Cultural Agenda Setting

Abstract

The recent transformation of cultural organizations has altered their relationship with the public, evidenced by their engaging in market research and analysis of public perceptions about their identity, mission, and services. Through the lens of agenda-setting theory, this paper examines whether, alongside other factors, media coverage of museums can drive museum visitation. We pursue a second level of agenda-setting analysis, examining the relationship between valence—the positive, negative and neutral portrayals of museums in the media—and visitation. We conduct an econometric analysis of monthly data for the period June 2004-June 2008 for a sample of major Greek museums. The study reveals that media coverage of cultural organizations and positive valence constitute significant elements of reputation, which subsequently drive museum visitation.

Keywords

cultural agenda setting museum salience museum attributes

Introduction

The agenda-setting theory has emerged as one of the most significant paradigms in the fields of journalism and political communication. In its 40-year history, agenda-setting theorists have dealt with two fundamental types of questions/hypotheses. At the theory’s “first level,” they have investigated the salience of “issues” or “objects”, while at the “second level,” they have tackled the question of “attribute” or “frame” salience. The theory’s key premise pertains to a “transfer of salience,” a clear and conclusive relationship between what the media portray and what the public subsequently perceives as important. The notion that the media attribute significance to certain issues, which is subsequently transferred to the public, has influenced scholars from many different disciplines. At the first level of the theory, researchers tackled the question of salience in terms of political issues (Funkhouser, 1973; McCombs & Shaw, 1972). Following the seminal Chapel Hill study (McCombs & Shaw, 1972), researchers accumulated evidence supporting the hypothesis that the media establish the salience of issues in the public mind. In the significant body of literature accumulated, different methodological tools have been employed to examine different hypotheses, which supported different facets of the foundational hypothesis of the Chapel Hill study.

In the 1990s, theorists developed a “second level” of agenda-setting analysis, turning their attention to “attributes” or “frames” of issues. The notion that frames or attributes of issues determine the possibility of salience gave new momentum to the theory, while opening up new thematic areas for research. One of the most promising areas of attribute agenda setting was again linked to political candidates, both because of the potential for theory development as well as the numerous applications for political practitioners. Previous research has paid particular attention to different categories of affective and cognitive attributes (Ghanem, 1997). The affective dimension deals primarily with tone: the positive, negative, or neutral media disposition toward the political personality. In the cognitive category, researchers have focused primarily on ideology, qualifications, and the personality traits of a political candidate (Kiousis & McDevitt, 2008; McCombs, 2004).

Organizations and Agenda Setting

A handful of studies have applied the agenda-setting theory using the organization as their “object.” For example, Carroll and McCombs (2003) examined both levels of the theory in the context of business organizations. At the second level, substantive and evaluative attributes of corporations were measured in media content, and the authors concluded that the theory can be applied very well as “images transmitted by the media to the public prime the attitudes and opinions that various publics hold about a firm” (p. 45). In a seminal study on corporate reputation, van Riel and Fombrun (2002) developed an index of reputation involving the general public providing a public salience measure for corporations. The Reputation Quotient (RQ) they developed has been used in different countries around the world. “The result is a list of the most visible corporate reputations in a country” (p. 297). The development of organizational reputation measurement instruments constitutes a useful device that allows researchers to measure salience both in terms of the “object”—in this case the organization—as well as different attributes that define different organizations.

Meijer and Kleinnijenhuis (2006) applied the agenda setting and issue ownership theories while investigating issues and attributes of corporate reputation. They found strong support for second-level agenda setting. They argued “if a lot of television news broadcast Shell and the environment, then the respondents were more likely to associate Shell with the environment” (p. 554). A similar approach was pursued by Deephouse (2000), examining the reputation of banks based in the metropolitan area of Minneapolis–St Paul, Minnesota. While encompassing an array of different variables and different theories, Deephouse linked the agenda-setting paradigm with other resource-based theories, concluding that “this study expands our knowledge of reputation as a resource. It develops a variant of the reputation concept called media reputation, defined as the overall evaluation of a firm presented in the media” (p. 1108).

