• No results found

Quality signals and the media reputation of corporate art collections

N/A
N/A
Protected

Academic year: 2021

Share "Quality signals and the media reputation of corporate art collections"

Copied!
53
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Quality Signals and the Media Reputation of

Corporate Art Collections

Lisa Klop 10363076 23-06-2017

MSc. in Business Administration: Creative Industries University of Amsterdam

(2)

This document is written by Student Lisa Klop who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

Firms can have several and different motivations for collecting art. Next to financial and intrinsic motivations, corporate art can be used to influence the behaviour of employees, to symbolize the corporate image and for marketing communication purposes. The latter two are ways for firms to enhance their corporate image and reputation. The media can play a key role in affecting and building a firm’s reputation. They shape information and reflect on corporate activities.

Acknowledging the power of the media might result in more external corporate communication activities. Therefore, this research investigates which quality signals that a firm conveys are most effective in influencing the media reputation of corporate collections. The media reputation is defined by review volume, valence, credibility and count of firms in one review. Previous research studied the art collection in relation to the corporate identity management framework. This study takes a step further by investigating the influence of corporate communication and symbolism on the reputation of the firm’s art collection. The empirical setting is the VBCN, the Netherlands Association of Corporate Art Collections, founded in 2005. Based on a longitudinal database an overview of variables is made per year between 2005 and 2016. It is found that published annual reports about the corporate collections have a positive effect on the review volume. No significant results were found for review valence, but generally the positive coverage is more prevailing than the negative coverage. Moreover, it is found that press releases have a negative effect on the review credibility. Overall, it is concluded that the newspaper coverage on corporate collections is hard to influence, which can be explained by the autonomy of Dutch journalistic media.

(4)

Table of content

1. Introduction p. 7

2. Literature review p. 9

2.1 Corporate art collections p. 9

2.2 Corporate identity p. 10

2.3 Media reputation p. 11

2.4 Signaling theory p. 13

2.5 Quality signals: Exhibitions and awards p. 15

2.5.1 Exhibitions p. 15

2.5.2 Awards p. 17

2.6 Quality signals: Annual reports and press releases p. 18

2.6.1 Annual reports p. 19 2.6.2 Press releases p. 20 3. Methodology p. 22 3.1 Research design p. 22 3.2 Data collection p. 22 3.3 Dependent variables p. 24 3.4 Independent variables p. 26 3.5 Control variables p. 27 3.6 Method p. 28 4. Results p. 29 4.1 Descriptive statistics p. 29 4.2 Correlations p. 30 4.3 Regression analyses p. 31 4.4 Robustness checks p. 34 5. Discussion p. 35 5.1 Theoretical foundations and hypotheses p. 35 5.2 Limitations, future research and contributions p. 37 5.3 Practical implications p. 38 6. Conclusion p. 39 References p. 40 Appendix p. 47

(5)

1. Introduction

Corporate art started when corporate headquarters incorporated the visual arts within the workplace (Matorella, 1990, p. xii). Before the 1940s the existing collections mainly represented the personal interests of the firm’s chief executive and artists where restricted to dealing with advertising themes (Matorella, 1990, p. 22). In the cultural boom of the 1960s new artistic styles were introduces which encouraged corporate support (Matorella, 1990, p. 22). Large corporate art exhibition made major players in the field more conscious of each other’s endeavours, and more aware of the prestige to be derived from collecting (Matorella, 1990, p. 22). Although Martorella (1990) discusses the trends in the specific context of America, it mirrors the overall developments seen in corporate collecting.

History shows that businesses have always used the arts, but it was never so publicized as it is today (Matorella, 1990, p. 21). An increasing number of art

collections are not restricted to the workplace anymore, but lend artworks to museums and make exhibition spaces available (Matorella, 1990, p. 20). Many firms are armed with their own curator and art department and tend to be cultural sophisticated and professionalized (Wu, 2003, p. 2). This professionalization is also visible in the Netherlands, where forty-nine Dutch cross-industrial businesses are members of the VBCN, the Netherlands Association of Corporate Art collections, founded in 2005 (VBCN, 2017). This research takes the perspective of the firm by elaborating on the motives of firms to have a corporate collection and by relating those motives to the quality signals they convey, as introduces by Spence (1973). The motivations of the firms determine the more internal or external communication about the corporate collection. Kottasz, Bennett, Savani, Mousley and Ali-Choudhury (2007) elaborate on three dimensions of the corporate identity management framework in which firms can utilize their corporate collection. These dimensions are: the behaviour,

communication and symbolism of the firm via its corporate collection (Kottasz et al., 2007). This study is mainly interested in the last two, most external, dimensions of the corporate identity management framework (Kottasz et al., 2007, p. 21). In the

communication and symbolic dimensions firms utilize their collection to

communicate values or to create a specific image in the mind of external stakeholders. Therefore it is expected that they convey quality signals, like public exhibitions and awards and communicate their activities externally via annual reports and press

(6)

releases. Annual reports and press releases are communication tools explained by one of the signaling strategies called impression management (Carter, 2006). Quality signals, conveyed by the firm, can influence the way third party intermediaries, like the media, write about the corporate art collection (Barnett and Pollock, 2012, p. 24). Subsequently, the research question of this paper is: How do quality signals,

conveyed directly by the firm, influence the media reputation of the corporate art collections in the Netherlands (from 2005 till 2016)? Examples of quality signals are awards, exhibitions, annual reports and press releases. The time perspective on reputation regards the evaluation of a firm in respect to its past, present (and future) (Helm et al., 2011, p. 7). Moreover, communication on firm’s corporate art collection may have effects that are only manifested over the long term (Helm et al., 2011, p. 7). It is thus more relevant to research the reputation of the collections overtime, instead of one point in time. This quantitative research will therefore cover a time period of twelve year, since the establishment of the VBCN in 2005, in a longitudinal dataset.

Existing literature on corporate art collections explains the origins (Wu, 2003; Matorella, 1990), and emphases the role of corporate art collections within the

corporate identity management framework, highlighting the motivations of businesses to invest in corporate art collections (Kottasz et al., 2007; Kottasz et al., 2008;

Hoeken en Ruikes, 2005; Wu, 2003; Shane, 1996; Jacobson, 1996). This study takes a step further by relating firm’s art collections to specific quality signals and by

investigating the influence of these signals on the media reputation. In other words, existing research examines the internal motivations and perceptions on art collections (corporate identity), while an analysis on external communication activities and external perceptions on the art collections has not yet been conducted (corporate communication and corporate reputation). A practical implication of this research is for firms to know which quality signals are most effective in influencing the media coverage, also in terms of the strength and the direction of the effect.

The outline of the paper is as follows: first, the existing literature on corporate art collections is reviewed, followed by its relation to the corporate identity

management framework. Second, media reputation construct is briefly introduced. Third, based on the signaling theory and related literature the hypotheses are

formulated. Next, the empirical setting is described together with the data collection and measurement. Subsequently, the method section exhibits the results from testing the hypotheses using the dataset obtained from the conducted database analysis. The

(7)

results from this quantitative study are then discussed and put into perspective. Finally, the last section answers the research question and concludes.

