Quality Signals in Uncertain Markets:
Prestigious Graduate Program Selection Criteria versus
Market Standards in Visual Arts
Master Thesis - MSc Business Administration
Entrepreneurship and Management in the Creative Industries Track
Name: Justyna Wiśniewska
Student number: 10826408
Thesis Supervisor: Monika Kackovic
Submission Date: 29
thJanuary 2016
Statement of Originality
This document is written by Student Justyna Wiśniewska 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.
ABSTRACT 4
INTRODUCTION 5
LITERATURE REVIEW 8
CREATIVE INDUSTRIES 8
INFORMATION ASYMMETRY AND MARKET UNCERTAINTY 9
QUALITY SIGNALS 10
SELECTION SYSTEMS 13
STATUS 15
ART MARKET 16
SELECTION SYSTEMS AND DECISION MAKING PROCESSES IN THE ART MARKET 19
SIGNALS AS FACTORS REDUCING UNCERTAINTY IN THE ART MARKET 21
DATA AND METHOD 21
RIJKSAKADEMIE AND ITS SELECTION PROCESS 22
ARTFACTS.NET RANKING 23
DATA COLLECTION 25
VARIABLES 28
RESULTS 31
PRE-TESTS DATA TREATMENT AND DESCRIPTIVES 31
HYPOTHESES TESTING 31
RELATIONSHIP BETWEEN THE ADMISSION CRITERIA AND ACCEPTANCE TO RIJKSAKADEMIE 31
RELATIONSHIP BETWEEN CREATIVITY SCORE AND BEING ACCEPTED INTO RIJKSAKADEMIE 32
RELATIONSHIP BETWEEN CRAFT SCORE AND BEING ACCEPTED INTO RIJKSAKADEMIE 33
RELATIONSHIP BETWEEN COLLABORATION SCORE AND BEING ACCEPTED INTO RIJKSAKADEMIE 34
RELATIONSHIP BETWEEN CRITIQUE SCORE AND BEING ACCEPTED INTO RIJKSAKADEMIE 34
RELATIONSHIP BETWEEN THE ADMISSION CRITERIA AND ATFACTS.NET RANKING SCORE 35
DIFFERENCE BETWEEN ACCEPTANCE RATE TO RIJKSAKADEMIE AND ARTFACTS.NET RANKING SCORE
38
DISCUSSION 40
GENERAL DISCUSSION 40
LIMITATIONS 42
IMPLICATIONS FOR FUTURE RESEARCH 43
CONCLUSION 44
Abstract
In creative industries, the product quality is very difficult to assess due to high market uncertainty. The visual arts market is a particularly good example of a market struggling with the fundamental issue of assigning value to products. This is why signals and selectors play a major role in reducing the uncertainty and facilitating the process of quality valuation. Education institutions are a perfect setting for exploring both signals’ and selectors’ participation in the valuation process. Focus of this paper is to investigate whether graduation programs have a positive relationship with the further careers of their students. In the empirical setting of the visual arts it is tested whether students of a prestigious residency program perform better after graduating than the ones who had been rejected. This analysis focuses on signals sent by both first and third parties in the visual arts and on how they correlate with artists’ performance. A unique dataset of the Rijksakademie applicants, spanning the years from 1990 to 2005 has been used, which covers career trajectories of 355 visual artists. For testing the artists’ performance the Artfacts.net Ranking scores of the rejected and accepted applicants have been compared. We found that the criteria used within the Rijksakademie recruitment process do not necessarily correspond to the criteria defining the success on the market. Moreover, the analysis showed that graduating from the prestigious residency program did not correlate with the increase in performance in the visual arts market, expressed in the Artfacts.net Ranking score.
Introduction
This dissertation aims to contribute to the theoretical understanding of the phenomenon of signaling in relation to selection systems theory and status transfer mechanism in the conditions of high uncertainty. For this purpose education in the visual arts market setting has been chosen, because of the high level of information asymmetry.
Signaling theory is looking into how the producer himself can signal the quality of his goods (first party signals) and how independent third parties give their testimonials about the quality of the good (third party signals). The main, overall goal of signaling is to transfer the information connected with the sender’s quality, reflected upon the level of his or her performance (Graffin & Ward 2010). The better the information is sent, the stronger the signal fit is and the better the necessary audience (selectors) is convinced (Connelly et al., 2011). The study takes place within the visual arts empirical setting, which serves here as an example of a highly uncertain market. Signals act as factors reducing information asymmetry – the disproportionate distribution of information about quality between the producer and the customer (Kirmani & Rao, 2000).
The aim of the following paper is to demonstrate the influence of signals on the overall artistic performance in Visual Arts. Both first party and third party signals are taken into account, with emphasis on the latter. It is also to be studied how well the signals emitted by first and third parties reach the necessary audience – the selectors. Selectors can be defined as entities that influence the behavior of the customers by expressing their opinions and judgment about the goods of interest (Wijnberg, 2014).
In the following thesis I would like to emphasize the exceptionality of education institutions and their selections processes, which are a combination of first party signals (artists choose a school with a goal to signal something) and third party quality signals
(schools communicating the quality of artists to the market). This study expects that graduating from prestigious education program is a signal facilitating the further career development of successful artists. It is surprising to note that though research exploring the effects of education as quality signals transferred to real life job market conditions is quite extensive (see, e.g., Pfeffer and Fong, 2002; Ehrenberg and Mavros, 1995; Groove and Wu, 2007), only few studies have examined the impact of elitist specialized programs, and nothing concerning the elite education in creative industries has been published yet.
In the study, two perspectives are being taken into consideration: the external perspective (how acceptance and rejection to an elite institution affects applicants’ overall future performance) and the internal perspective (studying the multiple criteria used for selection, and verifying how they predict being accepted into the educational program and how they predict overall future success, regardless of acceptance into the program).
The study will be conducted in the empirical setting of the Rijksakademie Residency Program and based on the data collected between 1990 and 2005 during the program entry interviews. It will be examined which selection criteria in the program predict the admission best and how strongly the scores correlate with the future success of the graduates in real life market conditions. For this research, as a measure of artistic success, I will be using the score in international artists’ ranking developed by Artfacts.net – a professional online database of key facts on the art world and special tools for analysis of activities, which together enable gaining insight into the dynamics of the art market.
The following research contributes to the literature mainly by giving a deeper understanding in the two following areas: elitist education programs and their selection criteria reducing uncertainty in the visual arts market, and the mechanism behind this selection process facilitating the prediction of the future performance of the graduates.
