensive, study. We adopted a conventional methodology for our research. For our survey instrument we generally used well-regarded and previously tested existing measures. Netemeyer, Bearden, and Sharma (2003) argue that the utilization of existing measures is an acceptable approach, particularly where there is the existence of suitable welltested measures, given the input required to develop new scales. For each construct, except transparency, we utilized/modified a range of items from well-regarded studies.
In the case of the construct transparency, we developed new items because suitable ones were not available to us. Specifically, we developed the new items through a series of discussions between members of the research team and with insight from external independent parties with an interest in our research domain. The new and existing items were then tested as part of the pilot survey (for the sake of brevity, not reported in full here).
Where we utilized existing measures we were mindful that some of the scales were developed in a context-specific manner, but we deemed them to be suitable for the purposes of our research because of the robust methodology undertaken to develop those items and their transferable nature. Following discussion between the research team, in some cases the item stems were slightly modified to fit our domain. Moreover, because we did not rely on one source for our item pool, we were able to maintain the robustness of the entirety of our item pool. The items upon which we drew were originally developed to reflect the conceptualization of the latent constructs in our theoretical framework. Table 1 shows the items that we utilized along with their sources.
Despite common agreement among the research team in identifying the items prior to the main survey taking place, we nevertheless pilot tested our research instrument among a group of postgraduate students at the host institution of the lead researchers. This group was deemed to be suitable because they were undertaking a sport marketing course and were keen followers of sport. The purpose of the pilot was threefold: (a) to make sure we were clearly measuring our constructs; (b) to assess research instrument quality, in terms of how it was being interpreted; and, (c) to ascertain whether there were any issues with respondent fatigue. As part of this pilot, we received 50 completed instruments.
Emergent from the pilot study was that the sample had no issues in terms of the measures. Furthermore, there was no evidence of respondent fatigue or issues with interpreting the survey. Thus, we made no further modifications to the survey instrument.
The items on the survey were anchored using a 5-point scale point where 1 = strongly disagree; 3 = neutral; and, 5 = strongly agree. We did not use anchors for points 2 and 4 on the scale.