The current project expands the above literature by examining whether museum media coverage influences their visitation. Cultural organizations are inherently different from business organizations, with their functionality linked to a list of unpredictable audience aesthetics and preferences, which depend on peoples’ cultural identities and their accumulated sets of values. Therefore, this line of research can provide insights for what has been referred to as a cultural agenda-setting hypothesis (Bantimaroudis, Zyglidopoulos, & Symeou, 2010), the hypothesis that media salience influences cultural organizations’ public salience. This study investigated, within the first-level agenda-setting theoretical framework, the impact that the media salience of museums has on their public salience, which was measured through a behavioral indicator (number of visitors/month). The study lends support to a cultural agenda-setting hypothesis at the first level, in other words that the media salience of museums—measured as the sheer volume of times a particular museum appears in the media (Kiousis, 2004)—did have a positive impact on museum visitation. The current study draws on and expands this earlier work as follows. First, it uses the same behavioral measures for public salience (museum visitors/month); second, it expands on the previous work by investigating the impact of the media on a larger sample of museums and by investigating the cultural agenda-setting hypothesis at a second level; and third, it replicates the earlier study by retesting its findings regarding first-level agenda setting. The following research hypotheses arise from the foregoing discussion. At the first level of agenda setting:

Hypothesis 1: Media visibility will be positively associated with the volume of museum visitation

At the second level of agenda setting:

Hypothesis 2a: Favorable media tone will be positively associated with the volume of museum visitation.

Hypothesis 2b: Unfavorable media tone will be negatively associated with the volume of museum visitation.

Method

Our sample consisted of 13 museums located in Athens and other Greek regions, both public and private ones, thus covering a wide array of cultural organizations. These museums are as follows:

The National Archaeological Museum (NAM), the largest public museum in Greece, housing more than 20,000 exhibits, providing, at the time of the study, an overview of Greek civilization.

The Archaeological Museum of Olympia (AMO), a regional public museum housing exhibits from ancient Olympia, the original location of the ancient Greek Olympic Games.

The Byzantine and Christian Museum (BCM), located in Athens, housing a large array of artifacts from the Byzantine and post-Byzantine eras.

The Athens Numismatic Museum (ANM), with its collection of more than 600,000 coins from most periods of Greek history.

The National Gallery (NG), located in Athens and housing more than 15,000 works of painting, sculpture, and engraving.

The Benaki Museum (BM), located in Athens, the most important private¹ museum of Greece, presenting a vast collection of cultural products.

The Archaeological Museum of Thessaloniki (AMT), a public museum, located in Thessaloniki, the second largest Greek city, housing treasures from the regions of Macedonia and Thrace.

The Museum of Greek Folk Art, located in Athens, housing a significant folk art collection from the early 1500s.

The Museum of Modern Greek History, located in Athens, a significant research center dedicated to the study of Modern Greek history and containing significant historical artifacts.

The Museum of Byzantine Culture, located in Thessaloniki, a modern research center dedicated to the study of the Byzantine era. It contains artifacts from the early Christian period, the Byzantine, and the post-Byzantine periods.

The Archeological Museum of Dion, located in Pieria, very close to Mount Olympus, housing significant collections from the prehistoric and classical Greek periods.

The Archeological Museum of Pella, located in the ancient city of Pella, the capital of Ancient Macedonia and the home of Phillip II and Alexander the Great, housing significant collections from the Macedonian era, including artifacts from the ancient agora and the private houses of the Macedonians.

The Archeological Museum of Delphi, located in Delphi, one of the most important Greek museums. It exhibits the history of the Delphic Sanctuary and houses rich collections of works of art donated to the sanctuary throughout its existence.

To gather the relevant information on these museums, we drew on two types of data sources. First, we drew on the General Secretariat of the National Statistical Service of Greece² for information on the monthly visitors and income of the museums. Second, to measure the visibility and tone of coverage that the above museums received from the media, we drew on the three major Athenian daily newspapers, Kathimerini, Ta Nea, and Eleftherotypia, because not only do they have a sophisticated digital archive, which made the search easier but also, given their readership (more than 600,000 daily readers) and distribution throughout Greece, they can influence other media. Overall, our search of these three newspaper databases resulted in 4,717 articles, where at least one of the above museums was mentioned. To assemble the content, we searched for the museum’s name (e.g., Benaki Museum) within a time frame of 1 month (e.g., from January 1, 2007 to January 31, 2007). The complete dataset covers a period of 49 months between June 2004 and June 2008 and pertains to 637 museum-month observations. A number of missing values reduces the usable observations to 580.