Comparing the information about accepted and rejected candidates is of great value to the proposed impact of the residency on artists and their careers. Thus, the main research question that this thesis will attempt to provide answer to is the following: how does the selection made by the Rijksakademie jury correlate with having a better score in Artfacts.net Ranking and which criteria play the crucial role in both accepting the candidates in the Rijksakademie, and increasing their score in the Artfacts.net Ranking?
Literature Review
Creative IndustriesCreative industries have been broadly defined by United Kingdom Department for Culture, Media & Sport as "industries which have their origin in individual creativity, skill and talent and which have a potential for wealth and job creation through the generation and exploitation of intellectual property" (DCMS 2001, p. 4). In line with this general definition, the following domains are currently classified as pertaining to the creative industries: advertising, architecture, art, crafts, design, fashion, film, music, performing arts, publishing, R&D, software, toys and games, TV and radio, and video games (Howkins 2001). According to P. Mœglin (2010), also the education industry can be classified as part of creative industries.
The recent development of the knowledge-based economy, and questions about the dependence between information, knowledge and creativity have led to an attempt of implementing a new market-based definition, taking into consideration how the demand and supply operate in complex social networks of creative industries (Flew, 2002). The definition is therefore being directed more towards innovation systems than as until now – the cultural policy paradigm. A new approach is to classify the social network markets as being the main defining quality of creative industries (Potts et. al, 2008). According to this approach, defining careers in creative industries requires participation of numerous stakeholders and their involvement in a complex social network process facilitating the link between the demand and supply.
The specific demand-supply relationship in creative industries comes from the particular types of goods they handle: the experience and credence goods. Experience goods can be
difficult to observe in advance, but these characteristics can be ascertained upon consumption. Experience goods are usually purchased based upon reputation and recommendation of the experts, since physical use of the good does not facilitate evaluating its quality (Aidan, 1988).
The credence goods (also called post-experience goods) are goods whose value consumers can never be sure of. However, what differentiates them from experience goods is the consumers’ difficulty to assess their quality also after having experienced them. The only way to ascertain their quality is to consult the experts either before or after the consumption (Emons, 1997). To a big extent, the value of credence goods is often a matter of faith or belief. The seller of the good knows the utility impact of the good best, which is creating the discrepancy between the knowledge of the buyer and the seller – the so-called information asymmetry. In some cases, the provider of credence good does not know its quality at all and has to rely on his or her intuition how to approach the process of convincing other stakeholders (Akerlof, 1970).
Information Asymmetry and Market Uncertainty
The notion of information asymmetry (Akerlof 1970) results from the fact that market consists of different parties in a transaction. The parties involved have different information and perceptions about product quality, which creates the discrepancy. In creative industries, the products are highly heterogeneous, thus it is very difficult to determine their quality, based on a simple observation (Dean & Biswas, 2001).
In the case of information asymmetry indices (the observable attributes) and signals (attributes manipulable by the applicant
)
can help to solve the evaluation problems. Using signals as proxies of quality can help fill in the information gaps in pre-purchase or pre-recruitment decision-making process (Spence, 1973).In all markets, entities face problem of receiving incomplete information, which might impede market participants from making decisions. Recent research in information economics focuses on signals as tools of solving the problems caused by the asymmetric information, or information gaps. Kirmani and Rao (2000, p. 66) define first party signals as “actions that parties take to reveal their true types (e.g. skill level) and that can be transmitted in many forms, including brand, name, price, warranty, and advertising expenditures”. The decisions concerning those signals aim in reducing the information asymmetry and convincing the consumers about the quality of the good they want to purchase. Investing in sending the relevant quality signals can decrease significantly the information gaps and the uncertainty by conveying credibility and increasing the probability of purchase (Rao et al., 1999; Johnson and Levin, 1985).
The main source of uncertainty is the difficulty to determine the value of a particular piece of work (Beckert, 2009). It is the case mainly for goods that are complex, inimitable, non-comparable, and have no objective measures (Yogev, 2010). The experience and credence goods are therefore particularly vulnerable to information asymmetry. The buyers are very quality-sensitive and incapable of assessing the quality and the value of the good at the same time. Value is greatly connected to quality, which in principle concerns the product’s ability to fulfill a specific consumer’s need. However, in the case of experience and credence goods, the evaluation of the quality can be described as the attribution of value to a certain good, based on not only the good’s functionality, but also its observable attributes and signals (Wijnberg, 2014).
Quality Signals
Quality within the framework of signaling theory is defined as “the underlying, unobservable ability of the signaler to fulfill the needs or demands of an outsider observing
the signal” (Connelly et al., 2011, p. 43). In other words, the quality is ungraspable and it is assessed based on the combination of first and third party signals (Connelly et al., 2011).
Signaling theory is trying to explain how the producer himself can signal the quality of his goods (first party signals, e.g. warranties) and how it is indicated by independent third party signals (e.g. rewards, reviews). The main, overall goal of signaling is to transfer the information connected with the sender’s quality, reflected upon the level of his or her performance (Graffin & Ward, 2010). By emitting the signals, the quality of the good is elevated and the buyer becomes more prone to choose one good over another. The extent to which the signal is correlated with quality is defined as signal fit, and it decides on how well the sender reaches the relevant audience (Connelly et al., 2011).
Signals originate from different sources and can be divided into first party signals and third party signals. The first party signals come from the producer of the good and it lies directly in his interest to improve the perception of the good he or she produces. Therefore, the signals emitted directly by the producer are often not considered to be credible and strong for the consumer (Dewally & Ederington, 2006). The third party signals originate from the impartial experts, having knowledge in the field and not being related to the producer of the good. Their independence in assessing the quality and rating the products makes them more credible and valuable than the first party signals to the consumers (Dean & Biswas, 2001).
The credibility of the third party signal depends highly on the credibility of the source. Art buyers often rest on the experts’ recommendations when making purchase decisions. The same mechanism takes place when art collectors consult the galleries and auction houses before they purchase any artworks (Emons, 1997). Naturally, higher status sources have a bigger influence on the quality perception than the low status ones. When a top gallery
exhibits an artwork of an unknown artist – it communicates a high potential of the artist and starts elevating his or her status (Annamma & Sherry, 2003).