2. Literature review

2.1 Corporate art collections

Consider this short quote from Wu (2003):

‘By the late 1980s law firms, which had no obvious relationship to contemporary art, had an entire department dedicated to the management of their art-related business’.

The quotation illustrates three interrelated aspect of contemporary corporate art collections. Firstly, the quotations indicated that collecting art is not the core business activity of the commercial firms. This raises the question why the law firms would collect art. Argued from a neoclassical model of the firm, the main purpose of a for-profit business is to increase returns (Kirchberg, 2003, p. 146). To increase returns the firm has to improve sales, create a positive corporate image and build a strong

reputation (Kirchberg, 2003, p. 146). Corporate art creates a pleasant surrounding for employees and potentially enhances their bond to the company, thus increasing the effectiveness of the company in yielding profit (Kirchberg, 2003, p. 146). In addition, the ability to buy expensive artworks creates the image for competitors and

stakeholders that this firm is in good shape (Kirchberg, 2003, p. 146). Moreover, firms can utilize the art as a kind of niche marketing to enter the mind of the sophisticated social groups by identifying with their taste (Wu, 2003, p. 9).

Secondly, elaborating on the previous argument, although collecting art is not the core business for these firms, it can closely integrated into the business practices by integrating it into a corporate cultural responsibility program (Maon and

Lindgreen, 2014). Following Kottasz et al. (2007) publicity on corporate art collections fits the corporate responsibility agenda and has the potential to counterbalance negative media coverage (p. 27). Following Maon and Lindgreen (2014) corporate cultural responsibility includes the cultural commitment of firms, such as the build-up of art collections and the sponsoring of art and culture (p. 760). Corporate cultural responsibility is partly in line with the corporate social

responsibility conceptualization (Maon and Lindgreen, 2014, p. 761). It includes actions of the firm that go beyond what is required by law and are developed to

(8)

maintain or contribute to cultural and societal welfare (Maon and Lindgreen, 2014, 761). Like CSR, these actions are integrated into the business practices to meet stakeholder’s expectations, enhance the firm’s reputation and to preserve a competitive advantage (Aguinis, 2011, p. 855).

Lastly, the law firms dedicate an entire department to the art-related business. This indicates the professionalization of the corporate art collections. Behnke (2007) discusses the significant changes between pre- and post-1990 corporate art

collections. The most significant change is the shift towards more art field based corporate collections, which implies the replacement of the corporate trained

administrators by the art-educated curators (Behnke, 2007, p. 225). Art administrators are in-house employees hired to maintain the collection and utilize it for employee or community related activities (Matorella, 1990, p. 157). The presence of professional curators rather serves to increase the reputation of the firm by applying inter-art criteria to their collection (Behnke, 2007, 234). The ability of firms to attract

prestigious curators insures the future value of the collection, visibility to the general public, and credibility within the artistic community (Matorella, 1990, p. 146). These curators buy work from young and emerging artists and tend to be more interested in innovative works like installations, sculptures, photography and prints (Behnke, 2007, p. 231). They also organize public exhibitions in central office halls or separate exhibition spaces outside the office (Behnke, 2007, p. 229). By exhibiting the

collection in separate spaces the corporates position themselves as relevant players in the art field and are able to generate media attention (Behnke, 2007, p. 230).

In short, the professionalized corporate collections with an art-educated curator tend to communicate their collection more externally. External

communication has become increasingly important due to stakeholder pressures. To improve the corporate reputation some firms integrate the art collection and related activities into a cultural responsibility program (Maon and Lindgreen, 2014, p. 760).

2.2 Corporate Identity

Cornelissen, Haslam and Balmer (2007) distinguish between three main domains on identity work: social identity (group level), organizational identity (organizational level), and corporate identity (institutional and customer level) (p. 2). This study is concerned with the corporate identity level, which is most external and symbolic (Cornelissen et al., 2007, p. 3). Corporate identity is the distinctive public image that a

(9)

firm communicates (Cornelissen et al., 2007, p. 3). The identity framework is based on three dimensions: the behaviour, communication and symbolism of the firm via the corporate collection (Kottasz et al., 2007, p. 21). In the first dimension, the corporate collection influences the behaviour of the staff and clients by creating a pleasant surrounding (Kottasz et al., 2007, p. 21). Particularly firms in the service industry are conscious of their image, because clients are in direct contact with the firm (Wu, 2003). Although some firms solely use their collection along this dimension, they can still be active in purchasing new works, educating employees about the collection,

organizing internal exhibitions and lending artworks to other institutions and firms1.

Secondly, the corporate collection can also be a way to communicate specific organizational values, like the inclusion of art collections into a corporate

responsibility program (Kottasz et al., 2007, p. 22). Lastly, corporate symbolism can be expressed through a corporate art collection (Kottasz et al., 2007, p. 22). In this dimension the art collection is utilized as a tool for niche marketing, by symbolizing cultural distinction wherein the firm is knowledgeable and sophisticated (Kottasz et al., 2007, p. 22).

The behaviour, communication and symbolic use of the corporate collection differ per firm. The VU Medical Centre, for instance, solely engages in the

behavioural dimension by creating a pleasant surrounding for their patients and staff

(VUmc)2, while AkzoNobel also utilizes the collection to communicate the

importance of art and culture in society (AkzoNobel)3. AkzoNobel has a separate

exhibition space, with frequent public offering. The firm’s corporate collection therefore becomes more visible for the public. Based on the activities and

communication of AkzoNobel a positive media reputation seems more relevant to them compared to the internal focus of the VU Medical Centre.

2.3 Media reputation

Corporate identity and reputation are distinctive but related concepts. They are related because successful corporate identity management results in an enhanced corporate image and, over time, an improved corporate reputation (Kottazs et al., 2008). The concepts are different because corporate identity addresses the perceptions of the internal stakeholders, while corporate reputation involves other people’s perceptions

1 Retrieved from: https://www.aegon.com/en/Home/About/Supporting-Society/Aegon-Art-Collection/ 2 Retrieved from https://www.vumc.nl/afdelingen/kunst/

(10)

of the organization, which is the result of information transmitted via the media and through interpersonal communication (Hooghiemstra, 2000, p. 58). The Social-constructionists view of the reputational theory is particularly relevant to this

research. The Social-constructionists argue that stakeholders interact with each other and exchange information (Barnett and Pollock, 2012, p. 20). Stakeholders of a firm are the opinion leader audiences such as press, investors and NGOs, and the general public such as consumers and local communities (Dawkins, 2004). These stakeholder groups collectively and socially construct a firm’s reputation, through attention and evaluation (Barnett and Pollock, 2012, p. 16). Following Deephouse, media coverage is a reasonable indicator of the public’s knowledge and opinions about firms

(Deephouse, 2000, p. 1096). Experts may be viewed as gatekeepers, as intermediate consumers, or as mediators of audience response (Shrum, 1991, p. 351). Thereofore, we further elaborate on the media reputation, as introduced by Deephouse (2000).

