Organization-level determinants of employees’ cultural consumption : a study of internal signalling qualities of corporate art collecting

(1)

Organization-level Determinants of Employees’

Cultural Consumption. A Study of Internal

Signalling Qualities of Corporate Art Collecting.

Master Thesis - MSc Business Administration

Entrepreneurship and Management in the Creative Industries Track

Name: Zuzanna A. Szczepecka Student number: 10828532

Thesis Supervisor: Monika Kackovic

(2)

Statement of Originality

This document is written by Student, Zuzanna A. Szczepecka, 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

This thesis examined the dynamics of corporate art collecting, focusing on internal (employee-oriented) process corporate collections facilitate and participate in. In a broader perspective, the project aimed to advance the understanding of organization-level determinants of employees’ behaviours, especially outside of their immediate work environment. In particular, it focused on the influence of firms’ involvement in art collecting practices on their employees’ behaviours as cultural consumers. The questionnaire-based research was conducted among a sample of 486 employees from five Polish companies - including three with and two without corporate art collections. The thesis studied the effect of firms’ collecting activities on the frequency of employees’ cultural consumption by comparing the two sample sub-groups. It also addressed the differentiating effect of collections’ specific profiles on the studied relationship. Finally, it looked into the moderating effect of corporate art collecting on employees’ educational attainment and occupational status, and their cultural consumption. Although the results did not support any of the proposed hypotheses, they, nonetheless, provided some meaningful implications for both future research as well as business practice.

(5)

1. Introduction

Corporate art collections (CACs) - ‘any and all art found within collections held by corporations’ (Martorella, 1990; 4) - are one of the most prominent forms of affiliation between the art and the business worlds. Over half of the Fortune 500 companies and many other major enterprises around the globe are currently found to be actively engaged in purchasing art and development of their own collecting programmes (Kottasz et al., 2007). Although, in a traditional perspective, CACs have mostly been seen as another source of tax benefits or diversification of corporate investment portfolios, firms’ increasing commitment to collecting practices and CACs’ rising professionalization has shifted researchers’ perception about the role and value of such activities for business (Martorella, 1990; Behnke, 2007). CACs are no longer assessed only according to their financial value but above all to their potentially lucrative symbolic nature (Cassullo, 1988; Wu, 2003; Coomber, 2006).

This thesis will investigate CACs’ dynamics, with a special focus placed on internal (employee-oriented) processes CACs participate in and stimulate. Although CACs have been operationalized as an important input into diverse corporate strategic and communications activities, the majority of the existing research has focused on their role in addressing firms’ external stakeholders (Barendregt et al., 2009). The internal perspective, despite being repeatedly mentioned as of key importance for the legitimization of art collecting activities in companies (e.g. Martorella, 1990; Wu, 2003), still lacks a more in-depth understanding and wider academic coverage.

In a broader context, the project seeks to add to the understanding of organization-level determinants of employees’ behaviours, not only at-work but also outside of their immediate work environment. More specifically, with the use of relevant theoretical constructs and empirical research, the thesis will argue for and examine the influence which firm’s art

(6)

relationship will be studied specifically in relation to the frequency of their cultural consumption, when compared to employees from firms not involved in collecting practices.

As briefly indicated above, thanks to their strongly symbolic nature, CACs are increasingly seen as important corporate artifacts serving a whole range of organizational functions and goals. Signalling theory - focusing on the use of observable quality signals for communicating unobservable attributes and values (Spence, 1973) - provides an interesting framework for contextualising these unique qualities of corporate collections. As an example of a first-party signal (a signal generated by a firm itself), a CAC can be seen as a manifestation of company’s wealth, sophistication or prestige but, above all, also the value which a firm attributes to arts and culture (Hoeken & Ruikes, 2005; Lindenberg & Oosterlinck, 2011). Although such signals are usually seen as addressed towards companies’ external publics, Kottasz et al. (2007) has argued that CACs are increasingly perceived as playing a crucial role in signifying key organizational values to firms’ most immediate audience - its employees.

While the available academic discourses have addressed a vast number of factors expected to influence individual’s cultural consumption, little is known about organization-level determinants likely to affect such behaviour. At the same time, the organizational values literature argues that employees’ personal values and those shared by a corporation often interact and align, and thus have a potential to influence workers’ attitudes and behaviours (Meglino & Ravlin, 1998; Gehman, Trevino & Garud, 2013). Whereas corporate values have been repeatedly found to affect employees’ at-work actions (e.g. Akaah & Lund, 1994), they have also been argued to determine the way employees behave outside of their immediate work environment (e.g. Cambra-Fierro, Polo-Redondo & Wilson, 2008). The influence has been contextualized mostly by looking at a socializing agent role of a corporation (Hemingway & Maclagan, 2004). This thesis argues that, taken together with the abovementioned signalling

(7)

theory, such perspective may provide a framework for understanding how and why organization-level determinants - like signals about firms’ values communicated by CACs - are likely to influence behaviours of employees, even when it comes to essentially not work related activities, such as their cultural consumption.

The signalling qualities of CACs are, however, expected to differ depending on collections’ specific features. The research - although quite limited in scope - suggests that characteristics such as collection’s size (Wu, 2003), its budget (Barendregt et al., 2009) or years of existence (Lindenberg & Oosterlinck, 2011) are important factors likely to affect the strength and success of the signals it is intended to communicate. Such features are expected to determine CAC’s position in organization and its strategies but also how the collection is being exhibited and managed. This, in return, is likely to influence the way the artworks are being perceived and interacted with.

Finally, the paper will also look into a potential moderating effect of corporate art collecting on two prominent determinants of cultural consumption, as identified in the relevant literature. This additional investigation, although slightly outside of the main focus of this thesis, will be conducted to provide a more in-depth understanding of the studied relationship and the developed framework of organization-level factors influencing employees’ behaviours. This will be done to highlight the multiplicity of levels and ways of interaction which could be taken into consideration in such a setting. First, the thesis will look into employees’ social status - one of the key variables identified as affecting individual’s behaviours as a consumer of arts and culture (Weber, 1968; Roose, van Eijck & Lievens, 2012). When operationalized as a social status of one’s profession, the relationship suggests that individuals will be found to be more culturally active as we move up the occupational status ladder (e.g. Chan & Goldthorpe, 2007a). At the same time, it is argued that employees’ adoption of organizational values will also be very much dependent on their position in a company - with, for example, managers’

(8)

values being more inline with corporate standards than employees’ at lower status positions (Meglino & Ravlin, 1998; Argandona, 2003). As a reflection of the value which firms attribute to art and culture, corporate collecting is then expected to enhance the relationship between one’s occupational status and cultural consumption, with the effect becoming stronger as we move up organizational hierarchy ladder. Cultural resources are another variable argued as a major determinant of one’s behaviours as cultural consumer (Bourdieu, 1984; Yaish & Katz-Gerro, 2010). When operationalized along the lines of schooling attainment, it is expected to influence one’s cultural consumption by determining individual’s experience in and propensity to enjoy diverse forms of creative expression (van Eijck, 1997). Corporate art collecting activities are often accompanied by various informational sources and educational programmes (Martorella, 1990; Barendregt et al., 2009). It is then argued that by offering an opportunity to not only interact with art on everyday basis but also to advance one’s knowledge on the topic, CACs will positively affect the relationship between employees’ existing cultural resources - defined by their educational attainment - and their cultural consumption.