In the job market the mechanism is similar: potential employees emit signals about their skills to the employer, e.g. by acquiring education testimonials. The informational value of the candidate’s received diploma comes from the fact that the employer believes it is positively correlated with being more capable of learning and performing well in the job. The more prestigious the school is, the more credible the signal is to the employer. Therefore, the education paths of the candidates become the basis for making the selection and distinguishing low ability workers from high ability workers (Spence, 1973).
The employers interpret signals that symbolize the skills and capabilities of the applicant. Since the recruiting entity is unaware of the real skills of the applicant, he or she has to base their judgment on both first party signals (candidate’s performance during the interview and his works) and third party signals (education, reviews, gallery affiliations, auction sales, etc.). This is why, the candidates always strive for acquiring as many quality signals as possible, such as graduating from prestigious schools, winning renowned awards, getting positive reviews and recommendations (Spence, 1973).
Moreover, since the visual arts market handles the experience products that may be affected by information asymmetry, using value signals helps determine the value of an artwork (Kirmani and Rao, 2000). Consequently, the effectiveness of the value signal in conveying product value becomes crucial. For a value signal to be effective it should be specific and accurate.
In this study, it will be examined which criteria fulfilled by the candidates during the interview fit best the criteria of the internal selectors in the Rijksakademie and the external
selectors in the visual arts market. Subsequently, it will be studied how the signal of having been admitted to the prestigious education program correlates with the artist’s performance later on.
Selection Systems
Selection systems theory interprets competitive processes in terms of the ‘selected’ (those who are being chosen), and ‘selectors’ (those who allocate value to the selected ones). It also argues that in order to compete successfully, agent’s resources have to be valued and recognized as valuable (Wijnberg & Gemser, 2000).
One possible way to examine how the value of cultural products is being determined is to do it from the perspective of the selection system theory (Wijnberg & Gemser, 2000). This theory claims that the value of a product is not given, but that it can be determined by decisions, actions and opinions of certain selectors. Therefore, the value of a good is being established within the context of a set of preferences of groups or individuals that can act as selectors (Wijnberg, 1995; Mol, Wijnberg, & Carroll, 2005).
In the creative industries, the role that selectors play in determining the value is generally more visible than in many other industries. The harder it is to estimate the value of individual product characteristics, the more valuable the necessary knowledge and expertise of the selectors is (Hirsch, 2000). However, since as stated before, the customers cannot assess the value of experience and credence goods independently; they very often arrive at estimations of the goods’ value with the help of selectors (Bhansing, 2013).
The selection system theory specifies the necessary characteristics of the selected and the selectors. The selected consist of the ones that are competing with each other for recognition and the selectors are agents whose decisions have an effect on the outcome of
the process (Wijnberg & Gemser, 2000). Therefore, selectors can be defined as “the actors whose judgment plays a part in determining the behavior of the eventual customers in a competitive process” (Bhansing, 2013, p. 92). In other words, the selectors influence the opinions of the consumers and properly addressed, can enhance the quality perception significantly, e.g. a reviewer by his publicly expressed positive opinion or a gallery promoting a particular new artist. Therefore, in order to create and maintain a competitive advantage it is important to create superior value in the eyes of the dominant selectors (Wijnberg & Gemser, 2000).
There are three main types of selectors (Wijnberg and Gemser, 2000):
• Market consumers (the end users are the selectors and the producers are the selected – users are independent in their assessment of resources)
• Peers (the opinion of fellow producers forms the ground on which end users make judgments about the value of the product)
• Experts (product value is driven by the viewpoints of selectors, who are neither users nor producers but who possess a particular knowledge and expertise).
Selection systems theory is a framework in which value is impossible to conceptualize without establishing particular sets of preferences and dependences between the actors, organizations and individuals interacting within one network. The economic role of selectors in the competitive processes is to determine the actions of the producers who strive to address the needs of their selectors. Based on the type of selection system they find themselves in, they opt for specific types of both first and third party signals (Wijnberg, 2004).
prevails, and the experts’ power has been increasing significantly, being almost as important as the artist himself. Experts, being the source of third party signals influence highly the success in the visual arts market. The artworks’ value is highly dependent on the status of the source, therefore the higher the status of the expert selector – the better the signal emitted by him or her influences the value of the producers’ goods (Wijnberg & Gemser, 2000).
Status
According to Podolny and Philips (1996) status can be defined as a stock of accumulated deference, having a dual foundation in past performance and an actor’s affiliations. Assuming that deference in this case is understood as flow of respect, status would be the stock that corresponds to this ‘flow’. Deference can have multiple forms represented by different behaviors, attitudes or quality signals (Podolny & Philips, 1996).
Possession of status entails a number of benefits for its owner. According to Podolny (1996) two different types of quality signals determine the organization's status: past performance outcomes and the status of the organizations affiliates. The better the past performance outcomes and the higher the status of the affiliates, the greater the organization's growth in status (Towse, 2001).
Usually, organizations as well as individuals strive for having higher status, which is associated with a wealth of social, material, and psychological benefits and let the entities increase their prestige and respect in the eyes of others (Anderson et al., 2012). One of the ways to achieve it is differentiating by gaining the appreciation signals from third parties, such as awards, reviews, imitation of its practices, etc. Therefore, individuals try to be part
of prestigious organizations and high status organizations try to attract already successful individuals or individuals with high potential.
Signals also contribute to establishing status hierarchies among numerous agents present in the market, conceptualized as a status order in which each producer’s status level defines the producer’s actions (Podolny, 1993). It might be stated that the status derives from signals and at the same time is a signal enhancer itself, carrying with it an attribution of superior quality and facilitating the deference mechanism (Stern, Dukerich & Zajac, 2013). Education programs are an example of institutions whose status depends highly on the third party signals and which are signals emitters themselves. In this study, the interdependence between the status of the Rijksakademie and the status of its graduates will be examined.
Art Market
In creative industries producers face a fundamental uncertainty. The simplest way to overcome it is to analyze the past performance of similar goods and take relevant actions to affect the expected quality and appeal of the product. However, due to extremely high heterogeneity of the goods, it is difficult to achieve the minimal predictive value based on the past performance (Goldman, 1984).
Since primary art markets are extremely heterogeneous and uncertain, the substitutes barely exist there and prices of works coming from one artist may vary significantly (Velthuis, 2003). Different types of buyers have different incentives for purchasing a work of art: author matters most (not the work itself), the specific work of art is desired or the incentive is a purely decorative and only the taste of the purchaser matters. In the last case, the substitutability is relatively high. Same mechanism applies in the case of beginning
buyers who still need to define their taste and preferred styles. The maximum of substitutability in the visual arts market is reached in the wholesale, art mass production (Towse, 2011).