Media reputation is the attention and evaluation of the corporate collections by the media (Deephouse, 2000, p. 1096). To determine the variables of the media

reputation concept the multidimensional construct media salience is used. Kiousis (2004) explains media salience from the agenda-setting theory. This theory uses the most external definition of salience, in which an object is salient relative to the situation of other objects (Kiousis, 2004, p. 73). The concept consists of three elements: attention, valence and prominence (Kiousis, 2004, p. 71). The level of attention can be measures by the number of reviews (Barnett and Pollock, 2012, p. 21). The evaluation is the attitude towards specific actions and attributes of the firm, or in other words the media valence (Barnett and Pollock, 2012, p. 22). Although journalists use objective methods for news reporting, they frame issues in positive or negative terms through selection and interpretation of events (Pollock and Rindova, 2003, p. 632). They thus provide public expressions of approval or disapproval

(Pollock and Rindova, 2003, p. 632). The third dimension Kiousis mentiones is media prominence. Media prominence is the placement of the story, its size, the pictures or other visual instruments used (Kiousis, 2004, p. 74).

Furthermore, the attribution theory adds a fourth dimension to the salience construct, the credibility of the review. The theory suggests that actors attempt to assess for themselves whether the source of the cue is credible or not (Gemser et al., 2008, p. 27). Cues from low-credibility sources will be given less weight by these actors than those from high-credibility sources (Eagle and Chaiken, 1975; Mizerski et

(11)

al., 1979) (Gemser et al., 2008, p. 27). The attribution theory emphasised that credibility of the newspaper source determines how much people rely on the review (Gemser et al., 2008, p. 27).

The media reputation construct thus includes several levels. This study will examine which quality signals have an effect on the different aspects of media reputation. The methodology will further elaborate on the operationalization of the dependent variable media reputation.

2.4 Signaling theory

This research focuses on first party signals, as introduced by Spence (1973). Spence utilized the labour market to model the signaling function of education (1973, p. 358). The company is not familiar with the qualities of the job applicant. The determination is therefore based on the assessment of the observable characteristics, like previous work experience and education (Spence, 1973, p. 356). Based on these signals the company can interpret whether that person is right for the job (Spence, 1973, p. 356). Signals in his example are thus the observable characteristics attached to the

individual that can be altered by the subject (Spence, 1973, p. 357). In applying the signaling theory to the market, quality signals are most useful for products whose quality are hard to determine and socially constructed, like artworks (Kirmani and Rao, 2000, p. 72). Beckert and Rössel (2013) argue that the value of an artwork or artist originates in an intersubjective process of evaluation by experts in the art field, such as gallery owners, curators, critics, art dealers, journalists, and collectors (2013, p. 180). These experts help establish the artistic reputation for the artist. The quality signals emerging from the art field enable buyers to assess the value of artworks (Beckert and Rössel, 2013, p. 180).

Firms striving to professionalize their collection appoint an art-educated and experienced consultants or curators to purchase the artworks in different quantities of museum-quality, regional or mass-produced art (Matorella, 1990, p. 160). Expertise of these curators and independent consultant allows them to spot young and

promising artists without a yet established reputation (Behnke, 2007, p. 232). Firms therefore play a key role in creating a young artist’s reputation (Behnke, 2007, p. 232). They can, for instance, support an artist’s style or medium that at first was rejected by museums (Matorella, 1990, p. 174). The substantial corporate collections

(12)

and the growing number of corporate sponsorships indicate that firms have a significant contribution to the growth of the arts (Matorella, 1990, p. 173).

Following Matorella (1990) curators and art consultants in their turn use dealers as their primary sources (p. 165). Given the dealer’s credibility, expertise, and gatekeeping function, curators feel they can choose from the widest selection of quality art by going through dealers (Matorella, 1990, p. 165). Reputable dealers and galleries are respected for their ability to screen and exhibit “quality” art, and

particular artists can only be purchased from them (Matorella, 1990, p. 165). The opposite is also true, prestigious dealers are in a position to be selective about clients, and prefer not to deal with the average corporate client (Matorella, 1990, p. 165). This attitude is expressed in reputable dealer’s commitment to maintain very high

standards and insure the future economic success of the artist it represents, which includes placing the work in important private collections (Matorella, 1990, p. 166).

In short, the art market is mainly built around the quality signals actors convey and the reputation they build (which is also a quality signal). A reinforcing feedback loop becomes visible. If the corporate has an established reputation of giving young artists recognition through exhibitions and media publicity, the dealer is more interested in selling to the corporate for the sake of the artist’s career (Matorella, 1990, p. 169). By supporting the careers of these artists the value of the collection increases, assuring the financial investments for the firm while simultaneously adding to the corporate responsibility agenda.

In an effort to generate more and positive publicity for the corporate collection the firm intentionally signals positive information to convey positive organizational attributes (Connelly et al., 2011, p. 53). Corporate art collections are not directly consumer directed, so no mass media advertisement is involved, but a variety of alternative communication devices allow firms to attract media attention. Examples are, among others, exhibitions, awards, press releases and annual reports. Each signal contains information that can influence how the media reports on firm behaviour and activities (Carter, 2006, p. 1150). The media in turn influences other external parties, which may ultimately result in a change in the firm’s reputation among various stakeholder groups (Carter, 2006, p. 1150).

By relating the signaling theory to the identity construct it becomes visible how the motivations for having a collection determine the signals conveyed. Like mentioned before, the VU Medical Centre solely focuses on the enhancement of the

(13)

work environment for the benefit of the patients and the staff (Kottasz et al., 2007). The VUmc therefore is reticent in conveying quality signals and will probably see media coverage on the corporate collection (if any) as a nice side-benefit of their internal activities.

2.5 Quality signals: Exhibitions and awards

Exhibitions and award both illustrations of the communication and symbolic

dimensions from the identity management framework (Cornelissen et al., 2007). The signals are forms of public communication that invoke a sense sophistication and professionalization of the firm among the public (Kottasz et al., 2007).

Professionalized corporate collections with an art-educated curator tend to be more based in the art-field (Behnke, 236). This means they adopt ceremonies that are specific to that organizational field accompanied with gestures, behaviour, symbols and settings to create a shared meaning (Trice and Beyer, 1984, p. 653). Examples of these ceremonies are the quality signals; awards and exhibitions.

Firm with motivations related to 1) corporate collections as investments, 2) corporate philanthropy and cultural responsibility and 3) corporate symbolism, most likely organize public exhibitions and award artistic endeavour (Kottasz et al., 2007, p. 22). For them media publicity increases the value of their collection, acknowledges and publicizes their responsible behaviour and helps them to build a reputation among the general public (Deephouse, 2000).

2.5.1 Exhibitions

The exhibitions organized by the firms vary from showcasing work from the collection to temporary exhibitions by artists (VBCN, 2017). They are given the opportunity to gain experience and receive publicity. Both the artists and the firm benefit from the publicity generated from exhibitions (Matorella, 1990, p. 22). Since the cultural boom commercial firms organized large art exhibitions. This made major players in the corporate world more conscious of each other’s endeavour, and more aware of the prestige to be derived from collecting (Matorella, 1990, p. 22).