The data for this thesis was obtained from a study conducted among five Polish firms - including three with and two without CACs. The firms represented three industries known already from previous research for being actively involved in corporate collecting, namely financial, pharmaceutical / cosmetics and real estate sectors (Martorella, 1990; Behnke, 2007; Kottasz et al., 2007). Poland has been identified as a potentially interesting research setting for this thesis as there has been no studies looking into the collecting phenomenon in the country yet (Rogozińska, 2010). Primary data used for the study was collected with an online-based questionnaire and complemented with secondary data from organizational publications, reports, websites and other sources. The studied sample included initially 486 employees - however, the number was reduced down to 349 in the context of study’s results. The gathered data was evaluated with a use of quantitative research methods.

(9)

The remainder of this thesis is organized as follows. The next section (Section 2) will provide a brief overview of the literature on signalling theory, organizational values and various cultural consumption concepts - all discussed in the context of corporate art collecting. The chapter will be used to plot the thesis’s key theoretical constructs and guide the development of its main hypotheses. From there, the thesis will progress into Section 3 discussing the data and empirical methods used to test the identified set of relationships. Section 4 will provide the results from the conducted research and analyses, which, in turn, will be discussed in the context of the previously described literature (Section 5). This chapter will also address the limitations of the conducted study as well as its implications for future research and practice. Finally, a brief conclusion and summary of the project will be provided in Section 6.

(10)

2. Theoretical Framework

2.1. Corporate Art Collecting - Signalling Perspective

The basic signalling framework was first put forward by an American economist, Michael Spence, in his seminal paper on job markets (Spence, 1973). Since then, it has been applied to a wide range of business concepts and research, from strategic management or entrepreneurship to organizational studies and marketing (e.g. Erdem & Swait, 1998; Kirmani & Rao, 2000; Connelly et al., 2011). Signalling theory, essentially concerned with situations defined by information asymmetry, focuses on the use of observable quality signals as a strategy for communicating differentiating yet unobservable characteristics and values (Spence, 2002). Depending on their source, signals can be classified as first-party signals (generated by a signaller himself) or third-party signals (originating from an external person / entity, such as an expert or a critic) (Dewally & Ederington, 2006; Connelly et al., 2011). Although the literature has operationalized signalling qualities of corporate art collections along both of the identified types, this thesis will focus on the first of the mentioned perspectives.

A significant part of the available research in the field has looked at CACs as examples of first-party signals used to communicate a firm’s desired qualities, values or attributes to its multiple stakeholders and publics. Wu (2003) argues that devotion to collecting is never an automatic or random choice. Purchased art becomes one of a company’s tangible assets which requires dedication of organizational resources (e.g. money, time, space) but also brings along certain challenges and risks (e.g. financial or reputational). In a similar manner, Lindenberg & Oosterlinck (2011) suggest that ownership of art represents a long-term commitment to the cause and is often seen as a stronger statement than, for example, one-off sponsorship of cultural events. Within this framework, CACs can therefore be seen as meaningful signals used

(11)

to serve firms’ various strategic goals and to convey certain organizational messages, such as the value which a company attributes to arts and culture.

Some of the literature in the field has looked at the corporate collecting phenomenon from the perspective of marketing or public relations practices - and thus activities addressed mainly at firms’ external stakeholders. CACs have been identified as an important aspect of a firm’s identity and image building practices, as well as a direct input to its marketing and differentiating strategies - a vehicle for communicating organizational wealth, attractiveness and prestige (e.g. Wu, 2003; Behnke, 2007). For example, having a collection might be used to project an image of a sophisticated corporation (Kottasz et al., 2007) or to enhance an overall corporate cultural capital - ‘the social status and value that accrue to a company engaged in cultural practices’ (Wu, 2003; 8). Hoeken & Ruikes (2005) argue that challenges such as increasing competitiveness of business environments or commoditization of firms’ offerings are some of the reasons why modern profit-maximizing and cost-minimizing companies are increasingly interested in engaging in art collecting - practices essentially superfluous to their every-day business activities.

However, with corporate art being exhibited mainly at work spaces, office corridors, canteens, meeting rooms etc. (Martorella, 1990; Smiraglia, 2014), it seems logical to expect that employees will be one of the key recipients of signals which collections are expected to communicate. Accordingly, (internal) employee-oriented processes and benefits are being increasingly mentioned as the main arguments used by companies to legitimize and justify corporate involvement in art collecting (Wu, 2003; Art & Business, 2004; Barendregt et al., 2009). CACs have been operationalized, for example, as an important human resource management tool influencing employees’ social interactions but also positively affecting their overall performance by stimulating creative thinking (e.g. Martorella, 1990; Barry & Meisiek, 2010). More importantly, however, corporate collecting activities are also increasingly seen as

(12)

a crucial part of internal strategic management and communication practices. One of the most prominent contributions has been a research by Kottasz et al. (2007) who argues that the act of corporate collecting - establishing an image of an art-loving and ‘cultured’ corporation - plays a vital role in signifying a corporate identity and core organizational values to firms’ employees.

2.2. Corporate Art Collecting and Employees’ Cultural Consumption - Organizational Values Perspective

Research in the field of cultural consumption exhibits a great diversity when it comes to the factors which are expected to influence the phenomenon. Although cultural consumption has been argued to be affected by factors as diverse as one’s gender (e.g. Katz-Gerro, 2006) or the wealth of one’s country of origin (e.g. van Hek & Kraaykamp, 2013), not much has been done so far to understand the influence of organization-level determinants on the individual’s behaviour as a consumer of arts and culture. At the same time, seeing a company as an important socializing agent, it seems interesting and valid to look at the role of a firm and a work environment in determining employee’s consumption behaviours. Due to the lack of relevant theoretical frameworks, this thesis proposes to examine this relationship from the perspective of the organizational values literature.