Furthermore, visual arts market – as most of the creative industries markets – is a market of credence goods. Their value cannot be objectively measured and the experts are always involved in their valuation, i.e. the attribution of worth. And since the value is derived from quality, the consumers are highly prone to change their opinions based on experts’ beliefs (Alexander & Bowler, 2014).
According to Bourdieu (1992), the essence of the capital in art market is experts’ signals and their widely expressed conviction about its value. It also means that “the art market cannot be understood properly without taking the role of various cultural institutions into account, such as museums, where artworks are exhibited and preserved, ‘alternative’ exhibition spaces like artist’s cooperatives, and the art press (art magazines, newspapers and book publishers that devote considerable attention to the arts). In addition to their cultural functions of exhibiting and reviewing art, putting artworks in a historical or critical context, and educating artists and their audience, these institutions participate in the art market.” (Towse, 2011, p. 37). In other words, those institutions are selectors and by allocating the quality signals to a pool of artworks and artists, they reduce information asymmetry and search costs for economic actors on the market, e.g. a gallery representative can save on these costs if he or she wants to collect art by taking the opinion of cultural institutions into consideration. Moreover, these cultural institutions produce ‘credibility’ or ‘belief’ in the artistic value of art among the audience of museum visitors and art collectors (Velthuis, 2003).
‘cultural judgment’ of curators and critics, on one side, and the ‘economic judgment’ of the market, on the other, are interdependent. Therefore, collectors extract additional advantages from devouring artworks that cultural institutions claim to be of good quality, or their taste is directly impacted by the selection of cultural institutions (Velthuis, 2003). Therefore, the status of the signal emitter is being transferred onto the good and improves its quality perception.
In the visual arts, awards, reviews, exhibitions, art shows, subsidies and gallery affiliation are common value signals. It is important that the value signals effectively convey their value message to decision makers. For a value signal to be effective, the message should be credible, clear, accurate and stemming from a credible source. Since experts dominate the visual arts market, the value signals originating from credible sources should have a greater positive effect on performance than value signals originating from less credible sources.
Furthermore, since it is impossible to measure the quality of visual arts objectively, the artist’s believability to the audience has to be achieved step by step, with engagement of both first and third party quality signals. This way, he or she can produce economic value. However, in order to verify this quality, the consumers need the help of selectors: experts, peers and consumers. As stated before, in creative industries, and in visual arts especially, the experts’ opinion is most valuable. They all base their judgment on a lifelong expertise and cultural knowledge, including both subjective and objective criteria, which makes them capable of assessing the goods’ value most appropriately (Wijnberg, 2004).
However, experts themselves also have to be credible to the audience. Since their believability is transferred to the particular oeuvre they are judging (Velthuis, 2003). This means they are at the same time playing the role of third party signals for the artworks they
are assessing and first party signals for themselves, choosing particular works over the others. The same mechanism works for the status transfer in visual arts: the high status cultural institutions pass on their status to its members and vice versa (Podolny, 1996). Therefore, the institutions considered to have high status select their members carefully, keeping in mind the new members need to fit and not deteriorate the status of the institution. Similarly, the ambitious individuals will always opt for affiliating with institutions considered to be of high status, hoping for elevating their own position among their peers.
Selection Systems and Decision Making Processes in the Art Market
The selection processes in all markets are influenced by signals. One of such signals is education, which influences the market success of the graduates and can be considered as a significant quality signal differentiating them in the market. However, it is difficult to assess to what extent the graduates had possessed the specific skills already before attending the school and how much the school they attended enhanced them.
According to Menger (1999), education in the uncertain markets is an “imperfect filtering device”. He explains it by the fact that learning by doing is crucial there and schools cannot assure universal competences. Both individuals and organizations make decisions built on publicly available information and the one coming from private sources. In the art market, it is the experts who have obtained both kinds of information through their regular engagement in the art scene and through a gradual learning process and who can judge best the quality of the artworks (Velthuis, 2003).
The end-users, on the contrary, do not have access to the same type of information. This is why they seek quality signals as a confirmation of their choices. The quality signals can be
divided into strong and weak ones, depending on how easy it is to recognize them (Gulati & Higgins, 2003). The more accurate information the signal brings, the more possible it is that the signal interpreter will refer to it in the future (Cohen & Dean, 2005).
The extent to which the signal is correlated with quality is defined as signal fit. It defines how strong the signal is and how well it reaches the necessary audience. It can be therefore defined also as the relationship between the public information (signal) and the private information (the signaler’s unobservable quality). The stronger the overlap between those two is, the better the signal fit is (Connelly et al., 2011).
In this dissertation, in the empirical setting of the Rijksakademie Residency Program, I will test what kind of first party quality signals the candidates sent during the interviews and what the jury took into consideration when making the final decision. In order to answer it, in the further part of this research, the following hypotheses will be tested:
1. Relationship between the Admission Criteria and Acceptance to Rijksakademie
H1a: A higher score in creativity increases the likelihood of being accepted to Rijksakademie.
H1b: A higher score in craft increases the likelihood of being accepted to Rijksakademie.
H1c: A higher score in collaboration increases the likelihood of being accepted to Rijksakademie.
H1d: A higher score in critique increases the likelihood of being accepted to Rijksakademie.
Subsequently, I will verify whether the same criteria have any relationship with an independent score in Artfacts.net Ranking:
2. Relationship between the Admission Criteria and Atfacts.net Ranking score
H2a: A higher score in creativity increases the artist’s Artfacts.net Ranking score.
H2b: A higher score in craft increases the artist’s Artfacts.net Ranking score.
H2c: A higher score in collaboration increases the artist’s Artfacts.net Ranking score.
H2d: A higher score in critique increases the artist’s Artfacts.net Ranking score.
Signals as Factors Reducing Uncertainty in the Art Market
Another interesting aspect of quality signals on the highly uncertain art markets is considering them as factors helping to overcome the information asymmetry and indirectly, the uncertainty in the decision making process itself. In order to verify it, I will test whether succeeding in the recruitment process of the education institution fits the overall subsequent market success of the graduates, when compared to the situation of the rejected candidates. As a measure of market success, I will use Artfacts.net Ranking scores.
H3: The candidates who went successfully through the Rijksakademie internal selection process, scored higher externally in the Artfacts.net Ranking than the rejected candidates.