Dependent Variable

This study examines the impact that museum media salience has on public salience. However, given the unavailability of data directly measuring the public salience of museums, we used a behavioral indicator of public salience based on research in cognitive psychology, which has shown that increased media salience does influence individual decision making (Iyengar & Kinder, 1987; Iyengar, Kinder, Peters, & Krosnick, 1984; Kadres, 1994). Namely, we measured public salience by counting the monthly visitors for each of the above museums for the period of the study. We considered such a behavioral indicator to be an appropriate measure of salience based on evidence from two lines of research on the role of information accessibility on decision making.

First, within the political cognition domain, many researchers found evidence that information accessibility does play a role in decision making. Iyengar et al. (1984) and Iyengar and Kinder (1987) found that television news programs did help define the standards by which presidents are evaluated. In addition, Higgins and King (1981) and Wyer and Srull (1986) found that accessibility did influence the weights that individuals assign to various considerations when expressing attitudes or making decisions (Iyengar & Ottati, 1994). Moreover, Behr and Iyengar (1985) demonstrated that news coverage did affect the level of concern with particular issues, a finding further supported by Iyengar and Kinder (1987), who found “that individuals habitually refer to issues or events ‘in the news’ when diagnosing current social and political ills” (Iyengar & Ottati, 1994, p. 156).

Second, the literature studying decision processes during consumer judgments reports similar findings. Kadres (1994), for example, reports that according to the accessibility-diagnosticity model (Feldman & Lynch, 1988), any increase in “the relative accessibility of a given piece of information . . . also increases the likelihood that the information will be used as an input for memory-based judgement” (Kadres, 1994, p. 439). This position has been supported empirically by multiple experimental studies. For example, Lynch, Marmorstein, and Weigold (1988) found “that when attribute information was accessible . . . subjects used recalled attributes as inputs to choice . . . and failed to use recalled evaluations as inputs to choice” (Kadres, 1994, p. 440). Moreover, Herr, Kadres, and Kim (1991) found that vividly presented information has a greater chance of being used in making judgments, a phenomenon that they labeled the “vividness effect.”

Even though our measure of public salience is not the one typically used within the agenda-setting literature, following McCombs (2004) we use a behavioral measure for salience. As McCombs (2004) points out in his latest book titled “Setting the Agenda,” there is evidence that a behavioral measure can be used as a reliable measure of public salience, and he refers to a number of studies that point in that direction. For example, as he reports “extensive news coverage of crime and violence, including a murder and a number of rapes, on the University of Pennsylvania campus a few years ago contributed to a significant drop in applications by potential first-year students,” mostly women (2004, p. 129). He also refers to the “successful use of entertainment television programming to spread the idea of “the designated driver,” that member of the party who abstains from drinking in order to drive his or her friends home safely afterwards” (2004, p. 129) to alter the behavior of young adults by Harvard University, as reported originally by Graber (2001). A third study cited by McCombs (2004) pertains to news on airplane crashes and risk avoidance behavior. As Bloj and McCombs (1974) found, news about airplane crashes made the danger of flying salient over a period of 5 years, data were divided in high salience weeks and low salience weeks and “as expected, ticket sales dipped in high salience weeks and, conversely, flight insurance sales increased” (2004, p. 130). Moreover, the use of behavioral measures as indicators of salience found support in the work of Moon (2008), who in an extensive study investigating the consequences of agenda setting on voting behavior and other types of civic engagement concluded that “the effects of agenda-setting may be viewed as related to the behavioral as well as the cognitive levels so that: What the public thinks about something can be extended to what the public does about something” (2008, p. 2)