Exhibitions inform the public and can change attitudes and behaviours of the

spectator (Dean, 2002, p. 2). This might also include a change of attitudes towards the firm, which refers to exhibitions as a promotional tool. Moreover, exhibitions are a

(14)

way to communicate and provide proof of professionalism and responsible handling of the collection, which attracts dealers and artists (Dean, 2002, p. 2).

Public exhibitions organized by the VBCN members come in several forms, both private and public. The private exhibitions are organized for employees and clients. This study is concerned with the public exhibitions with invited press and related press releases. They come in several forms: 1) The collection can be

temporarily showcased in an external public art space, like a museum, 2) exhibitions can be organized in (the central hall of) the office building, 3) or in the corporate gallery (VBCN, 2017). Loans to external parties are not included in this research, mainly because they are not always for public exhibitions purposes, like loans to other offices.

When corporations organize public exhibitions, they do so to garner media attention (Behnke, 2007, p. 230). Without any corporate communication and activities there is very little to impress the media with. To make sure the positive news

dominates the negative news, firms express their positive attributes (Connelly et al., 2011, p. 53). Exhibitions are an effective tool to create positive news and change overall perceptions on the firm. Negative news on corporate collections, like cutting

or freezing cultural budgets4, is compensated by positive news on corporate activities

like awards and exhibitions. Even more important; communication that make a firms activities transparent enables the public to appreciate the firm’s overall operations better, and so facilitates a better reputation (Fombrun and Rindova, 1998, p. 210).

H1a: The number of exhibitions has a positive effect on the media valence.

Sending more observable signals can enhance the effectiveness of the signal to be picked up by the media. Exhibitions and awards are more observable signals for the media than, for example, loans and private exhibitions (Connelly et al., 2011, p. 53). Also, increasing the number of signals, which is called signal frequency, can enhance the effectiveness (Connelly et al., 2011, p. 53). By organizing several exhibitions annually the collection stands a greater chance of being noticed (Connelly et al., 2011, p. 53). Thus, besides making sure the signals stand out, repetition can increase the effectiveness of the signaling process (Connelly et al., 2011, p. 54). Frequent temporal

4 See “Bedrijven bevriezen kunstbudget” (2009) Dagblad De Limburger, LexisNexis.

(15)

exhibitions mainly go together with a corporate gallery or separate art spaces5. Journalists writing the cultural agenda on local exhibitions, openings and other cultural activities are likely to include these exhibitions in their review.

H1b: The number of exhibitions has a positive effect on the media volume.

2.5.2 Awards

Awards are a widespread phenomenon. They can create social recognitions and status and can serve as a valuable tool to influence behaviour (Frey and Gallus, 2015, p. 190). Award come in different forms and are conveyed by actors from different fields like science, sports, the business sector and the cultural industries (Frey and Gallus, 2015, p. 190). In the cultural industries, established and broadly publicized awards have become global cultural icons signifying success of the awarded (Anand and Watson, 2004, p. 60). The studies of Anand and Watson (2004) and Gemser, Leenders and Wijnberg (2008) acknowledge the role of the award as a signaling device. Awards function as a signal of quality in different ways for different actors. For an artist an award or nomination signifies recognition of professional achievement (Anand and Watson, 2004, p. 60). For the general public it serves as a signaling device that assists in their selection process (Gemser, Leenders and Wijnberg, 2008, p. 26). For the firms the award establishes a relationship with the artist (Frey and Gallus, 2015, p. 190). The receiver accepting an award is expected to agree with the general goals of the giver. Similarly, the giver is expected to stay supportive towards the recipient (Frey and Gallus, 2015, p. 190). The corporate support is usually not simply philanthropic, but the product of a strategy that helps to advance the creative process and to gain access to it (Jacobson, 1996, p. 244).

Anand and Watson (2004) refer to awards as tournament rituals. Tournament rituals have a definite symbolic structure that is differentiated from the other activities in the field by taking the form of a public spectacle (Anand and Watson, 2004, p. 60). The symbolic structure is designed to attract attention within a field (Anand and Watson, 2004, p. 60). The role of awards as a signaling device seems particularly important in the cultural industries, where the quality of products is mainly socially constructed (Gemser et al., 2008, p. 26). Awards thus signal information of quality artworks, practices and trends, which makes it interesting for the media intermediaries

5 See corporate gallery and separate art spaces of DNB, Rabobank, LUMC, Akzo Nobel, AMC and

(16)

to communicate on this signal.

H2a: Having a corporate art award has a positive effect the media volume.

We argue that awarding artistic endeavour generates more and positive media

coverage as opposed to no award. Similar to exhibitions these firms create a sense of expertise and sophistication around their corporate collection by displaying positive firms attributes (Connelly et al., 2011, p. 53). Reviews about corporate art awards cover subjects like; winners, nominations, opportunities given by the firm, artworks,

artist careers and purchases for the corporate collection (ABN Amro)6. By conveying

an award, the firm influences the media coverage positively.

H2b: Having a corporate art award has a positive effect the media valence.

2.6 Quality signals: Annual reports and press releases

It is important to acknowledge the two-way interaction between the media and the firm. The activities and communication of the firm influences the media coverage, but this also occurs the other way around, where the media influences firm’s behaviour. As the media increase its attention on a firm, management may likewise view the media as a more powerful and legitimate stakeholder (Carter, 2006, p. 1152). Given the pervasiveness and agenda-setting power of the media, they can exert influence on the operations of the firm (Wartick, 1992, p. 33). The media has a societal role and monitors the impact of business practices and tend to be critical about the amount of attention companies pay to their responsibilities (Dawkins and Lewis, 2003, p. 187). More media exposure of firm operations increases the corporate responsible

behaviour, where criticized operations can be compensated with responsible activities (Wartick, 1992, p. 33). Corporate responsibility can thus be used to obtain a more favourable overall media reputation (Cahan et al., 2015, p. 410). Cahan et al. (2015) find that firms that perform better in CSR areas are viewed more favourably in the media (p. 419). By creating transparency in their cultural responsible operations and communication, they improve overall stakeholder evaluations (Burke, 1998, p. 8). When acknowledging the power of the media, the firm tends to focus more on influencing the media more directly via several communication tools.

(17)

One of the primary roles of corporate communication is to sustain and develop an organization’s reputation (Burke, 1998, p. 8). Reputation building involves trying to manage information through third parties (Jones, 2002, p. 219). This can be more easily done with signaling strategies that attempt to frame how signals are perceived (Jones, 2002, p. 219). One of these signaling strategies is impression management (Hooghiemstra, 2000). Through impression management the firm impresses the stakeholders and the media by its corporate responsible or philanthropic actions and so justifies its continued existence (Hooghiemstra, 2000, p. 57). Two ways for a firm to do this is by sending press releases and publishing annual reports.