Data and Method
I examined my hypotheses with a quantitative analysis of a renowned education institution within the visual arts industry. As stated before, the visual arts industry is a market characterized by high degree of uncertainty and no possibility of putting objective
measures on the quality of the work. This is why the dominant selection system is the expert system, where specialists, critics, rewards, education and other quality signals come into place in order to reduce uncertainty and facilitate the market choices (Bonus & Ronte, 1997).
For this thesis, I have studied 355 candidates that underwent the entire recruitment process to the Rijksakademie between 1990 and 2005. My sample was derived from the interviews held in the premises of the Rijksakademie, organized after CV and portfolio screenings. The results obtained within the Rijksakademie were subsequently analyzed together with the interviewees’ scores in the international Artfacts.net Ranking.
Rijksakademie and its Selection Process
In the self-evaluation report, the Rijksakademie describes its mission statement as “supporting the development of individual artists and artistic practice, its main goal: excellence” (Fall et al., 2010, p. 4). As the three main assets the institution mentions: a personalized artist residency, hosting the Prix de Rome and making their expertise available through library and collections. Rijksakademie is a unique institution in the Netherlands and in Europe, supporting advanced in their careers artists when further developing in customized programs. This is why it escapes any defined pattern of institutions for post-academic education. It is neither a scientific research oriented institution, or a museum, even though it does have some features of the two (Fall et al., 2010).
The organization attracts internationally recognized artists as its experts providing their opinions and feedback to the residents. This is why, the Rijksakademie aims to select promising artists with a strong potential to impact contemporary art. As for the recruitment
process per se, it takes place through the Internet and the network of former residents, experts and international and national art organizations. The first criterion to look at is the age of the candidates, which is preferred to be between 26 and 35 years old. The next condition is to be a graduate of an art school no longer than two years. The next step is the pre-selection, which focuses on choosing approximately 2.000 applicants (300 from the Netherlands, 1.700 from abroad). It is further narrowed down to 300 Dutch and 350 foreign candidates. After evaluation of their work, a jury consisting of five to six members decides to organize 25-minute interviews with a selection of 35 Dutch and 35 foreign candidates.
The Rijksakademie considers its selection process to be independent as there are no formal requirements or fees. Since the experts being part of the jury will also work with the artists during their residency, a lot of subjective fit-related criteria are taken into consideration as well. This is why, the Rijksakademie aims to choose a diverse jury consisting of advisors from different backgrounds.
Artfacts.net Ranking
The visual arts market is considered to be both a professional world with its own criteria of quality, and the world of the average art consumer, who is often confused and insecure about his or her own judgment of the quality of the works. The criteria taken into consideration by the professionals assessing the quality of art remain unclear and indefinable to a regular art consumer or potential art buyer. The discrepancy between the knowing of the experts and the average consumers comes from the impossibility of grasping the reasons why one work of art is good and the other is of poor quality. It also results in the lack of conviction of a large section of society about their taste. This explains why the art market continues to have such a unique and uncertain character when
compared to other markets.
Artfacts.net has made an attempt to invent a set of factors, objectively assessing artistic careers and listing the most important artists in the descending order. When putting together an international ranking, they took into consideration the success of artists’ exhibitions, assessing them with a points system. As a result they obtained a countable measure of attention each particular artist has received from renowned art institutions. Apart from defining the current auction and gallery sales of the artists, the Ranking aims also to forecast the future development of the artistic careers and track upcoming trends in the market (Artfacts.net, 2005).
In its ranking, Artfacts.net has evaluated around 525,000 international exhibitions operating in officially established, renowned structures - e.g. museums, art shows, and galleries – beginning in 1996. This evaluation has led to developing an algorithm ranking artistic performance relative to other artists in the database consisting of more than 100,000 artists.
As the base for the whole ranking idea, the concept of economy of attention has been used according to which, the art market is led by the common investment rules, similar to the ones defining capitalism. Since the objective criteria for the quality of art are missing, the attention (fame) is crucial for the investors (gallery owners) when making decisions which artworks to purchase and which do not deserve attention. The assumption is that the art buyers just like investors, expect return on their investment and make their decisions, based on quality signals building up the reputation of artists.
According to this concept, the art market can be divided into primary and secondary one. The primary market deals with newly produced art works, the secondary one handles the
‘old’ ones. The level of uncertainty on primary markets is extremely high. The galleries and their managers act as branding managers for the first type, for the second – auction houses and art dealers come into place and re-sell the works based on their hitherto performance and popularity. The prices of art works are also defined based on the market type. On the primary market the price is based on the size of the work and the reputation of the artist – no price differentiation based on quality is made. On the secondary market on the other hand, all the art works are priced independently and uniquely (Claaßen, 2005).
Apart from pricing, this classification also influences the behavior of the art galleries managers. On the primary market, the gallery owner strives to make his own name stand out, based on the reputation of ‘his’ artists. He has long-term contracts with artists based on mutual aesthetical taste and they work together on making the gallery and the artists’ names recognizable all over the world, building up their reputation. On the secondary market, the provenance of the work is crucial and high prices are achieved mainly by the reputation of the artist, his exhibition history, a good condition of the work itself and its authenticity.
In sum, there are two types of factors affecting the quality of contemporary art: reputable market participants (third party signals) reducing the uncertainty in the art market and quality signals produced by the artists (first party signals). The Artfacts.net Ranking takes into consideration both of them and this is why in this study the Ranking will be considered as an objective measure of artistic success.
Data Collection
The database used for this study consists of 15 years quantitative data concerning the acceptance to the Rijksakademie in the years 1990 to 2005. The study revolves around this period because of access to comprehensive selection criteria notes on applicants who were
invited to the last ‘interview’ round of the selection process in the Rijksakademie. In average, 65 artists are invited annually to this last round.
This data is carefully filtered for the specific goals of this thesis. Every artist who applied and reached the interview stage in the recruitment process received a unique number to ensure comparability and connectivity between the various datasets. In total, 355 artists born between 1941 and 1982 were included in the sample, 157 of who were accepted and 198 rejected after the interview. The total sample size for this study contained a percentage of 25% Dutch and 75% non-Dutch, 48% male and 52% female. After classifying all the candidates, their Artfact.net Ranking has been cross-checked: 254 of all the interviewed applicants were ranked in the Artfacts.net 2005 Ranking, 124 of whom were accepted and 127 rejected in the Rijksakademie. Their scores ranged from 277 to 136347.