Main survey and sample
To collect our data, information regarding the survey and a link to the survey site was sent out via social media, namely Twitter and LinkedIn. This reached a potential ∼8000 respondents. In addition, a CNN journalist also promoted the survey via a Twitter account with ∼15,000 followers. Collectively, we invited those with an interest in football to complete the survey. During the period when the survey was live, Twitter and LinkedIn were used to remind potential participants to complete the survey; this was done on three separate occasions. No incentives were provided to respondents to participate in the survey, which ran during February and March 2014, prior to the 2014 FIFA World Cup. In total, the survey achieved 503 usable responses with the data collected using the Bristol online platform.
The use of a web-based approach was important because of the transnational nature of our work. We were also appreciative of the fact that the use of an online approach was likely to result in the same degree of validity as a paper-based approach, because the variance comes with open-response questions, which our survey did not have (Deutskens, Ruyter, & Wetzels, 2006).
Grouping the responses into FIFA's regions, in terms of the realized sample distribution, 81% of the respondents were from Europe; 8% from Asia; 6% from North America; with the remainder from Oceania and South America. We received no responses from Central America. By gaining data from a number of regions, we were able to overcome the limitations of a single country study (Steenkamp, 2005), although the data is skewed towards one region.  Doney and Cannon (1997) 0.75 FIFA possesses the appropriate expertise to do its job properly (Adapted from Sekhon et al., 2014) 0.79 FIFA is generally competent in what it does (Adapted from Sekhon et al., 2014) 0.84 Integrity As far as I am concerned FIFA keeps its word (Adapted from Sekhon et al., 2014) 0.76 0.86 0.78 0.61 FIFA shows high integrity (Adapted from Morgan & Hunt, 1994) 0.86 FIFA seems to be honest in what it does (Adapted from Sekhon et al., 2014) 0.88 FIFA seems to be consistent in what it does (Adapted from Sekhon et al., 2014) 0.58 Communications FIFA informs fans immediately of new developments (Adapted from Morgan & Hunt, 1994) 0.87 0.86 0.76 0.67 FIFA communicates regularly with fans (Adapted from Sekhon, et al., 2014) 0.81 FIFA communicates openly with fans (Adapted from Sekhon et al., 2014) 0.77 Shared values FIFA has the same concerns as me (Adapted from Siegrist, Cvetkovich, & Roth, 2000) 0.89 0.94 0.90 0.83 FIFA has the same values as me (Adapted from Siegrist et al., 2000) 0.92 FIFA acts as I would (Adapted from Siegrist et al., 2000) 0.92 Benevolence FIFA does whatever it takes to make fans happy (Adapted from Hess, 1995) 0.90 0.93 0.80 0.88 FIFA acts in the best interests of fans (Adapted from Sirdeshmukh et al., 2002) 0.92 FIFA shows respect for the fans (Adapted from Sekhon et al., 2014) 0.87 Trustworthiness FIFA makes every effort to address fans' needs (Adapted from Sekhon et al., 2014) 0.90 0.95 0.94 0.84 FIFA has a reputation for being honest with fans (Adapted from Doney & Cannon, 1997) 0.92 FIFA has a reputation for looking after fans (Adapted from Sekhon et al., 2014) 0.93 FIFA has a reputation for having fans' interests at heart (Adapted from Sekhon et al., 2014) Sekhon et al., 2014) 0.93 FIFA seems to be very reliable (Adapted from Cummings & Bromiley, 1996) 0.92 Affective trust I trust FIFA to be truly concerned about me (Adapted from Sekhon et al., 2014) 0.92 0.96 0.95 0.86 I trust FIFA to be honest and fair with me (Adapted from Sekhon et al., 2014) 0.93 I trust FIFA to have my best interests at heart (Adapted from Sekhon et al., 2014) 0.94 Transparency (new items) FIFA takes time with fans to understand who we are 0.82 0.90 0.88 0.57 FIFA asks fans like me before making decisions 0.78 FIFA provides information that is complete 0.80 FIFA provides fans with information that could be verified by an outside source/auditor 0.82 FIFA provides fans with information in a language that it clear 0.58 FIFA provides fans with more than one side of a controversial issue 0.82 FIFA provides fans with information that can be compared with previous years 0.65 Word-of-mouth I say positive things about FIFA to other people (fans) (Adapted from Roy, 2013) 0.90 0.96 0.95 0.88 I often recommend FIFA to others (fans) (Adapted from Fullerton, 2003) 0.94 I encourage friends and relatives to become a fan of FIFA (Adapted from Roy, 2013) 0.94 I will continue to recommend FIFA to others (fans) (Adapted from Fullerton, 2003) 0

Findings
Initial analysis indicated that responses represented non-normal data distribution which we put down to the emotive nature of the research domain. Given that we used FIFA for our domain, the responses are a reflection of how the respondents view the governing body of football to be, which may be the result of the negative publicity that it has received in recent years. There is also anecdotal evidence that followers of football are more passionate and positive about the team they 'support' than they are about their national team.

Data analyses and results
To test our proposed hypotheses, given our data's non-normal distribution, we applied partial least squares structural equation modeling (PLS-SEM) using the statistical tool SmartPLS 3.2.3. We selected PLS path modeling because of its many advantages. First, PLS's flexibility in distributional assumptions and strength in handling complex causal models and constructs with a fewer number of items (Hair, Hult, Ringle, & Sarstedt, 2013). Second, PLS's algorithm does not assume equal weights for all indicators of a scale; it allows each indicator to vary in how much it contributes to the composite score of the latent variable. Third, PLS path modeling represents an alternative methodology for theory testing (Henseler, Ringle, & Sinkovics, 2009). We followed the guidelines of Hair, Sarstedt, Ringle, and Mena (2012) and Löhndorf and Diamantopoulos (2014) to analyze and interpret the proposed path model using PLS path modeling. We present the reliability and validity of the measurement model followed by the results of hypotheses testing.
We used PLS analysis with a repeated indicator approach to assess the measurement model of corporate image as a higher-order construct (Tenenhaus, Vinzi, Chatelin, & Lauro, 2005;Wetzels, Odekerken-Schröder, & Van Oppen, 2009). In this study we conceptualize corporate image as a Type I construct consisting of reflective second-order and reflective first-order constructs (Jarvis, MacKenzie, & Podsakoff, 2003). Empirical support for such a conceptualization can be obtained (i) when the indicator weights for each factor are significant; (ii) when the construct reliability and average variance extracted (AVE) exceed the cut-off values of 0.7 and 0.5, respectively; (iii) when the R 2 of each dimension is greater than 0.5; and, (iv) when correlations between the first-order factors are less than the second-order factor loadings (FLs).