2.6.1 Annual reports

Corporate communication is a management instrument. Both internal and external communication is balanced to create a favourable basis for relationships with groups where the company depends upon (Hooghiemstra, 2000, p. 57). Following the impression management theory, press releases and annual reports are tools used by managers to influence outsider’s perceptions of the firm (Bansal and Clelland, 2004, p.95). Most of the work in impression management is based on the study of Goffman (1959). He describes how firms use communication tools to create a particular impression for others (Osma and Guillamón-Saorín, 2011, p. 188). Through annual reports the company legitimizes its behaviour by intending to influence stakeholder’s and eventually society’s perceptions about the corporate collection (Hooghiemstra, 2000, p. 57). The firm impresses the stakeholders by its actions and so justifies its continued existence (Hooghiemstra, 2000, p. 57). Impression management is used to manage a firm’s reputation as well, where management makes key decisions

regarding actions used to promote or defend a firm’s reputation (Carter, 2006, p. 1148). A firm may, for example, attempt to weaken damaging signals through conveying many positive signals (Jones, 2002, p. 219). Corporate collections are a transparent activity, therefore firms generally don’t mind to disclose information in their annual reports. The press, in turn, is likely to proactively seek out information on responsible behaviour of a firm, and to use the company's annual report to get a comprehensive picture of the activities (Dawkins, 2004, p. 191).

H3a: Publishing an annual report about the corporate collection has a positive effect on the media volume.

(18)

Through annual reports firms can communicate and specify the positive impact of corporate collection and its related activities on society. To influence the perception on the firm positively, management is willing to report on positive information but reluctant to disclose the negative (Hooghiemstra, 2000, p. 57). This implies that disclosures on the corporate art collection are largely self-laudatory (Hooghiemstra, 2000, p. 57). The annual reports on corporate art collections are thus mainly a source of good news that influences the media reputation positively.

H3b: Publishing an annual report about the corporate collection has a positive effect on the media valence.

2.6.2 Press releases

A second communication tool used in impression management is the press release (Carter, 2006, p. 1146). Press releases represent an important instrument through which the firm communicates on the corporate art collection and related activities (Carter, 2006, p. 1150). Firms can use press releases to influence third party

perceptions for their own benefit (Bowen et al., 2005) (Osma and Guillamón-Saorín, 2011, p. 187). Press releases provide firm’s versions of information that is deemed to be of interest to the general public, in the hope that journalists will pass it on (Maat and de Jong, 2012, p. 349). Previous research showed that releases influence what reporters write (Hong, 2008; Morton and Warren, 1992; Walters and Walters, 1992; Turk, 1985) (Maat and de Jong, 2012, p. 349). Each press release contains

information that can influence how the media reports firm behaviour and activities and thus may ultimately result in a change in the firm’s reputation among various stakeholder groups (Paul, 2001) (Carter, 2006, p. 1150).

Press releases about the corporate art collections cover activities like

exhibitions and art space openings7. In general press releases are easily accessible by

the public and represent an important instrument through which managers

communicate firm performance (Osma and Guillamón-Saorín, 2011, p. 187). The press releases regarding the corporate collections, on the contrary, are generally not available online. They are communicated by email, in the form of an invitation with

(19)

additional information on the event and artist8. Curators or managers select the information to be released with the intention to influence the perceptions about the collection and the firm (Osma and Guillamón-Saorín, 2011, p. 187). Releases typically focus on specific events and activities of the company, ignoring the wider contexts (Maat and de Jong, 2012, p. 352). Moreover, they tend to tell good news or at least present their information as favourably as possible from the corporate

viewpoint (Bhatia, 2004; Catenaccio, 2008; McLaren and Gurǎu, 2005) (Maat and de Jong, 2012, p. 352). Firms thus report the most favourable items within the whole range of information available, focusing on positive outcomes and using positive language to convey a positive view of corporate activities.

H4a: The number of press releases has a positive effect on the review valence.

While annual reports and award can only signal quality annually, exhibitions and press releases have in common the advantage of more frequent signaling

opportunities. The repetition of the signal increases the effectiveness on the media reputation (Connelly et al., 2011, p. 54). Moreover, press releases are addressing the media directly, which makes them more observable and therefore even stronger signals to influence the media reputation. Hong (2008) finds that mainly newsworthy releases receive greater amount of coverage (p. 298).

H4b: The number of press releases has a positive effect on the review volume.

The local community forms an important stakeholder of the firm, especially when it comes to community involvement and contribution in the form of social responsibility (Dawkins, 2004, p. 110). Therefore it is expected that firms send their press releases mainly to local newspapers. Moreover, past research implies that local print media best covers local businesses (Deephouse, 2000, p. 1100). This means that press releases probably are most effective in influencing coverage when they are send to local newspapers. Morton and Warren (1992) found a high acceptance rate for the hometown releases of thirty-two percent in relation to general releases ranging from three to eight percent (1992, p. 389). They conclude that press releases coming from local businesses are an effective tool when send to local newspapers (Morton and

(20)

Warren, 1992, p. 389). Based on the calculated credibility score, as will be later explained in the methodology, local newspapers are generally less credible than national newspaper. Following this argument a third hypothesis for press releases is included.

H4c: Press releases have a negative effect on credibility of the source.

3. Methodology

In this section the research design and data collection are described. Subsequently, the sample and the variables are operationalized.

3.1 Research Design

The empirical setting of this study is corporate art collections in the Netherlands, specifically focused on the VBCN. The VBCN represents the largest share of the corporate art collections in the Netherlands with forty-nine cross-industrial members. From the total number of firms four were excluded from the research due to the lack reviews, leaving a sample of forty-five firms.

In order to answer the research question and to test the hypotheses introduced in the literature, an appropriate research method must be chosen. The data is collected from several sources and analysed using a panel data structure showcasing data on the VBCN members per year from 2005 till 2016.

3.2 Data collection

This research relied on five data sources. The primary sources are (1) the curators or managers of the corporate collections. The secondary sources are (2) the VBCN website, (3) the corporate website or the foundation website, (4) the LexisNexis database, (5) and the general web.

Information regarding awards, exhibitions, annual reports and press releases was obtained from the VBCN website, the corporate websites and personal contact with curators or managers. In general the corporate website displayed more loans, exhibitions and other activities, than the VBCN website indicated. Therefore, the corporate websites were used for more detailed information on exhibitions, awards and annual reports. All the members of the VBCN have been contacted by email or phone to confirm and complete the data.

(21)

VBCN – The VBCN is a platform for corporate collection in the Netherlands. They

stimulate collaborations between the cultural industries and non-cultural firms (VBCN, 2017). The values of the association are independency, dynamically and stimulating (VBCN, 2017). They organize the VBCN curator prize, member meetings and currently participate in a large UVA research project. The VBCN website

(http://www.vbcn.nl/NL/) offers specific information on members and their activities regarding yearly exhibitions, loans, publications, size of the collections, exhibitions spaces, contact person (curator or manager) and a reference to the corporate (art) website.