Organizational values - ‘general values that guide organizational members in their selection, evaluation or behaviour’ (Bourne & Jenkins, 2013; 497) - have been argued to have a major impact on employees’ attitudes as well as actions. The literature in the field has found evidence in support of the argument that organizational and individual values often interact and align. For example, Akaah and Lund (1994) have claimed that when it comes to ethical at-work behaviours, values shared by an organization and its members are of a bigger significance for the choices employees make than their own personal values. Moreover, Gehman, Trevino &

(13)

Garud (2013) have suggested adopting a more ‘performative’ perspective on corporate values by arguing that firms’ internal codes of conduct often result from complex processes of associations and interactions between employees’ private and shared corporate values.

Although most of the literature in the field has focused on the influence of corporate values on employees’ at-work behaviour, some researchers have attempted to adopt a wider view on the topic, including analysis of the effects of such values on employees’ actions outside of their immediate work environment. For example, Cambra-Fierro, Polo-Redondo & Wilson (2008) have studied the influence of organisational values on employees’ personal buying behaviours. The research has revealed that the majority of study participants internalized the organization’s environmental and ethical concerns to a significant extent - which reflected itself in loyalty to reliable environment friendly brands and suppliers in employees’ private (outside of work) purchases. An important insight for understanding the broader implications of organizational values can also be drawn from research conducted by Meglino & Ravlin (1998). The authors argue that the strength and extent of the influence of values shared by an organization will depend on the character of a behaviour and situation itself. The more public - open for social scrutiny - the situation is, the more likely an individual will be to adopt collective (and thus socially validated) values, instead of personal ones. In a similar manner, Hemingway & Maclagan (2004) have suggested that a firm often operates as an important socializing referent for individual’s choices and actions. Organizational values are likely to interact with individual’s attitudes and behaviours by providing an opportunity for employees to align themselves with a particular social category - in this case, an organizational membership (Ashforth & Mael, 1989).

The abovementioned theoretical perspective(s) can be seen as applicable to the studied relationship in multiple ways. They provide a framework for understanding how and why organization-level determinants have the potential to influence one’s choices and activities

(14)

outside of a work environment - with cultural consumption being an example of such behaviour. By highlighting the importance of the socializing agent role of a corporation, they point out to the likely process of internalisation of organizational values, such as those related to a firm’s appreciation of arts and culture (reflected in its collecting activities). Moreover, cultural consumption involves behaviours which are of a high symbolic and social nature (Throsby, 2001; Peltoniemi, 2014). Following Meglino & Ravlin’s (1998) and Hemingway & Maclagan’s (2004) reasoning, it can then be inferred that employees will be even more inclined to internalize organizational values in this particular consumption context and setting. Finally, an organization, which clearly states its commitment to the support of arts and culture, is likely to create an environment in which employees feel encouraged to express their affiliation with arts but also actively advance their interest in the topic. With an ‘institutional’ support and encouragement, employees from such firms might be expected to actively engage themselves into higher levels of cultural consumption.

Combining the discourses put forward by the above two sections, this thesis would like to argue that employees from companies involved in corporate art collecting will be more culturally active than employees from firms not engaged in such practices. It is being implied that organizational signals communicated by CACs - highlighting, above all, the art-loving character of a corporation and the value which a company attributes to arts and culture - are likely to be internalised and acted upon by firm’s employees and lead to higher frequency in their cultural consumption. Hence, the following hypothesis is proposed:

H1: Corporate art collecting has a positive influence on the frequency of employees’ cultural

consumption. Employees’ from companies with CACs will be more frequent cultural consumers that employees from companies without CACs.

(15)

2.3. Differentiating Effect of CACs’ Profiles

From the corporate collecting literature, it can be inferred that the effect of a firm’s involvement in collecting activities on employees’ cultural consumption will depend on a CAC’s profile. The available discourses and research - although very limited in number and mostly theoretical - hints upon differentiating qualities of collections’ characteristics which are expected to influence the strength of their signalling qualities. CACs’ specific features are argued to determine, for example, a collections’ position, structure as well as their overall dynamics - from administration models and exhibition conditions to type of collected works.

It can be implied that the signalling qualities of a CAC will be influenced by, for example, its size - especially in relation to the way it is being managed and presented (Behnke, 2007). Wu (2003) argues that bigger collections - signifiers of economic vitality of a company - will often be made more accessible and visible than smaller ones. Companies are more likely to make an effort to manifest the presence of such collections through well-designed and facilitating exhibition spaces in their everyday work environment. Importantly, the form in which the collection’s artworks are being shown as well as their accessibility can be expected to have a significant influence on the way they are interacted with and perceived by employees and all other publics (Martorella, 1990).

The size of a budget, which a company allocates to the maintenance and development of a collection, can be another important factor affecting CACs’ signalling qualities. As a form of internal legitimization, a collection’s budget indicates organizational commitment to and interest in collecting activities, and thus can be expected to influence the strength of the signal a particular CAC communicates (Barendregt et al., 2009). Financial support is likely to be reflected in quality and prestige of collected works but also in overall structure and diversity of company’s art programme - including its promotional efforts (collaborations, publications etc.)

(16)

or, for example, accompanying events (workshops, guided tours, guest lecturers etc.) (Martorella, 1990).

Finally, signalling qualities of a CAC might also be determined by collection’s date of establishment. Although the number of years of existence of a CAC is definitely not an indication of, for example, its quality, it is still an important signal about a status and position which the collected art works hold in an organization (Barendregt et al., 2009). A long tradition of art collecting indicates organizational commitment and devotion to the programme - especially that such ‘superfluous’ non-business activities are often dropped first when the times get more challenging or difficult (Wu, 2003; Lindenberg & Oosterlinck, 2011).