Independent Variables

Our main independent variables are media visibility and media tone. Media visibility, denoting the total coverage of a museum in the media, was measured by the sum of the monthly mean-corrected total number of times a museum appeared in the three newspapers (Ta Nea, Eleftherotypia, & Kathimerini). A Cronbach’s alpha of .965 suggested a strong intercorrelation and consistency among the three newspapers supporting the construct development. Researchers have often measured media salience through the sheer volume of coverage (Bantimaroudis et al., 2010; Kiousis, 2004). However, the current study, within the context of second-level agenda-setting theory, investigates the role of the media tone of the coverage that a museum received for a given month. Drawing on prior work by Deephouse (2000), we classified the media tone of an article as “positive,” “negative,” or “neutral.” For example, an article was classified as referring to a museum in a favorable manner when there were references to its well-maintained infrastructure, its helpful personnel, the good quality of its exhibitions and events, or other services. The coders were instructed to search for references on active organizations involved in different types of cultural activities such as periodic exhibitions, presentations, speeches, debates, concerts, and so on. The “positive” category was to be selected if a museum were described as expanding, undergoing renovation, moving to better location, acquiring new exhibits, or being recognized as an active and progressive institution.

An article was classified as referring to a museum in a negative manner if there were references to an organization that visitors do not care to visit or had poor visiting experiences, including low-quality exhibits or services provided by the institution. Furthermore, articles were coded as negative if the relevant organization received poor reviews by critics and visitors alike, including references to complaints, problematic behavior, such as lawsuits, delays, low-quality services, economic problems, low-quality infrastructure, and so on.

Articles were classified as neutral when the museum was mentioned, but it was portrayed in neither a favorable nor an unfavorable manner. In other words, certain events pertaining to the museum were simply reported, without further elaboration. From a population of 4,717 articles, 1,261 were coded as positive, 72 were coded as negative and 3,384 were coded as neutral. From those three categories of tone, in line with established practice (Deephouse, 2000), we included in our regression models only the “positive” and the “negative” articles, while we excluded the “neutral” ones.³

The coding was conducted by one of the authors, and to assess intercoder reliability of media content, two experienced coders recoded 6.8% of the total population of media articles (a total of 325 articles). According to Lombard, Snyder-Duch, and Bracken (2002), a size of 300 articles is usually adequate to assess intercoder reliability. In this case, the level of agreement was 0.92 (Holsti measure) and 0.79 (Scott’s Pi measure), which also accounts for agreement by chance. Three constructs were respectively estimated by dividing each newspaper’s monthly tone (e.g., the monthly number of positive mentions) by the newspaper’s mean tone and adding the results from all newspapers. Similarly, the Cronbach’s alphas were all high, supporting the consistency of the constructs.

Control Variables

A number of control variables were included in the model to account for museum-specific characteristics, environmental effects, and seasonality in the data. First, we controlled for the promotion activities of a particular museum by measuring the relationship among museum income, number of visitors, and ticket price. Income is divided by the number of visitors multiplied by the ticket price. This ratio is always below 1 because a number of tickets are being sponsored or subsidized by the museums or affiliate organizations. Therefore, smaller ratios might imply that a larger number of visitors have visited the museum using a sponsored ticket. Correspondingly, the ratio acts as a proxy for the magnitude of promotion the museum receives. Promotion’s relationship with visitation is expected to be positive. The variable is included in a logarithmic form.

Second, we control for the fact that the Olympic Games took place in Athens in August 2004, through the inclusion of a dummy variable that takes the value of 1 for the months of July or August 2004, and 0 for all other months. Third, the governance of the museum is denoted by a dummy variable that takes the value of 1 if the museum is a “private” institution and 0 if public. Certain museums such as the Benaki Museum are “private” institutions and take active initiatives for their promotion and fundraising that likely have a positive effect on visitation. This variable recognizes these governance differences between public and “private” cultural organizations.

Fourth, we control for museum location, given that the location of a museum might play a significant role in its attractiveness. Museums located in Athens, a city with a metropolitan population of 3.9 million in 2007, might have had an advantage over those in rural areas. Anticipating that Athenian museums will have higher visitation levels, we include a dummy variable “Capital” in our models to distinguish between Athenian and non-Athenian museums. Fifth, in order to account for people’s disposable incomes, the model includes monthly GDP per capita in a logarithmic form. GDP per capita is widely used in empirical studies in the social sciences to control for the financial capacity of agents (i.e., museum visitors) to complete an activity that bears an economic cost. The variable is expressed in constant Euro prices of 2005, accounting for price inflation. The expected relationship between people’s disposable income and museum visitation is positive.