LexisNexis - Following the methodology proposed by Deephouse (2000) the

media reputation is measured by a content analysis of a newspaper archive (p. 1092). This study retrieved articles from LexisNexis (Zhang, 2015). The LexisNexis

database offers access to newspapers, legal and financial information of ten thousand international sources (LexisNexis, 2017). The database contains data of more than 3000 newspapers, including Dutch and foreign coverage from 1980 until today (LexisNexis, 2017). LexisNexis is freely accessible for students, and accurate; new content is added daily. LexisNexis is the only database in the market offering

extensive information on media coverage over a longer time period. This makes it the most appropriate and efficient database for this research. To search the news articles in the local newspapers, the following search method was used: Select ‘all news’ on the homepage (http://academic.lexisnexis.nl), in the next option you specify on ‘all available dates’ and ‘all Dutch news’. The search bar offers a maximum of five search terms. The more search terms are used simultaneously, the more specific and narrow the search. All firms have been individually searched for in combination with one keyword each time (Appendix, table 5). Every pair of words (firm + keyword related to art collection) creates a new search task. This approach is taken to keep the

database as broad as possible but still related to the art collection of the specific company. Every article is carefully read to confirm its relevance.

Corporate websites – The information on the corporate website differs per

firm. Twenty-eight of the VBCN members have included information about their art collection on their corporate website, or a separate website in case of the NOG collection of SNS Reaal. This leaves twenty-one firms who did not incorporated any information about the art collection on their website. The corporate websites can be found in the reference list.

(22)

3.3 Dependent variables

As mentioned in the theory on media reputation, the media salience construct of the agenda-setting theory consists of three elements: attention, valence and prominence (Kiousis, 2004, p. 71). The third dimension, media prominence, is not included in this study, mainly because the NexisLexis database doesn’t show any pictures, colours or special layouts of the reviews. Furthermore, the placement of the articles is not always indicated or clear, because the reviews are taken out of their initial context of the newspaper of magazine. Besides, if information on pages and sections is given, the information is still hard to compare, since the sections and total number of pages differ per source. The other dimensions of media salience, as explained in the theory, are included in the database and clarified below.

The dependent variable captures the firm’s media reputation per year, signaled through newspaper reviews from fifty-one Dutch newspapers and operationalized by, (1) volume of reviews, (2) valence of reviews, (3) count of group or count of solo, and (4) credibility of reviews.

Volume – Zhang (2015) in “Measuring Media Reputation” discusses the

quantitative analysis on media reputation based on research of Fombrun and Shanley (1990) and Wartick (1992). In both studies media reputation is measured by visibility and tonality of the media coverage. Zhang (2015) concludes that media favourability (measured by the values of the tonalities) together with media visibility (measured by the number of news items) are better indicators of media reputation than solely focusing on media favourability (Zhang, 2015, p. 898). Visibility is the number of reviews that address the corporate art collection and is counted yearly per company. All the reviews are read to double check the content of the review. Some articles refer to more than one collection of the VBCN. Therefore, two extra variables were

created; count of group and count of solo. Count of group represents reviews mentioning a minimum of two VBCN members, while count of solo reviews cover only one VBCN collection. The total number of reviews per year is the sum of the count of group and count of solo.

Valence - Media tonality or valence was calculated by an online linguistic

inquiry and Word Count (LIWC). The program calculates and records the positive and negative emotions within a text. The valence of the text shows the attitude of the review toward the corporate collection. Especially when measuring the reputation, the

(23)

attitude is an important indicator (Zhang, 2015, p. 898). LIWC2015 is a computerized text analyser. It is accurate, easy to use, and provides a broad range of social and psychological insights. LIWC reads a given text and counts the percentage of words that reflect different emotions, thinking styles, social concerns, and even parts of speech. Per firm the mean valence is calculated on a yearly basis.

Credibility – Third party signals from the media are perceived as far more

valuable when compared to quality signals that stem from firms themselves (Dean and Biswas, 2001, p. 44). Third parties, like the media, are able to independently rank, rate and review corporate collections (Dean and Biswas, 2001, p. 44). But also in the case of independent reviews, actors do not simply accept the content of the review. Based on salience concept of the attribution theory, the credibility of the review is included as a dimension of media reputation (Gemser et al., 2008, p. 27). An article about the corporate collection from the Vrij Nederland is more credible and signals stronger validity compared to a reviews from a regional newspaper like Noord Hollands Dagblad. The credibility scores of the newspaper sources are based on an earlier conducted credibility score test.

Monika Kackovic offered a list of 353 Dutch and foreign written media sources with related credibility scores. These scores were calculated using a multi-item and validated scale for the source credibility construct (Ohanian, 1990). Two experts, with around twenty years of experience were chosen from a random group of experts, and the third is one of the authors. To measure the credibility level of these third party sources, the selected experts were asked to rate these media sources based on a six items scale: trustworthiness, honesty, dependability, experience, expertise and how knowledgeable a source is (Ohanian, 1990). A 5-point Likert measurement was used. The experts were allowed to base their rating on a quick Internet search, whenever the judge was less familiar with the source. If he would still not be able to judge the source, the ‘do not know source’ option could be checked. The internal validity of the six items measured was high, with a Cronbach’s alpha of .98. Inter-rater reliability was calculated by using an intraclass correlation coefficient (ICC), specifically a two -way random effects model (Shrout and Fleiss, 1979). This resulted in an intraclass correlation (ICC 2,3) of .744 and significant at the .001 level. This indicates the judge’s ratings to be internally consistent, which suggests low random and specific errors (Bravo and Potvin, 1990).

(24)

Eleven additional newspaper sources had to be validated: Almere Vandaag, BN/De Stem, De Persgroep, Dagblad van het Noorden, de Peperbus, De Nieuwsbode Zeist, Forum, Het Nieuwe Stadsblad, Reformatorisch Dagblad, Quote and Spits. Two experts were chosen from a random group of journalists to validate these additional sources. The first expert was an UVA academic, teaching the course Journalism in the Netherlands. The second expert was an AD journalist. The two judges individually scored the nine sources based on similarity with the rated newspapers from the credibility score list. In an email they received the credibility score list and background information on the method. An average score was taken in case of a different assessment by the two judges. The credibility score list used for this research can be found in the appendix (table 6). For each firm the mean credibility scores per year were calculated.

3.4 Independent variables

The four independent variables are: awards, exhibitions, annual reports and press releases.

Awards - Information on awards was retrieved from the corporate websites.

Award is a dummy variable that takes the value 0 or 1 to indicate the absence or presence of an award in a given year. Leaseplan, Menzis, KPN, ING, ABN Amro, F. van Lanschot and since 2017 the NN group all award artists. These awards are related to both the corporate collection and its exhibitions. For instance, the KPN Kunstprix in the past purchased one work of the winning artists for the KPN collection every

two years (PrixdeRome)9. Furthermore, the ABN Amro Art Award invites the winner

to exhibit in the Hermitage Amsterdam (ABN Amro)10.

Exhibitions - The VBCN website, corporate websites and contact persons were

valuable sources for gaining information on the yearly public exhibitions. Some collections run a separate art space, where external parties are invited to exhibit work. These exhibitions are included in the count, because they are part of the related activities of the corporate collection. The curator or manager of the corporate

9 retrieved from:

http://prixderome.nl/2011/07/27/pilvi-takala-wint-prix-de-rome-beeldende-kunst-2011/

10 retrieved from:

https://www.abnamro.com/en/newsroom/press-releases/2016/abn-amro-art-award.html

(25)

collection organizes them and these exhibitions are related to awards and purchases. For each firm, the number of yearly public exhibitions was counted.