Measurement properties
The construct reliabilities were evaluated using two criteria, i.e. Cronbach's alpha and composite reliabilities (CR). The Cronbach alpha for each of the constructs is greater than 0.75 and all the CR values are greater than 0.85 that are above the recommended cut-off values (Henseler et al., 2009) to indicate appropriate internal consistency. All the measurement items loaded significantly onto their intended construct with a FL greater than 0.50 as shown in Table 1.
We assessed the convergent validity by examining the AVE. The AVE values for all the constructs exceeded the threshold value of 0.50 which supports convergent validity (Fornell & Larcker, 1981). Finally, the discriminant validity was assessed by comparing the square root of the AVE value of each construct with its correlations with other constructs (Fornell & Larcker, 1981). Since in each of the cases the square root of the AVEs is greater than the inter-construct correlations it supports the discriminant validity of the constructs. The discriminant validity of the constructs is shown in Table 2.
We tested the significance of the outer-measurement model using bootstrapping tvalues computed on the basis of 504 cases and 1500 samples (Hair et al., 2012). We obtained the bootstrap critical ratios greater than 1.96 to indicate significance of the reflective outer-measurement model. In addition, the path coefficients (β) from each of the first-order dimensions to the higher-order corporate image dimension indicate strong FLs, ranging from 0.92 to 0.95. The CR and AVE of the dimensions exceeded the threshold levels of 0.70 and 0.50, respectively. The corporate image dimensions exhibited high R 2 -values, exceeding the threshold levels of 0.50.
As shown in Table 3, the correlations between the first-order dimensions were found to be less than those of the second-order FLs. These findings lend support to our proposed dimensionality of the corporate image construct. In addition, each of the first-order factors had high and significant loadings onto the second-order corporate image (see Table 4). Thus, we conclude that the operationalization of corporate image resulted in a reliable higher-order construct, with four reflective first-order factors of trustworthiness, transparency, reputation and responsibility.
We used Harman's one-factor test to mitigate the risk of common method bias in our sample. This was conducted by entering all the measurement variables in an exploratory factor analysis using SPSS. The sample would have a common method bias problem if a single construct explains more than 50% of the extracted variance (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Our results show that the factor obtained only reached 25.5% of the extracted variance for the sample. We also tested for the presence of common method bias using another method proposed by Liang, Saraf, Hu, and Xue (2007). We used a common method factor in the PLS model with indicators of all the constructs. We calculated each indicator's variances explained by the original constructs and the method factor. The results showed that the average of the variance explained by the constructs was 0.75 while the average method-based variance was 0.03 and the ratio of the two was 25:1. In addition, most of the method FLs were not significant. These results indicate that common method bias was not a major threat to our analyses.

Results of hypotheses testing
After demonstrating the validity and reliability of the measurement model, in the second stage we examined whether the structural model supports our proposed relationships. We are informed by the recommendations of Hair et al. (2012) and Henseler et al. (2009) in our analysis of the PLS path modeling results. The evaluation of the prediction-oriented PLS path modeling results for the structural model centers on the R 2 values with the R 2 value representing a measure of the variance explained in a variable by a model. Results indicate that the hypothesized model accounted for 88% of corporate image, 66% of cognitive trust, 80% of affective trust and 68% of positive word-of-mouth, suggesting good overall predictive power. All the R 2 values are greater than the recommended cut-off value of 0.10 (Falk & Miller, 1992). Average variance accounted for is defined as the mean R 2 of the model, which is 0.74. Since the average variance accounted for value is high, acceptable (Falk & Miller, 1992) and the individual R 2 estimates of the endogenous constructs are greater than the recommended cut-off values, it is appropriate to examine the significance of the relationships between the exogenous and the endogenous constructs.
We assessed the relative importance of the proposed relationships in our model by using the standardized path coefficients (β). Results are shown in Table 5. We applied a non-parametric bootstrapping method using 1500 subsamples with 504 cases and individual sign changes (Henseler et al., 2009). In accordance with hypotheses H1-H5 not all the antecedents of corporate image have a significant and positive impact on corporate image: integrity (β = 0.26; t = 6.68), communication (β = 0.20; t = 5.79), shared values (β = 0.19; t = 5.24) and benevolence (β = 0.39; t = 8.89) accepting H2-H5. Impact of expertise on corporate image was not significant and hence H1 was rejected.
Corporate image has significant and positive effects on cognitive trust (β = 0.81; t = 31.08) and affective trust (β = 0.78; t = 14.97) as shown in Table 6 which supports hypotheses H6 and H7. Finally, there is a significant effect of cognitive trust on affective trust (β = 0.14; t = 2.35); cognitive trust on word-of-mouth (β = 0.28; t = 5.18) and affective trust on word-of-mouth (β = 0.58; t = 10.44), which lends support to H8, H9 and H10.  Moreover, we used the blindfolding procedure to evaluate the predictive ability of our conceptual framework (Tenenhaus et al., 2005). The means of the cross-validated-communality and redundancy (i.e. Stone-Geisser Q 2 ) were used to measure the global quality of the measurement model and the structural model (Tenenhaus et al., 2005).
The results indicate that the quality of both the measurement model and structural model are well above the recommended standard of 0.3 (Tenenhaus et al., 2005). The average cross-validated-communality is 0.67 and the average cross-validated-redundancy (Q 2 ) is 0.64. This further supports the predictive power of our model.