Following these arguments, it seems logical to reason that the effect of corporate collecting on employees’ behaviours as consumers of arts and culture will differ depending on CAC’s characteristics. However, since the abovementioned dependencies and effects are derived solely from theoretical discourses, untested assumptions and individual case studies (limitations highlighted by Barendregt et al. (2009)), they provide a rather weak background for the development of more specific hypotheses. As we can only assume that certain characteristics (independently and / or in interaction) will affect the relationship in question, individual profiles of studied collections will be developed and treated as specific cases. This will be done instead of hypothesising around particular dependencies and interactions of identified features (age, size etc.). This part of the analysis should be then seen as rather explorative in nature. Keeping that in mind, inferring from the existing literature, which suggests differentiating qualities of CACs’ characteristics on their signalling effects, the following hypothesis is suggested:

H2: The effect of corporate art collecting on the frequency of employees’ cultural consumption

(17)

2.4. Moderating Effect of Corporate Art Collecting

Occupational Status

Arts and cultural goods have a long tradition of marking social distinction. It has been argued that cultural preferences as well as consumption choices - a key form of stratification of post-modern styles of living - are very much associated with and conditioned by an individual’s social positioning, especially one’s status (Weber, 1968; Alderson, Junisbai & Heacock, 2007; Lizardo, 2008; Roose, van Eijck & Lievens, 2012). Although researchers have used multiple proxies of social status, occupation has been repeatedly highlighted as one of the most salient variables when it comes to situating individuals in the modern society (e.g. Peterson & Simkus, 1992; Kwon & Kwon, 2013). Job-based status categorization originates from the social ‘honour’ or symbolic position of an occupation and derives its order from ‘relations of perceived, and in some degree accepted, social superiority, equality and inferiority among individuals’ (Chan & Goldthorpe, 2007a; 514). Occupational status has been found to be successful at not only grouping people with similar job conditions but also economic and cultural capital - and thus to operate as a potentially meaningful determinant of one’s behaviour as cultural consumer (e.g. Kraaykamp & van Eijck, 2010). Most of the literature in the field has focused so far on the so-called ‘homology argument’ which suggests a direct relationship between one’s occupational status (low - high) and attendance to certain forms of culture (lowbrow - highbrow, respectively) (e.g. Weber, 1968). An important advancement to the above mentioned view has been, however, the concept of ‘omnivore’ introduced by Peterson and his colleagues (1992; 1996). The hypothesis suggests that, in contrast to the traditional perspective, individuals in the upper end of the social stratum, when compared to their lower status counterparts, have been discovered to not only enjoy more of highbrow categories but, essentially, more of activities from the whole spectrum of cultural offerings

(18)

(Holbrook, Weiss & Habich, 2002; Peterson, 2005). Such findings provide a new perspective on the relationship between social stratification and cultural consumption and argue that, as we move up the social status ladder, we can expect to find people defined by more frequent consumption of diverse forms of art and culture.

Looking back at the studied context, the literature points out several ways in which corporate art collecting can be expected to moderate the relationship between one’s occupational status and cultural consumption. The previously discussed corporate values literature argues that the influence of values shared by an organization on employee’s behaviours will be very much dependent on the position one holds in an organization (Meglino & Ravlin, 1998; Argandona, 2003). It is suggested that, as we move upwards an organizational occupation-based status ladder, employees will be found to be more responsive but also more likely to internalise values of the organizations they work for (Bourne & Jenkins, 2013). A similar relationship can be expected to take place at the level of signals communicated by a CAC - including that of an organizational support for and appreciation of arts and culture. It seems logical then to imply that corporate collecting activities - a reflection of certain organizational values - will be likely to strengthen the relationship in question at all levels of organizational hierarchy, with the effect becoming stronger as we move upwards. Additionally, the relationship between organizational art collecting, one’s occupational status and cultural consumption can also be contextualized within the theoretical framework of the upper echelons (Hambrick & Mason, 1984; Hambrick, 2007). The essential premise of the upper echelons theory is that, in order to understand why an organization engages in certain activities, we should look at the profiles of their most powerful actors (Bhansing, Leenders & Wijnberg, 2012). Organization’s strategies and performance are contextualized as a reflection of its decision-makers’ own cognitions, perceptions as well as values (Jensen & Zajac, 2004; Carpenter, Geletkanycz & Sanders, 2004). Interestingly, the corporate collecting literature has

(19)

widely discussed the role of managerial elites in firm’s art collecting efforts. Managerial interest in corporate collecting activities is repeatedly highlighted as being of a pivotal role for both initiation as well as lasting success of firm’s art programmes (e.g. Kottasz et al., 2007; Barendregt et al., 2009). Most importantly, however, collecting is often seen as a reflection of management’s own values and appreciation of art, with the most personalised collections revealing leaders’ specific tastes and interests (Martorella, 1990). Wu (2003) extends this argument by suggesting that managerial elites - a group traditionally dominated by socially prominent individuals - often proactively use corporate art programmes as an extension of their personal efforts to maintain their dominant social position and distinction. It can then be expected that employees from companies involved in art collecting at positions defined by high occupational status will already exhibit a greater interest in consumption of culture than, for example, their counterparts from firms without CACs. Taking the above arguments together, the following hypothesis is proposed:

H3a: Corporate art collecting moderates the relationship between employees’ occupational

status and the frequency of their cultural consumption. If a company has a CAC, employees’ occupational status will be more positively related to the frequency of their cultural consumption than when a company does not have one.

Educational Attainment

Individual cultural resources are another variable recognized in the literature as being key for social stratification of attitudes and behaviours, such as cultural consumption (e.g. Bourdieu, 1984; Yaish & Katz-Gerro, 2010). An achieved schooling level has been repeatedly highlighted as one of the most accurate proxies of such resources, having a significant influence on individual’s behaviours in the realm of arts and culture (van Eijck, 1997; DiMaggio & Mukhtar, 2004). A level of education is expected to determine individual’s

(20)

propensity - or ‘capacity’ - to both enjoy as well as consume certain cultural offerings and goods by determining one’s previous exposure to arts as well as the ability to interact with its diverse forms (Kraaykamp & van Eijck, 2010). Within the traditional perspective, the relationship has been mostly studied in association to highbrow culture - with interest in more ‘legitimate’ offering expected to increase as one moves upwards on the education ladder (e.g. Bourdieu, 1984). Schooling attainment has been hypothesised as being linked to increased highbrow consumption along the lines of, for example, information processing capacity and complex reasoning which high-end cultural offerings often require (e.g. DiMaggio & Mukhtar, 2004). However, this view has also been challenged by the previously discussed omnivore argument. Van Eijck (2001) has reported a decreasing correlation between high educational levels and highbrow cultural consumption and pointed out to more omnivorous tendencies found in the relationship. Additionally, Chan & Goldthorpe (2007b) have argued that the level of educational qualifications has a significantly positive effect on one’s probability of being an omnivore. In a similar manner to what has been argued in the previous section regarding one’s occupation status, it can be expected therefore that individuals with higher educational levels will consume, on average, more of the whole spectrum of cultural offerings than individuals with lower schooling qualifications.