Annual reports – In total nine firms from the sample signal their corporate

collection through an annual report. Three firms including the ING, Erasmus

University and SNS Reaal, incorporated information on their corporate collection in

the general annual report, by relating it to art and culture or societal relevance11. In the

other six cases the firm related art foundations published the annual reports. Without the foundation the collection might be closed one day, as a result of management changes or mergers (Behnke, 2007, p. 231). The foundations posting their annual report online are: Stichting Kunstcollectie KPMG, Stichting Kunstcollectie Essent-Enexis, Stichting Beheer Kunstcollectie Zuyderland MC, Kunst en historisch bezit ABN Amro en de AkzoNobel Art Foundation. These foundations are not the only ones holding a corporate collection, but the remaining foundations do not post their annual reports online. This either implies that they do not have an annual report or they intentionally do not disclose them. The corporate websites and foundation websites are consulted for the annual reports. A corporate annual report is presented once a year, which means it is a dummy variable that takes the value 0 or 1 to indicate the absence or presence of an annual report in a given year.

Press releases - Press releases on corporate art collections are barely

published online, therefore the number of press releases per year has been confirmed by all firms via email or by phone. Most press releases seemed related to events like exhibitions and openings. These press releases are generally not found online, but send to the press in the form of an invitation with background information on the event and artist.

3.5 control variables

This study controlled for three variables that can be important predictor of media reputation: Collection size, firm size and collection age.

Age of the collection - The age of the collections is presented on the VBCN

website. The oldest collections are the DSM, DNB and F. van Lanschot collections from above a hundred years old. The age of the collection is expected to influence the media reputation. On the one hand, the oldest collections had the longest time frame

11

Retrieved from: http://www.beheersnsreaal.nl/statuten

Retrieved from: https://www.eur.nl/eur/corporate_publicaties/jaarverslagen/ Retrieved from: https://www.ing.com/Investor-relations/Annual-Reports.htm

(26)

to receive reviews. On the other hand, these collections already exist for a long time and might get out of fashion, which makes them less interesting for the media to cover than newer collections (Hirsch, 1972, p. 649). Following Hirsch (1972) mass-media gatekeepers serve as institutional regulators of innovation in the cultural industries (p. 649).

Firm size – The size of the firm is the second control variable that might

influence the media reputation. Firstly, larger firms are more visible to the public and the media (Graafland et al., 2003, p. 45). Secondly, large firms are able to invest heavier in their corporate collections and (cause related) marketing communication to attract media attention. Because of their larger scale, the costs involved are relatively lower (Graafland et al., 2003, p. 45). Firm size is operationalized in the number of employees. Hart and Oulton (1996) argue that employment, assets, sales, market value, and value added are the most common measures for company size, but each has its limitations. In practice, the choice of the type of measurement is based on the available data (Hart and Oulton, 1996, p. 1243). A limitation of the method used for this study is that the number of employees is a discrete variable, meaning the variable doesn’t have an infinite number of values. Moreover, some employees, like

freelancers, are not always recorded. Therefore, it is generally better to base the measurement on total fte, fulltime-equivalents, but these numbers are not always reported (Hart and Oulton, 1996, p. 1243). The number of employees per year is stated in the corporate annual reports posted online. The most recent annual reports available are used as a source for this variable.

Collection size - The number of works per collection is given on the VBCN

website. For the same reasons as firm size, it is expected that larger collections

generate more attention than smaller collections. Because of the greater news value, it is expected that larger collection have a stronger media reputation.

3.6 Method

The dependent variable media reputation has several dimensions. Review volume,

review positivity and negativity, review credibility, count of group (shared reviews)

and count of solo (reviews dedicated to one specific corporate collection of the VBCN). Four independent variables and the three control variables are all measured along these dimensions. Multiple linear regressions were performed to test the hypotheses. The independent variables award and annual report are dummy

(27)

variables. Exhibitions and press releases are count variables; these variables can only take on zeros and positive discrete values (Coxe et al., 2009). The dataset included many zeros, meaning there were no signals conveyed. Therefore, the filter variable was used to exclude the zero-cases. Furthermore, the dataset included some missing values, mainly on press releases. The discrete missing values (9999) were excluded. The main cause of the missing values was the dependency on older corporate databases that, in some cases, could not deliver the data requested. The remaining dataset of 217 cases was used for further analyses. SPSS 24 was used for all statistical analyses.

4. Results

In this section the results are presented. First, the descriptive statistics of the sample are described. Second, correlations are presented. Third, the results of the regression analyses are discussed.

4.1 Descriptive statistics

Table 1 presents the descriptive summary of the sample. Firstly, the quality signal

exhibition has a mean number of 2.92 per year with a standard deviation of 3.26.

Secondly, the mean of the quality signal award is .07 with a standard deviation of .26. Thirdly, the mean number of the quality signal press releases is 1.39 per year with a standard deviation of 2.18. Lastly, the mean of the quality signal annual reports is .20

(28)

with a standard deviation of .40. Exhibitions and press releases can be compared because they are both count variables. The data shows that exhibition is a relatively more conveyed quality signal than press releases. Award and annual reports are dummy variables. By comparing these variables it becomes visible that annual report is a relatively more conveyed quality signal than award.

Furthermore, descriptive statistics also indicates a higher mean for positive

reviews .79 with a standard deviation of .50 (minimum 0.00; maximum 3.89), than for negative reviews .46 with a standard deviation of .34 (minimum 0.00; maximum

2.17). The mean count of solo 2.48 with a standard deviation of 3.63 (minimum 0; maximum 27) is higher than we mean counts of group 1.40 with a standard deviation of 2.40 (minimum 0; maximum 21). This indicated that the reviews more generally dedicate the article to one corporate collection of the VBCN, instead of several collections.

4.2 Correlations

Table 2 presents the correlations between all variables. The table shows that annual

report, award, size collection and size firm have significant positive correlations with review volume (resp., r= .23, p < .001; r = .31, p < .001; r= .28, p < .001; r= .22, p <

.001). This could suggest that larger firms and larger collections receive more coverage and that the collections with an art award or annual report receive more reviews. Other than review volume the table shows that press releases are negatively correlated with review credibility (resp., r= -.26, p < .001). This could suggest that

(29)

newspapers with lower credibility scores cover firm sending press releases. It could also suggest that firms intentionally send it to local newspapers, which have lower credibility scores than national newspapers. Size firm is significantly and positively correlated with review credibility (r= .12, p = .048). This could suggest that larger firms receive coverage from more credible sources, like national newspapers. Firm

size is also positively correlated to review positivity (r= .15, p= .017). This could

suggest that the coverage on larger firms is more positive. Moreover, annual report,

award and collection size have a significant and positive correlation with count of solo (resp., r= .16, p = .01; r= .24, p < .001; r= .17, p= .01). Annual report, award, size collection and size firm are also positively correlated with count of group (resp.,

r= .25, p < .001; r= .30, p < .001; r= .35 < .001; r= .30, p= .001).