Sixth, we control for the number of foreign visitors that Greece received for a particular month, given that Greece is a very popular tourist destination with more than 18.7 million foreign visitors in 2007.⁴ Visitors engage largely in sea tourism, but other attractive forms of tourism include mountain holidays, ecotourism, sports activities, and visits to archaeological sites, museums, and monuments all over the country. To control for the number of museum visitors that are unlikely to have been influenced by a Greek newspaper article, the model includes the variable tourism that measures the number of foreign tourists that visit Greece over a period of a month. This variable is expected to have a positive influence on museum visitation. Finally, our econometric models include a set of month-dummy variables to account for seasonality. Seasonality pertains to periodic fluctuations, and in this paper, it is treated as possible noise contaminating the underlying relationship between media visibility and museum visitation. For example, tourism tends to peak for the summer season (June-September) and decline after the holidays. It is very likely that a time-series of visitation will typically show increasing values during the summer period and declining values between October and May. Notwithstanding, the seasonal effect on visitation might be counterbalanced by the increased number of schools visiting the museums over the October-May period. Hence, it is essential that month-dummies be incorporated in the model.

Analysis

For the analysis of the multitudinal data comprising 13 cross-sections (museums) and 49 time-series (months), we use a multiple regression model, the generic form of which is as follows:

Y_{i t} = α_{i} + β_{k} x_{i t} + e_{i t} + u_{i},

where i = 1,2, . . . , 13 is the subscript for the cross-sectional dimension and t = 1,2, . . .,49 is the subscript for the time-series dimension. Y_it is a T × 1 vector representing visitation; α_i is a 1 × 1 scalar constant; β is a series of coefficients corresponding to x_it series of exogenous variables (see Table 1 for variable description); u_i is a T × 1 vector of the effects of omitted time-invariant museum-specific variables; and ϵ_it is a random disturbance variable assumed to be distributed identically and independently with zero mean and finite, constant variance. The model was subjected to the Hausman test to choose the most suitable estimation method among the Fixed Effects or the Random Effects transformations (Baltagi, 2005). The test indicated that a Fixed Effects model was more appropriate because the χ² statistic of the test was quite large (107.4). Also, the authors estimated the regression model with robust standard errors to account for possible autocorrelation and heteroskedasticity in the disturbance tersm ϵ_it.

Table 1.

Variables Involved in the Econometric Analysis

	Description	Source
Dependent variable
Visitation	The number of museum visitors	The General Secretariat of National Statistical Service of Greece
Independent variables
Media visibility	A construct denoting the total coverage of a museum in available newspapers	Developed by the authors. We added up the visibility of the museums in 3 newspapers, “ta nea,” “eleftherotypia” and “kathimerini”
		Cronbach alpha .965
Positive	A construct denoting the amount of coverage with positive tone a museum received for a given month by the media	Developed by the authors drawing on prior work of Deephouse (2000)
Negative	A construct denoting the amount of coverage with negative tone a museum received for a given month by the media	Developed by the authors drawing on prior work of Deephouse (2000)
Control variables
Promotion	Museum income divided by the product of the number of visitors multiplied by ticket price, transformed in logarithms	The General Secretariat of National Statistical Service of Greece
Olympics	A dummy variable denoting the event of the Olympic Games in Athens in July and August 2004	Developed by the authors
Capital	A dummy variable indicating whether the museum was located in Athens or not (this is multiplied by a time trend as otherwise it would be dropped by the Fixed Effects transformation)	Developed by the authors
Month-dummies	Dummy variables for every calendar month	Developed by the authors
GDP per capita	GDP in constant Euro of 2005 transformed in logarithms	Eurostat
Governance	A dummy indicating whether the museum is private or public (this is multiplied by a time trend as otherwise it would be dropped by the Fixed Effects transformation)	Developed by the authors
Tourism	Number of monthly foreign visitors	Eurostat
Month-dummies	Dummy variables for every calendar month	Developed by the authors

Findings

A preliminary examination of summary statistics for the data reveals that the media coverage variable does not follow a consistent seasonal pattern of coverage like the visitors variable (Table A1 in the appendix). Coverage tends to be higher during fall and spring, but not in a consistent fashion. Furthermore, not all museums receive similar coverage by the media. By far the most visible museum in the media is the Benaki Museum, with more than 54 monthly references in all three newspapers, while the least covered is the Museum of Dion, with only 0.02 monthly references. The correlations (Table A2 in the appendix) show that the overall visibility of a museum is more likely to consist of positive (tone) articles rather than negative articles. The number of positive articles has a stronger correlation with visitation compared to the negative counterparts. Moreover, people’s disposable income does not appear to have a relationship with people’s museum visits.