It is interesting, however, to investigate if the studied relationship is in any way affected by a firm’s engagement in art collecting. It seems valid to argue that employees from companies involved in such activities will be given additional opportunities for advancing their cultural resources (determined initially by their highest education level achieved), when compared to employees from firms without CACs. First, in the most basic perspective, by being exposed to signalling qualities of a CAC, employees are more likely to increase their awareness of value of culture and arts, as communicated by a company (e.g. Kottasz et al., 2007). Similarly to the arguments put forward in the first sub-section of this chapter, this might

(21)

provide them with additional encouragement for higher levels of cultural consumption, despite their initial propensity in this area determined by the educational level achieved. More importantly, however, corporate collecting efforts will often be accompanied by special programmes and informative initiatives contextualizing firms’ collecting efforts. This might include very simple practices such as labels under the artworks or newsletters issued by a company but also more advanced schemes such as workshops, guided tours or lectures (e.g. Martorella, 1990; Wu, 2003). It can be implied that by being given access to such resources as well as actual artworks on a daily basis, employees will develop certain attitudes and ways of interacting with arts and culture - a process likely to affect their behaviour as cultural consumers (e.g. Barendregt et al., 2009; Smiraglia, 2014). By expanding their knowledge on the topic and abilities to respond to diverse forms of creative expression (usually achieved through institutionalised schooling efforts), employees can be expected to become more aware and variety-seeking but also more frequent consumers of cultural goods and activities (Art & Business, 2004; Barry & Meisiek, 2010). Hence, taken together, the above arguments lead to the development of the following hypothesis:

H3b: Corporate art collecting moderates the relationship between employees’ educational

attainment and the frequency of their cultural consumption. If a company has a CAC, employees’ educational attainment will be more positively related to the frequency of their cultural consumption than when a firm does not have one.

(22)

3. Data and Methods

3.1. Research Setting

The data collection for this study was conducted among a group of firms based in Poland. Importantly, the literature has repeatedly highlighted the significance of understanding the impact of a national context on a research process as well as its results (e.g. Katz-Gerro, 2004; UNESCO Institute for Statistics, 2009). For this particular study, the country-specific setting had, above all, significant implications for contextualizing the main research construct - corporate art collecting. The national context had an influence on some other aspects of the study as well - however, these will be addressed in specific sections below.

Although the beginnings of some of the most famous corporate collections around the world date back already to the 1950s and 1960s (Martorella, 1990; Wu, 2003), it seems like the trend has not developed at the same pace in Poland. The oldest collection from the identified Polish CACs has been founded in 2000, which indicates that the phenomenon in the country is still very much at its early stage of development (Rogozińska, 2010). The novelty and the rather limited scope of CACs in Poland had significant effects on all aspects of the research - from the size of the population to, for example, characteristics of the studied collections (such as size) - and thus should be taken into consideration while evaluating both the methods used as well as the eventual results.

3.2. Sampling procedures

The previously mentioned definition of a CAC by Martorella (1990) was employed to guide the initial identification of the study’s population. This was done, above all, to distinguish the actual corporate art collecting efforts from casual purchases of art made by companies for purely decorative purposes. Additionally, the firms with CACs taken under consideration for

(23)

this research were expected to hire at least one art professional (e.g. curator, art market specialist) and have a separate budget designed specifically for art purchases and maintenance of the works. Such operationalization reflects the most common perspective on corporate art collecting, as defined in the relevant literature (e.g. Wu, 2003; Kottasz et al. 2007).

In contrast to countries such as, for example, the Netherlands, where corporate art collecting is already a prominent trend and thus national directories of local CACs are available (e.g. the Netherlands Association of Corporate Art Collections index), Poland does not have such a listing or a reference institution yet. As a result, online research looking at companies’ websites, available academic publications as well as media coverage had to be applied to define the population in question. In the end, a group of 10 firms - with CACs satisfying the previously described conditions - was recognized. Importantly, the number of identified cases was significantly smaller than that reported in similar studies in other countries (e.g. Barendregt et al., 2009; Lindenberg & Oosterlinck, 2011). The initial list was further consulted with an art market professional - who specialises in advising Polish individual as well as corporate and institutional collectors - which, however, resulted in no further additions. All of the identified companies were contacted by email and / or phone, from which 5 replied and 3 agreed to participate in the project.

Table 1. presents an overview of the CACs included in the study. The participating firms were representatives of three industries - financial, real estate and pharmaceutical / cosmetics. All three sectors have been identified in previous research as leading in the trend of corporate art collecting - with financial institutions being often mentioned as pioneers in such practices (Martorella, 1990; Wu, 2003; Kottasz et al., 2007; Lindenberg & Oosterlinck, 2011). The studied CACs were relatively young - with the oldest one founded in 2000 and the newest one in 2012. In terms of their size (number of artworks), the collections were smaller than in the case of CACs studied in previous research (e.g. Martorella (1990) - an average American

(24)

CAC in the sample had around 600 artworks) - and ranged from 50 to 250 pieces. Importantly, only administrative employees working in firms’ headquarters (and thus excluding, for example, production staff) were involved in the study. This was done to address the staff which actually had an access to the collected and exhibited artworks and interacted with them on everyday basis.

Table 1. The profiles of the corporate art collections (CACs) owned by the firms participating

in the study, including firms’ industry as well as collections’ founding date and size (number of artworks).

Firm’s industry CAC’s founding date CAC’s size (number of artworks

as in June 2015)

CAC 1. financial 2000 160

CAC 2. pharmaceutical / cosmetics 2012 250

CAC 3. real estate 2011 50

Since, initially, the research was supposed to be based on the twin study design, the next round of sampling included identifying possible ‘sibling’ organizations of the pre-selected ones. The twin model allows for controlling a set of subjects’ shared characteristics and thus guides their comparison at the level of effects of a differentiating phenomenon, such as corporate art collecting (see, for example, Seifert, Morris & Bartkus, 2003). The profiles of pre-selected companies were developed according to their industry, annual turnover for 2014 as well as size, operationalized as a number of employees. A purposive sampling approach was employed, with 4-5 twin firms identified for each of the participating companies. Again, all of them were contacted via email and / or phone. However, despite multiple attempts and efforts, only two of them expressed willingness to participate in the study - twins of the real estate and

(25)

pharmaceuticals / cosmetics firms. The failure to identify the third sibling company did not allow for an actual twin model to be applied in the research. Despite this, the sibling companies were still used to develop a comparative data set applied to hypotheses testing. In the end, a group of 1,880 potential participants (from both groups of firms) were identified. Appendix 1.1 provides a summary of the information about all participating companies.