Post hoc mediation analysis
Corporate image is identified in the literature as one of the most influential factors affecting consumers' trust (Flavián et al., 2005). Based on this we posit a direct relationship between corporate image and two forms of trust (cognitive and affective). Likewise, literature also underscores a direct relationship between trust and word-of-mouth (Barreda, Bilgihan, & Kageyama, 2015;Urban, 2005). In light of these findings we posit that cognitive and affective trust mediates the relationship between corporate image and word-of-mouth.
In the post hoc analysis we examined the mediation effects of cognitive and affective trust in the corporate image-word-of-mouth linkage. We adopted the multiple mediator analysis with bootstrapping as recommended by Preacher and Hayes (2008). We used the latent variable scores obtained during the PLS analysis and conducted the mediation analysis using the SPSS macros developed by Preacher and Hayes (2008) to obtain the 95% bias-corrected bootstrap confidence intervals for the mediator variable (Shrout & Bolger, 2002). The indirect effect is significant when the interval for the mediating effect contains no zeros. This study uses 5000 resamples in accordance with the suggestions of Hayes (2009). Table 6 presents the results of the post hoc tests of mediation. Results show  that cognitive trust (β indirect effect = 0.18, p < .05) and affective trust (β indirect effect = 0.34, p < .05) mediate the relationship between corporate image and word-of-mouth.
In summary, the results show that corporate image is a multi-dimensional construct which impacts fans' cognitive trust and affective trust as well as word-of-mouth.