Model 1 in Table 2 treats the main independent variable as a comprehensive media visibility factor. The combined media visibility variable has the expected positive relationship with visitation, which is statistically significant at the 0.001 level. From the list of the control variables, seasonality demonstrates the anticipated outcomes with higher visitation during the summer period (relative to January, the base month). GDP per capita has a positive relationship with visitation, whereas the number of foreign visitors has a negative effect, both statistically significant. An important finding concerns the impact of promotion. Promotion (the variable represents the ratio of actual revenue to potential revenue, suggesting that higher ratios denote lower promotional activities) has a negative coefficient, which suggests that lower promotion is associated with fewer museum visitors. The coefficient is statistically significant at the 0.001 level. Furthermore, the location of the museum (capital) does not impose any limitations to its visitation. Instead the Olympics and museum governance variables have negative relationships with visibility, which are statistically significant at the 0.05 level.

Table 2.

Regression Results: Dependent Variable (Visitation)

	Model 1 Visibility	Model 2 Visibility with lags	Model 3 Valence	Model 4 Valence with lags
visibility	0.031***	0.026**
	(0.009)	(0.009)
lag_visibility		0.029***
		(0.008)
positive			10.680**	6.595*
			(3.758)	(2.658)
lag_positive				16.908***
				(3.232)
negative			-3.876***	-1.302
			(1.164)	(1.121)
lag_negative				-1.293*
				(0.514)
promotion	-0.272***	-0.193**	-0.278***	-0.202
	(0.068)	(0.069)	(0.068)	(0.131)
lag_promotion		-0.260***		-0.264
		(0.070)		(0.129)
gdp_cap	1.603**	1.882***	1.727**	1.851*
	(0.543)	(0.500)	(0.544)	(0.739)
capital	-0.002	-0.001	-0.004	-0.002
	(0.003)	(0.003)	(0.003)	(0.006)
governance	-0.011*	-0.016***	-0.006	-0.009
	(0.004)	(0.004)	(0.004)	(0.005)
Olympics	-0.295*	-0.089	-0.256	-0.063
	(0.134)	(0.128)	(0.133)	(0.135)
tourism	-0.000*	-0.000**	-0.000*	-0.000**
	(0.000)	(0.000)	(0.000)	(0.000)
Month-Dummies
February	-0.023	-0.018	-0.021	-0.018
	(0.101)	(0.092)	(0.100)	(0.058)
March	0.554***	0.560***	0.541***	0.547***
	(0.107)	(0.099)	(0.107)	(0.091)
April	0.958***	0.977***	0.934***	0.932***
	(0.140)	(0.132)	(0.139)	(0.139)
May	1.586***	1.809***	1.566***	1.726***
	(0.281)	(0.269)	(0.281)	(0.187)
June	1.520***	1.827***	1.500***	1.758***
	(0.324)	(0.319)	(0.323)	(0.217)
July	1.628***	1.969***	1.553***	1.842***
	(0.362)	(0.347)	(0.361)	(0.250)
August	1.899***	2.178***	1.822***	2.033***
	(0.374)	(0.358)	(0.372)	(0.234)
September	1.786***	2.117***	1.738***	2.012***
	(0.322)	(0.307)	(0.321)	(0.208)
October	1.190***	1.404***	1.162***	1.338***
	(0.182)	(0.172)	(0.182)	(0.189)
November	0.256*	0.375***	0.243*	0.358*
	(0.103)	(0.096)	(0.103)	(0.119)
December	0.043	0.063	0.046	0.050
	(0.101)	(0.091)	(0.101)	(0.070)
Constant	-3.915	-6.142	-4.727	-5.879
	(3.809)	(3.506)	(3.814)	(5.156)
R-squared (within)	0.362	0.445	0.360	0.452
N	580	564	580	564

Note: Standard errors in parenthesis.

p < .05. **p < .01. ***p < .001.