3.3. Data collection

Most of the study’s primary data was collected using an online questionnaire including a set of closed-end questions (see Appendix 2). The questionnaire was distributed online, in the form of a link, using companies’ internal networks. As the response rate for online surveys tends to be quite low (van Selm & Jankowski, 2006), employees were offered a chance to join a small lottery with rewards to encourage participation. All companies sent two reminders regarding the questionnaire, at weekly intervals. Although the same methods and incentives were used for all participating firms, employees’ response rates per company differed quite significantly and ranged from 6% to 80% (the exact percentages can be found in the last column of Appendix 1.1). In total, the system registered 494 entries. However, 8 cases were identified as invalid - participants entered the questionnaire but did not attempt to fill it in before the link was deactivated. The final sample comprised of 486 respondents from 5 firms.

Specific data about the participating firms and the collections was gathered from companies’ websites and publications (ex. collections’ catalogues) but also during e-mail consultations, Skype calls and face-to-face meeting with firms’ representatives and CACs’ curators. This additional research was conducted to gain a better understanding of the research setting, develop the previously mentioned profiles of the participating companies and identify key characteristics of the selected CACs - including their size and founding year (for reference, see again Table 1.). Unfortunately, the companies were not willing to share information about

(26)

the size of the budgets assigned for their collecting activities and thus the variable had to be excluded from the analysis.

3.4. Questionnaire

Translation

The initial questionnaire was written in English (as the items were developed according to English-speaking academic literature) and thus it had to be translated to Polish. This was done to avoid possible problems caused by the language barrier and to make the questionnaire accessible to as many employees as possible. The survey was translated in parallel by the researcher herself and an independent translator - a PhD student of English linguistics, who is also a native Polish speaker. The two versions of the questionnaire were compared and evaluated in terms of their consistency. The final version of the translated document was also reviewed by a professional Polish language editor in order to confirm the clarity of translated constructs and the questionnaire’s overall grammatical correctness. Both English and Polish versions of the questionnaire can be found in the Appendix section (Appendix 2.1 - English version, Appendix 2.2 - Polish version).

Variables and Measurements

The questionnaire was created on the basis of the review of existing research and theoretical discourses, which guided the development of measures relevant to the studied relationships. Its 38 questions were organized in 3 sections - with some questions included exclusively in the questionnaire sent to the employees from companies with CACs (marked with * in the Appendix copy). Importantly, only a part of the collected data - specifically, from the first two sections of the questionnaire - was used in the analyses conducted for this particular thesis. As a result, only the sections relevant for this research are presented in Appendix 2. and discussed

(27)

below.

Employees’ Cultural Consumption

The main hypotheses of the thesis addresses the relationship between the effects of corporate art collecting (independent variable) on employees’ cultural consumption (dependent variable). Employees’ cultural consumption was studied using measures adopted from the 12-item scale originally developed by van Eijck & Knulst (2005). The authors divided the studied categories of cultural offering into highbrow and lowbrow sub-groups, which reported Cronbach’s alphas of 0.74 and 0.51, respectively. Although rather low, the reliability scores of the selected items reflect a broader problem which the writer came across while searching for the most adequate scales. The research process revealed that the majority of them fall below or just at the commonly accepted threshold of 0.7 (especially the lowbrow categories). Rimmer (2012) argues that the reason behind such a limitation might be the permanently changing character and symbolic meaning of cultural offering but also blurring of the boundaries between the traditionally clear-cut categories.

The items identified by van Eijck & Knulst (2005) include a diverse set of activities in the sphere of visual arts, performance, music, cinema and entertainment across the lowbrow-highbrow spectrum. Such selection of items deals with one of the key limitations of research in the field, which often disregards non-highbrow activities as a part of cultural consumption behaviour (Katz-Gerro & Jægar, 2011). Moreover, it acknowledges the importance of studying cultural consumption in its totality (Olivier, 2008) and allows to address the abovementioned increasingly omnivorous tendencies in cultural consumption behaviours in the modern society (Peterson, 2005; Roose, van Eijck & Lievens, 2012).

The items were reviewed regarding their relevance to the studied setting and other research in the field (e.g. Alderson, Junisbai & Heacock, 2007; Yaish & Katz-Gerro, 2010). As

(28)

a result, small changes to the original scale were made. Two categories were merged into one question (e.g. classical music performance and opera) and one category was expanded by adding an option of watching a film on DVD or online (and not only on TV or on a videotape) to reflect the most recent changes in the area. The final version included then 11 items.

Additionally, van Eijck & Knulst’s original ‘yes-no’ response scale was replaced by a 5-item Likert frequency scale (including options from ‘never’ to ‘very frequently’). The alternation was made to allow for measuring the actual frequency of employees’ attendance to and consumption of identified categories. The studied time scale of employees’ behaviours was, however, kept the same and referred to the frequency of consumption in the 12 months. The items related to the discussed variable can be found in Q.16 – Q.26 of the questionnaire (see Appendix 2).

Employees’ Demographic and Socio-economic Profile

In order to develop a better understanding of the profile of the studied sample but also to address the last two hypotheses, participants were also asked to answer a set of demographic and socio-economic questions. Although in the majority of the cases, measurement of such variables was self-explanatory (e.g. questions regarding one’s age or gender), some of them were more complex and thus will be discussed briefly below.

Taking into consideration the national context of the conducted research, the variable asking about participants’ income was developed based on publications of the Polish Central Statistical Office (Central Statistical Office, 2014b). It resulted in 4 household per capita income categories (Q.9 of the questionnaire). In response to the requirements placed by the participating firms, an additional option - ‘do not wish to share this information’ - was included in the answer list.

(29)

and used to study the relationship addressed in Hypothesis 3a. The 10 job categories used in the item (see Q.10 of the questionnaire) were created and ordered on the basis of 31 occupational status groups developed by Chan & Goldthorpe (2004; 2007a). The applicability of these categories for the national setting in question was again verified against the measures used in Central Statistical Office’s reports and in consultation with the HR departments of participating companies. This was done to ensure that the item included all employees from specific internal hierarchies and structures. Finally, to address Hypothesis 3b looking at employees’ educational attainment a question about participants’ highest completed schooling level was asked (Q.5 of the questionnaire). The review of a relevant CSO report resulted in the development of a list of 7 key subsequent educational levels (Central Statistical Office, 2014a).