The focus on the interviewees and not on all the candidates happened for a few reasons: first, it is already confirmed, these applicants have signaled potential excellence the most, and therefore, the criteria for accepting the candidate or not are the most strict at this stage. Moreover, the interview round has the most precise selection criteria and therefore it enables posing the concrete questions about the predictive value or the influence of selection processes on the future career success. And lastly, the miscellaneous and administrative problems such as missing application deadlines are eliminated at this stage of recruitment, which makes the results more reliable.
Two different jury groups consisting of three evaluators and one facilitator conduct the interviews of the candidates. All of them conduct an individual interview of around 30 minutes. After the whole round the jury discusses the following aspects of the candidate: work content, technical craft skills, creativity, motivation behind the candidates’ work, communication skills, ability to link practice and theory, future potential and their fit
within Rijksakademie organization. The facilitator takes notes during the interview and final points for each of the abovementioned criteria are given to the candidate (i.e. +, +/, -). After summing those up, the jury makes the final decision on admitting or rejecting the candidate.
The transcripts of these in-depth interviews have been used to draw out the selection criteria and weigh them accordingly to their influence on the final decision. Since the Rijksakademie does not have an official list of admission criteria, the notes helped to understand better the selection procedures. Subsequently, the criteria have been classified into the following sets that were given a score from 1 to 3:
• Work related content (artistic development over time) • Context (candidate’s self-positioning in art history) • Craft (technical artistic skills)
• Creativity
• Education and expertise (educational background and experience) • Critique (openness to assessment and criticism)
• Communication (professional awareness and confidence) • Career fit (rightness of the moment in artist’s individual career) • Collaboration (jury member’s eagerness to work with the candidate)
• Contradiction and final assessment (level of agreement in making the final decision).
Variables
As mentioned before, the focus of this paper is to explore the three following dependences:
1. The relationship between four different criteria and being admitted or rejected to the Rijksakademie.
2. The relationship between artists’ scores in four different criteria within the Rijksakademie recruitment process and their score in Artfacts.net Ranking in 2015.
3. The occurrence of a change in Artfacts.net Ranking received by artists who after the interview for the residency at the Rijksakademie got accepted for the program or not.
Since I did not use more than one item per variable, the Cronbach’s alpha reliability check was not needed.
Criteria scores variables. All the criteria have been derived from interview notes and coded according to the following rule: 0 = empty (jury has not mentioned the criterion during the interview), 1 = negative (jury has expressed the unsatisfactory level of compliance with the particular requirement), 2 = neutral (neither a negative nor a positive opinion on the criterion has been expressed), 3 = positive score (the jury found the candidate fulfilling the required skill or characteristic). Out of ten criteria taken into consideration during the interviews, four have been analyzed in this research. I chose them as best probable variables based on the correlations between each criterion and the positive decision of the jury and the Arfacts.net Ranking score, which are the following: creativity, craft, collaboration and critique (Table 1). However, no participants of the study have been deleted, thus the randomization has been maintained and selection bias has been avoided. The next step was to analyze the relationship between those four and the admission to the Rijksakademie or the Artfacts.net Ranking score of the studied artists.
Table 1 Correlation matrix M SD 1. 2. 3. 4. 5. (1) Rijksakademie Admission .44 .50 (2) Artfacts.net Ranking Score 35867.22 38925.15 -.024 .701 (3) Creativity 1.15 .97 .251* .000 -.003 .968 (4) Craft .67 .87 .138* .009 .023 .715 .344* .000 (5) Critique .35 .67 .158 .003 -.047 .453 .278* .000 .226* .000 (6) Collaboration .17 .56 .148* .005 .020 .749 .190** .000 .061 .253 .206* .000 *Correlation is significant at the 0.01 level (2-tailed)
Jury’s verdict variable. Jury’s verdict on accepting the candidate or not is in this study is a categorical binary variable, i.e. it can only have two values: positive or negative one. First, I have examined whether any particular criterion has an outstanding relationship with the fact that the candidate has been admitted or not. Furthermore, I studied the relationship between being accepted in the Rijksakademie and the subsequent artistic success, measured by the score in Arfacts.net Ranking.
Artfacts.net Ranking score variable. The Arfacts.net Ranking score is a continuous dependent variable in this study. Two dependences concerning this variable are examined: whether a particular criterion or the fact of being admitted to Rijksakademie has any relationship with the artistic performance measured by Artfacts.net Ranking score.
The scores in the 2015 ranking range from 1 to 5 561,186 points and the highest score belongs to Andy Warhol. The minimum score of the artists examined in this study is 117 and the maximum score is 136,347.
Artfacts.net takes into consideration the following factors when estimating the scores:
• Artist points (each artist with international long term relations receives points according to the number of galleries, museums and countries he or she cooperated with) • City points (based on the geographical location of the institution representing or collecting ‘international’ artists. Capital cities like Paris or New York with vast numbers of museums and galleries receive more points than small cities)
• Institution points (it sums up the artists’ points for each institution as well as considering the geographical location of the institution by multiplying the city points to the institutions value)
• Exhibition points (the exhibition points decide on the final position of the artist in the Ranking. The system assigns points to different kinds of exhibitions held in major art institutions, such as solo and duo shows in private galleries or public institutions, group shows in private galleries or public institutions and other regular exhibitions, and auction previews held in an auction houses).
Controls. All the results have been controlled for age, gender and nationality of all the candidates. Age of the artists studied ranged from 33 to 77, taking 2015 as the year for calculations. The two genders of the artists have been operationalized into men being 0 (n = 171) and women being 1 (n = 184). The nationality has been grouped into two sets: Dutch being 1 (n = 90) and non-Dutch candidates being 0 (n = 265).
Results
Pre-tests Data Treatment and Descriptives
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.
The variables of year of birth have been recoded into age, considering 2015 as the analyzed year. The gender of the studied artists has been recoded to male being 0 and female being 1, nationality to Dutch being 0, international being 1, accepted to Rijksakademie being 1 and rejected being 0.
Hypotheses Testing
Relationship between the Admission Criteria and Acceptance to Rijksakademie
H1a: A higher score in creativity increases the likelihood of being accepted to Rijksakademie.
H1b: A higher score in craft increases the likelihood of being accepted to Rijksakademie.
H1c: A higher score in collaboration increases the likelihood of being accepted to Rijksakademie.
H1d: A higher score in critique increases the likelihood of being accepted to Rijksakademie.