Model 3 examines valence, a more sophisticated measure of media visibility, by distinguishing between the positive or negative stance of the media toward the museums. Since overall media visibility and valence are highly correlated, leading to a problem of multicollinearity, Model 3 does not simultaneously consider overall media visibility and valence. Besides, the valence measure is derived from a disentanglement of the numerous positive and negative tone articles that jointly comprise the media visibility variable. In essence, the two new variables (positive and negative) account for visibility and valence. This refined measure of media visibility pertains to a second-level agenda-setting analysis. Both measures of valence, the positive and the negative category, are statistically significant at the 0.01 and the 0.001 level respectively and exhibit the expected relationships with museum visitation. Positive valence is positively related to visitation, while negative valence is negatively related to visitation. Very similar results to Model 1 are obtained with regard to the pool of control variables.

To examine whether the independent variables have an impact not only in the present period but also in subsequent periods, we include time-lags in the analysis. Specifically, we estimate models which additionally involve 1-month and 2-month lags for media visibility, the two valence variables (positive and negative), museum promotion, and GDP per capita. The inclusion of these additional variables yields a substantial contribution to the explanation of the variance of museum visitation (i.e., the R² increased roughly from 0.36 to 0.45 for both the visibility and the valence models). For the sake of model parsimony, a series of Wald statistics were estimated that allowed unimportant variables to be dropped from the model.

This process culminated in Model 2, which now includes 1-month dummies for media visibility and museum promotion, and Model 4, which now involves 1-month dummies for the two valence variables, and museum promotion. Model 2 indicates that the lagged effects of visibility and promotion are both important and statistically significant. Namely, media visibility and museum promotion activities in the current month have a positive impact on visitation both in the current month and in the following month. The impact for the following month is even stronger. With regard to Model 4, positive articles in the current month have a positive effect on visitation both in the current and the following month. Instead, negative articles are more likely to have a negative effect on visitation in the following month rather than in the current month, as the coefficient for the latter variable is statistically insignificant. It is noteworthy that promotion loses its statistical significance under this specification.

The nonlagged effect indicates that two possible states of delay pertaining to a delay between media exposure and museum salience and a delay between museum salience and museum visitation are not substantial suggesting that salience and behavior may converge and take place concurrently. In the lagged-effect case we cannot explicitly determine that the statistically significant effect is solely due to a delay of an agenda-setting effect, though we cannot rule it out, either. Since our analysis is based on monthly observations, a conclusion that can be drawn is that museum visibility can lead to museum visitation within a maximum period of a month, regardless of when salience takes place and how long it lasts.⁵ These results fall within the findings of the existing literature, which examines how long an issue will remain salient in people’s minds. For example, Winter and Eyal (1981) suggested that the “optimal effect span” is between 4 and 6 weeks. Stone and McCombs (1981) reported that changes in the media agenda require 2 to 6 months to be fully translated to the public agenda. Shoemaker, Wanta, and Leggett (1989) found that coverage that recurs for 3 or 4 months may have the most impact on public opinion. Specifically for regional newspapers, Wanta and Hu (1994) found optimal time lags of 3 weeks.

The initial Models 1 and 3 establish that media visibility has a positive influence on museum visitation, lending support to Hypothesis 1. Furthermore, the tone, or valence of the media toward the organizations, displays positive relationships when the coverage is positive, supporting Hypothesis 2a, and negative relationships when the coverage is negative, supporting Hypothesis 2b. In addition, controlling for seasonality, the two models show that museum promotion, location, organizational governance, consumers’ disposable income and the size of tourism exert a degree of influence on visitation. Empirical analysis of the two research hypotheses benefits from the inclusion of time lags, as these are introduced in Models 2 and 4. These augmented models shed more light on the time aspect of the effects of focal variables.