Other Variables

The questionnaire also included a set of additional questions exploring employees’ active involvement in arts and their awareness as well as attitudes towards firms’ CACs. The items were developed as a result of the themes identified in the literature but also issues and concerns reappearing in conversations with collections’ curators. Q.14 of the questionnaire asked employees if they practiced any art-related hobbies and Q.27 if they collect any art themselves. The participants were also asked if they were aware of their firm’s involvement in corporate art collecting practices (Q.29 of the questionnaire), if there were any artworks exhibited in their direct work environment (Q.30) - and if yes, what their attitude towards them was (Q.31). All of the above mentioned questions were addressed in a straightforward ‘yes-no’ manner, apart from Q.31 which used 5-item Likert scale (‘strongly like’ - ‘strongly dislike’) as response options.

(30)

4. Results

4.1. Pre-tests Data Treatments

In order to prepare the dataset for the following analyses, all studied variables were checked for missing data by running frequencies tests. However, no missing values in the relevant cases were identified. Additionally, a scale mean for employees’ cultural consumption was computed to facilitate the hypotheses testing. The new variable was tested for its reliability to identify how successful the items were at measuring the constructs they were intended to address. The 11 items (including both low as well highbrow offering) reported satisfactory level of internal consistency, scoring Cronbach’s alpha of 0.74 so above the required threshold of 0.70.

Appendix 1.2 provides a table with an overview of the relationships between all studied variables - including the means, standard deviations, inter-correlations of the studied categories and the scale’s reliability coefficient.

4.2. Descriptives

The initial dataset of 486 respondents included two groups of participants: employees from companies with corporate art collections (N = 357) and those from companies without (N = 129). In the survey sent to the first group an additional question was asked about employees’ awareness of their firms’ collecting practices (Q.29 of the questionnaire, see Appendix 2). Interestingly, 137 participants (over 38% of the sub-sample) were found to be unconscious of their firm’s involvement in corporate art collecting. Considering the implications of such a finding for the theoretical arguments put forward by this thesis - based essentially on the interaction between signalling qualities of CACs and employees, as their conscious receivers - a decision was made to exclude those cases from the following analyses. Although dismissing the aforementioned group of participants significantly reduced the sample size - down to 349

(31)

cases - it was actually found to improve the overall correlation of studied variables and performance of the data.

The total sample consisted of 204 females and 145 males, which represented 58.5% and 41.5% of the sample, respectively. The sub-sample of employees from companies with CACs was found to be more female-dominated - with 144 females (65.5%) and 76 males (34.5%) (see Appendix 1.3). The average age of employees oscillated around 36 years old, which hinted upon a rather young character of the studied group. However, the majority of the participants (over 81% of the total sample) were actually found within an older age category, 41-60 years old, with similar tendencies found in both sub-groups of the samples.

The total sample was dominated by highly educated individuals, with 316 employees (90.5% of the sample) holding a university degree, although the size of this group was slightly smaller in case of the participants from companies without CACs - 114 employees (88.4% of the sub-sample) (see Appendix 1.4). The homogeneity of the educational profile of the participants could have been related to the type of studied companies and the fact that the research included only administrative workers. It also reflects an important educational trend in Polish society, with an increasing majority of people holding (at least one) university degree (Central Statistical Office, 2014a). The main three occupational groups included employees working at specialist and expert positions (150 participants - 43% of the sample), middle- and low-level managers (70 participants - 20.1% of the sample) and sales / customer service staff (46 participants - 13.2% of the sample) - all three categories being classified as of average to high occupational status (according to Chan & Goldthrope’s (2007a) schema). Similar distribution of frequencies was found in the sub-group of employees from companies with CACs. The second sub-sample, however, included only 4 sale / customer service staff (3.1% of the group). Appendix 1.5 provides a detailed overview of the descriptives for all occupational categories in both the total sample and its sub-groups. Finally, 87 participants (24.9% of the

(32)

sample) fell into the middle per capita household income category (2001-3000 PLN), which corresponds with the national average (Central Statistical Office, 2014b), and 75 (21.5% of the sample) into the highest one - more than 4000 PLN (see Appendix 1.6). It is important to highlight, however, that 82 participants (23.5% of the sample) refused to share information on this question.

As for the main studied construct - cultural consumption - employees exhibited a rather low level of activity across the identified cultural offerings. The average result on the cultural consumption scale measured against 11 items (with scores ranging from 1 to 5) was found to be 2.58 (SD = .56), which falls between ‘rarely’ and ‘sometimes’ categories of the frequency of attendance in the last 12 months (Table 2.). Specific descriptives revealed differences in mean frequencies of yearly attendance between different cultural offerings. Individual items’ means ranged from 1.31 (ballet performance) to 4.18 (watching a movie on TV, DVD, videotape or online). The items representing lowbrow cultural consumption offerings, such as previously mentioned TV watching or going to movies, reported the highest mean scores.

Additionally, 185 respondents (53% of the sample) defined themselves as people interested in arts and culture, 120 (34.4% of the sample) were found to be actively participating in arts (having an art-related hobby, playing an instrument, being a part of a dance group etc.) and 70 (20.1% of the sample) to be collecting art (paintings, photography, sculptures etc.). The scores followed a similar pattern for employees from companies involved in collecting practices. However, they were distributed slightly differently for the second sub-group, with only 37 employees (28.7% of the sub-group) having an art-related hobby and 19 (14% of the sub-group) collecting art by themselves. An overview of the descriptives discussed above can be found in Appendix 1.7.

Finally, the results also showed that 158 of employees from companies with CACs (71.8% of the sub-sample) had artworks exhibited in or nearby their office. However, the group

(33)

displayed rather negative attitudes towards the pieces with 32 employees strongly disliking and 80 disliking the artworks exhibited in their workspace (together over 70% of the group). The summary of these frequencies can be found in Appendix 1.8 and 1.9.

Table 2. The means (M) and standard deviations (SD) of the frequency of employees’ cultural

consumption in the last 12 months - in total and per scale item.

M SD

Visiting a museum 2.54 .96

Visiting an art gallery and / or attended an art exhibition

2.19 .99

Going to a classical music and / or opera performance

1.98 1.08

Visiting a historical and / or monumental architecture site

3.32 1.00

Going to a professional theatrical play 2.56 1.15

Going to a ballet performance 1.31 .66

Going to a popular music concert (rock, pop, hip-hop music etc.)