Finally, the control variable size firm had significant positive correlation with

annual report (r= .24, p < .001) and significant negative correlations with exhibitions

and press releases (resp. r= -.24, p < .001; r= -.21, p= .002). This could suggest that larger firms engage more in conveying annual reports on the art collection, but generally engage less in press releases and exhibitions. The control variable size

collection positively correlates with annual report, award and exhibitions (resp. r=

.22, p < .001; r= .23, p < .001; r= .29, p < .001). The last control variable age

collection positively correlates with exhibitions (r= .17, p= .007) and negatively

correlates with annual report (r= -.21, p = .001). This could suggest that older

collections engage more in exhibitions and less in publishing annual reports about the corporate collection.

(30)

Table 3 presents the seven regression analyses performed. In the table the Unstandardized B, Standard Error, the significance level, R2 and R2 Change are included. Model 1 of each regression presents the results of the control variables size

firm, size collection and age collection. In model 2 the independent variables are

added.

The first two regressions both calculate the predictors for review volume; the first regression with dependent variable review volume and the second regression with a logarithmic transformation of review volume. Logarithmic transformation is a convenient way to transform a highly skewed variable into one that is more approximately normal (Benoit, 2011, p. 2). In model 1 of the first analysis a

significant regression was found (F (3, 213)= 13.67 p < .001) with a R2 of .16, which indicates that 16% of the differences in review volume are explained by the control variables. Model 2 allows us to test hypotheses 1b, 2a, 3a and 4b; resp., exhibitions, awards, annual reports and press releases have a positive effect on the review volume. A significant regression is found (F (7, 209)= 7.38, p < .001) with a R2 of 0.20, indicating that 20% of the differences in review volume are explained by the independent variables and control variables. The R2 Change is .04, so 4% is explained by the independent variables. The model calculates a positive significant effect for size collection (R= 3.45, p= .001) and for award (R= 2.22, p = .03). Only hypothesis 2a, a corporate art award has a positive effect the review volume, is supported. The review volume increases 2.97 with each corporate award.

The first regression on review volume calculated a Skewness of 4.59, which means it is skewed to the right. The right tail is longer and the mass of the distribution is concentrated on the left side of the figure. A Skewness of 0 means perfect

symmetrical, so 4.59 is very high. The variable review volume has to many outliers. These outliers make any prediction unreliable. Therefore a second regression analysis of review volume was performed, this time transforming the variable with LN

(logarithmic transformation) to reduce the Skewness.

Table 3 presents the results of the second regression with dependent variable

LN review volume. Model 1 includes the control variables with a significant

regression (F (3, 213)= 16.18, p < .001). The R2 is .19 which means 19% of the differences in review volume is explained by control variable size collection.

Model 2 shows a significant regression (F (7, 209)= 8.24, p < .001) with a R2 of .22. 3% of the differences in review volume are explained by the independent

(31)

variables. The model shows a positive significant effect for size collection (R= 3.90, p < .001) and age collection (R= 2.02, p= .045). Which suggests that larger and older collections receive more reviews. The coefficient of annual report is positive and significant (R= 2.60, p= .01). The review volume increases .40 with each annual report. Hypothesis 3a is supported: Publishing an annual report about the corporate art collection has a positive effect on the review volume.

Both regression 1 and 2 are included into table 3. Statistically LN review

volume is more reliable, it enables us to make generalizable statements about all

VBCN members. But we could argue that, although regression 1 is statistically less reliable due to the outliers, these outliers seem to attract most attention from the media compared to the rest of the collections and attracting this media attention is exactly what this research is about.

The third regression tests the predictors of review positivity, one dimension of review valence. It allows us to test hypotheses 1a, 2b, 3b and 4a. The results show that awards, exhibitions, annual reports and press releases are not significant

predictors of review positivity. The hypotheses are not supported. Review valence can be both positive and negative, therefore review negativity was also tested in a fourth regression, but no significant results were found.

The fifth regression tests the effect on the review credibility. In model 2 the control variables and independent variables are good predictors of review credibility (F (7, 209)= 3.09, p = .004) with a R2 of .09. Model 2 calculated a negative

significant effect of press releases (R= -3.63, p < .01). Hypothesis 4d is supported: press releases have a negative effect on the credibility of the source. The independent variable press release significantly explains 6% of the difference in review credibility. This could suggest that firms only send press releases to local newspapers that

generally have a lower credibility score than national newspapers. The result could also suggest that the press releases are send to several newspapers and that only local newspapers respond to the press releases sometimes.

The last two regressions test the effect on the count of group and count of solo. Model 1 of the sixth regression shows the control variables are good predictors of

count of group (F (3, 213)= 18.55, p < .001) with a R2 of 0.21.This means 21% of the

change in count of group is explained by the control variables. Model 2 also shows a significant regression (F (7, 209)= 10.39, p < .001) with a R2 of .26, which means 5% of the change in count of group is explained by the independent variables. Like model

(32)

1, model 2 shows significant results for size firm and size collection (resp., R= 2.32, p= .02; R= 3.83, p < .001). Furthermore, model 2 shows a positive significant result for award (R= 2.58, p = .01). Large VBCN collections or collections from large firms are more mentioned in reviews together with other VBCN members. The same

applies for awards.

The last regression with the dependent variable count of solo shows that the control variables are good predictors of count of solo (F (3, 213)= 5.41, p = .001), with a R2 of .07. The collection size has a positive significant effect on count of solo (R=3.65, p < .001), which explains 7% of the change in count of solo. Model 2 calculated a significant regression (F (7, 209)= 3.68, p < .000) with a R2 of .11. The coefficient of press releases is positive and significant (R=2.01, p =0.05). This result indicates that press releases about the corporate collection influence the count of solo, meaning these reviews are solely focused on one specific VBCN collection.

Furthermore, the collection size in model 2 is still significant (R= 2.13, p= .03). Together with the first regression and regression six the results indicate that larger collections predict a higher total review volume, both in groups and solo.

Referenties

GERELATEERDE DOCUMENTEN

So, chromosome cohesion defects can result in premature sister chromatid separation, which bypasses the spindle assembly checkpoint and leads to incorrectly attached and

The qualitative review helped to contextualize and interpret the results found in the meta-analysis, which revealed a large effect for mobile technologies in

The research questions addressed how attitudes toward Muslim immigrants are affected by news framing (RQ1), and questioned the moderating roles of political knowledge and

Drawing upon the example of arts &amp; humanities scholarship, a domain that suffered greatly under economic reductionism, Eleonora Belfiore and Anna Upchurch use their new

In the present study is the general public defined as: the individuals within the external environment, were the organization has direct interest in, or individuals have

De studiepopulaties (n=641 en n=696) betroffen patiënten met door opioïden geïnduceerde constipatie (OIC) bij behandeling met een stabiele dagdosis opioïden over een periode van

More precisely, the influences of environmental CSR (ECSR) dimensions comprising the three pillars of Asset4 (emission reduction, resource reduction, product

Results indicate a significant and positive relationship between firm’s social media popularity and market capitalization and market-to-book ratio, while controlling for firm size,