Conclusions and managerial implications
Given the highly emotive nature of sport, we set out to evaluate the nature of the relationship between corporate image, trust and other constructs using a sport's governing body. As posited earlier, governing bodies fulfill an important role because they serve as the custodians of the sports for which they are responsible (Misener & Doherty, 2013). This role is consequently bound up in key principles of good governance such as integrity, responsibility and communication. FIFA and a number of sport governing bodies have come under increasing scrutiny and attracted growing criticism in relation to aspects of their governance (Jennings, 2011), leading to a poor corporate image and, anecdotally at least, trust levels being undermined.
As such, we believe there are two main actionable points from our study: firstly, the research serves to highlight the specific role of corporate image as a basis upon which fans may trust a sport governing body; and, secondly, it signals key variables that sport governing bodies should address in managing their corporate image and stakeholder engagement (primarily fans). Among these implications, similar to Goodwin and Grix (2011), for FIFA the creation and maintenance of strong image management allied to trust is a process that needs to be actively addressed. Moreover, we contend that in strategically managing matters pertaining to image and trust, FIFA, as a governing body, must engage with fans in an effective, meaningful way. Furthermore, we anticipate extensions of this study to account for other stakeholders (for example, sponsors) will reiterate the benefits that can be achieved through a positive corporate image.
Our findings have revealed that four constructsintegrity, communication, shared values and benevolenceare directly linked to corporate image. In its broadest sense, integrity is a key foundation upon which good organizational governance is built (Evans, 2012). This, therefore, indicates that FIFA needs to take the issue of maintaining good governance seriously (Geeraert, Alm, & Michael Groll, 2014) which potentially implies the need for cultural change, strategic management and more effective fan engagement. This may come about if the senior leadership is new and untainted by recent allegations of wrongdoing. Strong governance standards are closely aligned to open and regular communication, another feature of the research findings presented here. For FIFA, this presents a particular challenge given the global nature of the governing body's constituency. One of the ways for FIFA to conceive of the need for communication improvements might be through the use of social media and with greater transparency through the various channels.
In identifying shared values as influencing corporate image, we infer that FIFA should demonstrate a set of values that are somewhat congruent with the values of fans. At one level, this may indicate that FIFA needs to reposition the organization, undertake cultural transformation or communicate in a more strategic fashion. In a similar way, FIFA should also seek to identify means through which benevolent behaviors can be demonstrated, such as showing it has the fans' best interests at heart. While there may be links to the growing commercialization of sport that might explain a perceived lack of benevolence among fans, the governing body still needs to accentuate its role as being 'for the good of the game'.
Congruent with other researchers (e.g. Norman, Avolio, & Luthans, 2010) we find that corporate image is linked to both cognitive and affective trust, thus suggesting that FIFA needs to foster relationships whereby fans become more confident about the way in which the governing body performs its role. We estimate that it is also about the way in which FIFA is perceived to be accomplished and reliable in fulfilling its role as a custodian. Furthermore, perceptions of care and concern among fans need to be enhanced by FIFA, confirming the affective dimensions of trust in relations between fans. Some examples of how this might be achieved may include fan representation on FIFA's boards and committees because outside independent representation on a board can act as form of capital (Johnson, Schnatterly, & Hill, 2013) while diversity can improve the organization's performance (Mahadeo, Soobaroyen, & Hanuman, 2012). The creation of more effective communication channels between the governing body and fans may help to improve trust levels, or the clearer accentuation of missions and strategy by FIFA.
Adding to the findings of Lacey and Kennett-Hensel (2010) we identify a link between corporate image and affective trust. This could be premised on the care and concern shown on FIFA's part; alternatively, fans may perceive feelings of insecurity or weak relations. In summary, the consequences of this are likely to be that fans do not trust FIFA to be responsible for the sport, or in the way it approaches relations with them. Among the implications for FIFA, therefore, are: the promotion of key governance principles such as democracy and openness; and, the inclusion of fans and other appropriate stakeholders into decision making mechanisms employed by FIFA.
Our study supports the idea of cognitive trust leading to affective trust. This happens because FIFA is able to show it has the ability to deliver before trust becomes more idealized in the way of affective trust. Finally, irrespective of whether it is cognitive or affective trust it leads to positive word-of-mouth with this form of advocacy behavior being a capital asset for FIFA. It may, because of this positive word-of-mouth, lead to enhanced revenue opportunities via improved sponsorship deals. Indeed, this is important given that some of FIFA's principal sponsors are calling for independent representation on FIFA's board as a way to safeguard their sponsorship returns.

Research limitations and future research directions
As we have seen, this article complements the body of literature but we recognize the limitations of our study. Given the most recent allegations of corruption and misbehavior among FIFA's executive, if our study is replicated we would expect differences to emerge. Using FIFA we assessed our model at the broad level, but different indicators may emerge if the same survey instrument is applied at the narrow scope level, i.e. the national footballing bodies.
Given the emotive nature of our assessment domain it would hardly be surprising if the model outcomes are different for other sporting contexts. The sample we involved in our study was both highly involved in sport and interested, and, thus, it may have influenced their responses with the potential to create bias. In expanding our work, more recently the IAAF has faced questions regarding integrity and thus research in that sporting arena is likely to result in different findings. In our model we positioned word-of-mouth as an endogenous outcome variable but in other contexts it may behave differently. We would suggest that other behavioral outcomes such as loyalty are included in future models. We examined our model in its entirety and thus we would encourage future researchers to build models with moderating variables, as indicated by Palmatier, Dant, Grewal, and Evans (2006). Lastly, similarly to numerous other studies, we measured our model on a cross-sectional basis and a longitudinal panel study may identify other patterns.

Disclosure statement
No potential conflict of interest was reported by the authors.