A logistic regression was performed to ascertain the relationship between creativity, craft, collaboration and critique criteria and the likelihood that participants are accepted to the Rijksakademie, when controlling for age, nationality and gender.
Firstly, it has been tested if the control variables individually have any relationship with the fact of being accepted or not to the Rijksakademie. The result of the analysis was that the controls have no direct effect.
Secondly, a logistic regression analysis has been conducted in order to ascertain the effects of each of the criteria separately, together with covariates, on the likelihood of being accepted to the Rijksakademie. In all the cases all the assumptions have been met.
Relationship between creativity score and being accepted into Rijksakademie
The logistic regression model was statistically significant, χ2(4) = 30.344, p < .05. The model explained 11.6% (Nagelkerke R2) of the variance in Rijksakademie admissions and correctly classified 65.0% of cases. Sensitivity was 50.7%, specificity was 76.9%. Of the three covariates and one predictor variables only one covariate and one predictor were statistically significant: nationality and creativity respectively (as shown in Table 2). Dutch applicants had 1.7 times higher odds of being accepted than the international applicants. High score in creativity criterion increased the likelihood of being accepted 1.8 times. Therefore, the H1a has been accepted.
Table 2 Logistic regression predicting likelihood of being accepted to Rijksakademie based on Age, Nationality, Gender and Creativity
Variable B S.E. Wald p Odds
Ratio
Gender -.085 .232 .134 .714 .919
Age .028 .022 1.673 .196 1.029
Nationality .540 .267 4.099 .043 1.716
Creativity .611 .124 24.135 .000 1.843
Relationship between craft score and being accepted into Rijksakademie
The logistic regression model was statistically significant, χ2(4) = 10.746, p < .05. The model explained 4.2% (Nagelkerke R2) of the variance in Rijksakademie admissions and correctly classified 58.7% of cases. Sensitivity was 35.5%, specificity was 78.0%. Of the three covariates and one predictor variables only the predictor was statistically significant: craft (as shown in Table 3). High score in craft criterion increased the likelihood of being accepted 1.4 times. Therefore, the H1b has been accepted.
Table 3 Logistic regression predicting likelihood of being accepted to Rijksakademie based on Age, Nationality, Gender and Craft
Variable B S.E. Wald p Odds
Ratio
Gender -.144 .225 .408 .523 .866
Nationality .457 .258 3.149 .076 1.580
Craft .329 .130 6.339 .012 1.389
Relationship between collaboration score and being accepted into Rijksakademie
The logistic regression model was statistically significant, χ2(4) = 12.400, p < .05. The model explained 4.9% (Nagelkerke R2) of the variance in Rijksakademie admissions and correctly classified 59.0% of cases. Sensitivity was 25.7%, specificity was 86.8%. Of the three covariates and one predictor variables only the predictor was statistically significant: collaboration (as shown in Table 4). High score in collaboration criterion increased the likelihood of being accepted 1.8 times. Therefore, the H1c has been accepted.
Table 4 Logistic regression predicting likelihood of being accepted to Rijksakademie based on Age, Nationality, Gender and Collaboration
Variable B S.E. Wald p Odds
Ratio
Gender -.147 .226 .422 .516 .864
Age .029 .021 1.945 .163 1.030
Nationality .342 .259 1.748 .186 1.408
Collaboration .604 .230 2.797 .009 1.829
Relationship between critique score and being accepted into Rijksakademie
correctly classified 59.6% of cases. In the mode, sensitivity was 30.9% and specificity was 83.5%. Of the three covariates and one predictor variable, only the predictor was statistically significant: critique (as shown in Table 5). High score in collaboration criterion increased the likelihood of being accepted 1.6 times. Therefore, the H1d has been supported.
Table 5 Logistic regression predicting likelihood of being accepted to Rijksakademie based on Age, Nationality, Gender and Critique
Variable B S.E. Wald p Odds
Ratio
Gender -.109 .226 .235 .628 .896
Age .030 .021 1.961 .161 1.030
Nationality .350 .259 1.824 .177 1.419
Critique .500 .177 8.022 .005 1.649
Relationship between the Admission Criteria and Atfacts.net Ranking score H2a: A higher score in creativity increases the artist’s Artfacts.net Ranking score.
H2b: A higher score in craft increases the artist’s Artfacts.net Ranking score.
H2c: A higher score in collaboration increases the artist’s Artfacts.net Ranking score.
H2d: A higher score in critique increases the artist’s Artfacts.net Ranking score.
Since none of the correlations between the criteria and the Artfacts.net Ranking score were statistically insignificant (see Table 1), it was highly unlikely that any of the hypotheses will be supported. However, I have run a linear regression on all of them together in order to check whether any dependence can be seen. The analysis showed that the correlations and significance levels were virtually the same as in the initial correlations test. Only two covariates: nationality and age had a statistically significant effect on the Artfacts.net Ranking socre. Therefore, the analysis confirmed that the hypotheses could be rejected.
Table 6 Linear Regression Correlations Matrix
M SD 1. 2. 3. 4. 5. 6. 7. (1) Artfacts.net Ranking Score 35867 38925 (2) Nationality .26 .44 .120* .029 (3) Gender .52 .50 .023 .356 .074 .120 (4) Age 42.6 5.43 .153* .008 -.001 .496 -.061 .170 (5) Craft .67 .87 .024 .351 -.018 .388 .035 .289 .297* .000 (6) Creativity 1.15 .97 -0.12 .426 -0.60 .170 -.120* .029 .036 .287 .321* .000 (7) Collaboration .17 .56 0.33 .304 .152* .008 .077 .112 -.025 .345 .073 .123 .181* .002 (8) Critique .35 .67 -0.33 .301 .073 .126 .030 .316 -0.60 .172 .254* .000 .328* .000 .146* .010 *Correlation is significant at the 0.05 level (1-tailed)
A linear regression has been run: with the controls of age, nationality and gender and including the admission criteria: creativity, craft, collaboration and critique. All assumptions have been met except for the homoscedasticity and normal distribution but the number of participants was more than 250, therefore the analysis could still be conducted.
The first model regression has shown that the controls can predict statistically significantly the Artfacts.net Ranking score, F(3,247) = 3.289, p = .021. However, the variance explained is only 3.8%.