2.49 1.22

Watching a film in a movie theatre 3.38 1.08

Going to a live sport event (e.g. football match) 2.47 1.37

Going to an amusement park 1.98 .99

Watching a movie on TV, DVD, videotape and / or online

4.18 .93

(34)

4.3. Hypotheses Testing

Corporate Art Collecting and Employees’ Cultural Consumption

The first of the conducted analyses set to explore a hypothesized difference in the frequency of cultural consumption in the last 12 months between employees from companies involved in corporate art collecting (N = 220) and those from companies which do not collect art (N = 129) (Hypothesis 1.). The independent-samples t-test, looking into the significance of a difference in the means of two independent groups on a continuous variable - in this case, cultural consumption - was identified as most appropriate model for studying this relationship. Although the number of cases in the two studied sub-groups was not equal, all other key assumptions of the test were satisfied: there were no outliers in the data (as assessed by inspection of a boxplot), employee’s cultural consumption scores were normally distributed and fulfilled the requirement of homogeneity of variance.

The t-test analysis pointed to a slight difference in the sub-group means, as scored on the cultural consumption scale (Table 3.). Indeed, employees from companies involved in corporate collecting were found to be, on average, more frequent consumers of art and culture (M = 2.59, SD = .56) than those from companies without CACs (M = 2.56, SD = .55). However, the identified difference in the mean cultural consumption scores between the two groups was not found to be statistically significant, t(347) = .51, p = .61. Thus, Hypothesis 1 was rejected.

Additionally, although not directly addressing the main hypothesis, independent-samples t-test analyses were also run for all individual cultural consumption items. This was done in order to identify if there were any meaningful differences in the mean consumption scores for specific cultural offerings. The t-test procedures were repeated for all 11 items used to assess employees’ cultural consumption in the last 12 months. The results of the tests can be

(35)

reviewed in Appendix 1.10. The analyses revealed that employees from companies with CACs scored, on average, higher in the majority (5 out of 6) of categories referring to highbrow cultural consumption, i.e. visiting museum, M = 2.61 (SD = 1.00), and art galleries, M = 2.27 (SD = 1.04), or seeing a play in a theatre, M = 2.61 (SD = 1.15). On the other hand, employees from companies without CACs were found to be more frequent consumers of the lowbrow / popular categories (4 out of 5), such as attending a popular music concert, M = 2.51 (SD = 1.23), or watching a movie on TV, DVD, videotape or / and online, M = 4.22 (SD = 0.88). The last category was also the most frequently consumed item of the whole scale. However, none of the identified differences was found to be statistically significant and thus no specific conclusions could be drawn from the conducted analyses.

Table 3. The t-test results for the effect of firm’s involvement in collecting activities on the

frequency of employees’ cultural consumption in the last 12 months. Employees’ count per sub-sample (N) as well as the means (M), standard deviations (SD), t-value (t), degrees of freedom (df) and test’s significance (p) scores are provided.

T-test

N M SD t df p

Employees from companies

with CACs 220 2.59 .56 .51 347 .61*

Employees from companies without CACs

129 2.56 .55

(36)

Differentiating Effect of CACs’ profiles

The second of the stated hypotheses addressed a potential differentiating effect of CACs’ profiles on the studied relationship. The one-way ANOVA test was run on the data sub-set including employees from companies with CACs to determine whether there were any differences in their mean cultural consumption frequency scores, when categorized depending on a specific profile of each of the collections. Again, the data was tested for compliance with the model’s assumptions. All of the requirements, apart from the equal number of cases in the studied categories, were satisfied. Data was normally distributed and there was homogeneity of variances. The analysis of boxplots identified one outlier - however, considering the size of the sample, the case was kept in the dataset.

Data entries were classified against three categories, representing the studied CACs: CAC1 (N = 169), CAC2 (N = 26) and CAC3 (N = 25). The profile of each CAC was plotted on the basis of the previously described characteristics, such as when the collection was founded and how big it was (for details, refer back to Table 1, p. 24). The employees from the company with the smallest of the collections (CAC3) were found to be most culturally active, M = 2.75 (SD = 0.53) and those from the company with the middle size collection (CAC1) the least, M = 2.56 (SD = 0.58), with the employees from the firm with the biggest one (CAC2) oscillating in between the two, M = 2.68 (SD = 0.43) (see Table 4). Accordingly, when looking from the perspective of collections’ years of existence, the analysis revealed similarly surprising results - with the employees from a company with middle-age collection (CAC3) scoring the highest in the group, from the youngest one (CAC2) being the average and from the oldest collection (CAC1) being the least active. The identified differences between the collections, however, were found to be statistically insignificant, F(2, 217) = 1.74, p = .18, and thus did not allow to proceed with any post hoc or contrasts procedures. The conducted tests showed no support for the argument put forward in Hypothesis 2.

(37)

Table 4. The one-way ANOVA results for variance in the frequency of employees’ cultural

consumption in the last 12 months per collection. Employees’ count per group (N), means (M) and standard deviations (SD) as well as sum of squares (SS), degrees of freedom (df), mean squares (MS), F-value (F) and test’s significance (p) are provided.

N M SD CAC1 169 2.56 .57 CAC2 26 2.68 .43 CAC3 25 2.74 .43 One-way ANOVA SS df MS F p Between groups 1.07 2 .54 1.74 .18* Within groups 67.60 217 .31 Total 68.69 219

(38)

Corporate Art Collecting as a Moderator

A final set of hypotheses addressed potential moderating effects of corporate art collecting on two relationships which emerged after reviewing the literature in the field. The analyses set out to identify if a firm’s involvement in collecting activities moderates the relationships between employees’ occupational status (Hypothesis 3a) as well as employees’ highest educational level achieved (Hypothesis 3b) and the frequency of their cultural consumption. As in the case of both models the independent variables were ordinal - and not continuous, as required in the standard moderating analysis - a two-way ANOVA was identified as the most applicable statistical test to be used. Although an interaction effect studied by two-way ANOVA is essentially symmetrical, the model also allows for one of the independent variable to be theoretical contextualised as a moderator of the focal relationship (Field, 2009).

Data relevant for both analyses was tested for the ANOVA’s assumptions. There was homogeneity of variance and, although normality of data distribution was slightly violated, the tests were still performed (ANOVA is often being perceived as quite ‘robust’ to deviations from normality, especially when the sample size is large enough (Maxwell & Delaney, 2004)). Additionally, because there was an unequal number of cases in the cells of the study design (unbalanced study design), the unweighted marginal means (ignoring the number of participants in each cell) were used to interpret the tests’ results.

Occupational Status

The first of the studied relationships hypothesised that the effect of employee’s occupational status on their frequency of cultural consumption will be different (more positive) for employees from firms with CACs than for employees from firms without CACs - in other words, that corporate collecting will positively moderate the relationship in question. Table 5 provides an overview of the results of the conducted analyses. First, the main effect test for

Organization-level determinants of employees’ cultural consumption : a study of internal signalling qualities of corporate art collecting