Table 7 Criteria VIF scores
Criterion VIF score
Creativity 1.213
Craft 1.141
Collaboration 1.158
Critique 1.051
Dependent variable: Arfacts.net Ranking Score
The second model regression analysis (including the criteria) revealed that the criteria taken altogether, whilst controlling for age, nationality and gender, cannot predict Artfacts.net Ranking score statistically significantly, F(7, 243) = 1.509, p = .165, variance explained being only 4.2%. Therefore the hypotheses H2a – H2d cannot be proven.
Table 8 Linear Regression predicting likelihood of having higher Artfacts.net Ranking score based on Age, Nationality, Gender, Creativity, Craft, Collaboration and Critique
Model df F p 1* Regression Residual Total 3 247 250 3.289 .021 2** Regression Residual Total 7 243 250 1.509 .165
*Predictors: Age, Nationality, Gender
**Predictors: Age, Nationality, Gender, Creativity, Craft, Collaboration and Critique
Difference between Acceptance Rate to Rijksakademie and Artfacts.net Ranking Score H3: The candidates who went successfully through the Rijksakademie internal selection process, scored higher externally in Artfacts.net Ranking than the rejected candidates.
Independent samples t-test analysis without covariates
The following analysis has been conducted in order to explore a hypothesized difference in the Artfacts.net Ranking score between the candidates accepted (n = 124) and rejected (n = 127) by Rijksakademie. 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, Artfacts.net Ranking score - 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, the Artfacts.net Ranking Score was normally distributed and fulfilled the requirement of homogeneity of variance.
The group of the accepted candidates (n = 124) had a mean of 36786.12 and standard deviation of 40808.326. The group of the rejected candidates (n = 127) had a mean of 34903.87 and standard deviation of 36988.935. However, the difference between the candidates accepted and rejected by Rijksakademie in the Artfacts.net Ranking score was not statistically significant t (df) = .385, p = .701. Therefore the hypothesis 3 can be rejected.
Univariate ANOVA with covariates analysis
Subsequently, a univariate ANOVA analysis has been run in order to check whether the difference changes when controlling for age, gender and nationality.
The ANCOVA has been run to determine the effect of the two different sets mentioned above, after controlling for age, gender and nationality. There was homogeneity of regression slopes as the interaction term was not statistically significant, F(1,246) = .895, p = .345. Standardized residuals for the interventions and for the overall model were normally distributed, as assessed by Shapiro-Wilk's test (p > .05). There was homoscedasticity and homogeneity of variances, as assessed by visual inspection of a scatterplot and Levene's test of homogeneity of variance (p = .097), respectively. There were no outliers in the data, as assessed by no cases with standardized residuals greater than ±3 standard deviations. Therefore all the assumptions have been met. However, there was no statistically significant difference in the Artfacts.net Ranking score between the candidates accepted and rejected by Rijksakademie, F(1,246) = .895, p = .345 > .05, thus, the H3 is also rejected, when controlling for age, gender and nationality.
Discussion
Although a more in-depth evaluation of the limitations of the above study will be discussed below, some general comments on the validity of the research and its results have to be made before any conclusions are drawn. There have been multiple factors identified in the literature as likely to influence the studied relationships and differences - but which were not directly addressed and controlled for in the statistical models used. As a result, the following discussion of the findings should be seen as rather explorative and its deductions should not be treated as in any way conclusive.
General discussion
The prevalence of quality signals is very high in the creative industries, making them an attractive empirical setting to study signals’ influence on selection and status transfer processes. For this paper, the setting of education in visual arts has been selected.
This thesis has studied the proposed positive relationship between selection criteria within the Rijksakademie and criteria used in the visual arts market. It also explored how differently the rejected and accepted candidates performed in the visual arts market and what the relationship between being accepted to the Rijkskakademie and reaching high score in the Artfacts.net Ranking was.
As expected, all the criteria differentiated based on the Rijksakademie recruitment interviews transcripts were positively correlated with being admitted to the program – hypotheses 1a-1d have been supported. The strongest correlation has been shown by creativity criterion, which was most predictive of the candidate’s admission.
rejected. This might imply that the signal fit within the internal perspective was definitely stronger than the signal fit within the external perspective. It also could mean that the Rijksakademie does not value the signals that are valued later on in the visual arts market.
Finally, there was no statistically significant difference in the Artfacts.net Ranking score between the accepted and rejected by the Rijksakademie candidates. This leads to a conclusion that in this particular case the status transfer between the prestigious Rijksakademie and their graduates has not taken place and graduating from the Rijkskakademie is not necessarily a signal facilitating achieving higher performance on the visual arts market.
The above mentioned research results do not confirm the findings of previous studies that underline the importance of signals (especially those emitted by experts) in the visual arts market. However, this tendency might be explained by the specificity of an academic education in the field of arts, which is sometimes perceived as a deficient filtering device (Menger, 1999). Therefore, it could be hypothesized that being accepted at an academy or obtaining a degree from an institution may be seen as a weak quality signal and not a good predictor of success in the visual arts market.
Moreover, it has to be taken into consideration that quality and value in visual arts are built gradually through complex social processes and they depend on the social interactions and the social context, in which the product is placed (Currid, 2007). Furthermore, the past performance of the candidates’ influences highly the overall performance of the artists – their education being only a very small fraction of all the signals they produce.
Limitations
Although I believe that this dissertation contains valuable insights into the market for visual arts and the artists’ quality and value determination process, I recognize that some shortcomings limit the explanatory power of the thesis.
The results show the relationship between quality signals and the overall performance of an artist. Mix evidence is found for the hypotheses, but in essence they reject the core assumption that the signals artists sent by graduating from a prestigious residency program increase their performance on the visual arts market. However, the limitation is that there are no compatible indicators evaluating the overall performance in visual arts and the artistic performance is difficult to measure objectively since it is often heavily biased.
Another important limitation to the research is the under-explored moderation effect. Spence (1974) differentiated between indices as the inflexible entities’ features (e.g. age, gender, nationality) and signals – observable characteristics attached to the individual that are subject to manipulation. The indices might have a significant effect on how the signals are being perceived by the receivers. Even though all the hypotheses have been tested with controls of age, gender and nationality, the results still miss the proper exploration of the moderation effect of each of the indices on the signals studied. A good example of how it can affect the overall results is the element of the recruitment system in the Rijksakademie. It gives the favor to Dutch candidates by requiring an equal balance between residency places for artists from abroad and from the Netherlands (25 – 25). Since the pool of international applicants is bigger and more various, the foreign residents tend to perform better than the Dutch.