Discussion

This paper empirically examines modified first and second-level agenda-setting hypotheses applied to the field of culture. Major Greek museums enjoying international recognition were involved in an econometric analysis of the effects of media visibility on museum visitation. The research design attempted to clarify the agenda-setting hypotheses, controlling for a number of variables that might influence the process of the transfer of salience: the nature of the organization (“private” vs. public museums); the degree of seasonality (seasonal fluctuations of visitation over the year); promotional activities; museum location; prospective visitors’ disposable income; the number of foreign visitors; as well as one-time events, such as the Olympic Games that took place in Greece in the summer period of 2004.

As regards the dependent variable, the assessed influences could be further examined if the authors were able to identify the origin of museum visitors. However, statistical services measuring museum visitors do not keep records on the origin of visitors—that is, domestic vs. foreign. To disentangle the effects of visibility (and valence) on visitation, the econometric analysis attempts to control for museum visitors that are less likely to have been influenced by the Greek newspapers through the number of foreign visitors. The addition of this variable in the models is believed to have allowed for a more accurate measure of the agenda-setting process.

Both hypotheses of this study receive empirical support from the analysis. Namely, at the first level of agenda-setting, museums that are more visible in the media receive higher visitation by the public. At the second level of agenda setting, museums that receive more favorable mentions in the media enjoy higher visitation, whereas those that receive more negative mentions in the media manifest lower visitation. The paper also examines the time-lagged effects of focal variables on museum visitation. The time-lagged effects of media visibility and media valence are statistically significant, though we cannot explicitly suggest the presence of lagged agenda-setting effects as they can also be attributed to a delay between museum salience and museum visitation. We leave the disentanglement of the two effects to future studies, which may utilize more detailed and distinct measures for salience and behavior. Notwithstanding, our analysis suggests that museum visibility can lead to museum visitation within a maximum period of a month, regardless of the timing and duration of salience. These findings have serious implications for museum managers. Namely, the emergent transformation of traditional museums necessitates that museum managers take into consideration the impact that media exposure and promotional activities might have on their museum’s visitation and the timeliness of these effects.

We were unable to include a separate measure for museum reputation in the analysis which could have had a positive effect on visitation. In the corporate environment, reputation is defined as “the overall estimation in which a particular company is held by its various constituents” (Fombrun, 1996, p. 37) reflecting the path dependent history of the corporation and its past performance (Dierickx & Cool, 1989). Management researchers describe reputation as a multidimensional construct (Zyglidopoulos & Phillips, 1999) that constitutes a valuable resource, which contributes substantially to the firm’s financial performance and the sustainability of this performance over time (Fombrun, 1996; Roberts & Dowling, 2002).

In the context of cultural markets, we have argued that museum reputation could be built by various marketing processes, drawn from the corporate environment, such as public relations activities and advertising, which are expected to increase the museum’s attractiveness to the media. We made attempts to collect related information on museum advertising spending, but this information was not separated from general expenditure, and in addition most museums are reluctant to disclose their financial accounts. In the current research design, two variables could be seen as proxies of museum reputation: museum visibility and valence (Media tone toward museums) and museum promotion (free tickets distributed in the form of promotion). Museums that pursue promotional activities and aim at higher media visibility and positive coverage probably have higher reputation. However, future research should take into account museum reputation by measuring it in a more direct way.

The project controlled for the nature of governance of cultural organizations; namely, whether they are publicly or “privately” managed. Certain “private” museums, like the Benaki Museum, are more active in organizing public events that generate media publicity. However, the econometric analysis indicates that their “private” governance per se has a slight negative impact on their visitation. This might be attributed to the prevalence of large public museums, which account for the majority of visitors in our sample. This calls for further analysis of the differences between public and “private” organizations as it can offer useful explanations with regard to the agenda-setting process.

Furthermore, museum location does not register as a significant variable in the transfer of salience. Museums based in Athens, the largest metropolitan area in Greece, due to their location, might be expected to attract more visitors than museums based in rural Greece or in smaller cities. This was not necessarily the case. This factor also merits further examination. When disposable income is higher, it appears to enable the transfer of salience depicted by an increase in museum visitation. Whilst the negligible negative relationship between the number of tourists and museum visitation is contrary to our initial expectations, it can be attributed to their seasonal effects, which are already captured in the model by the month-dummies. The latter are positive and statistically significant for the months that tourism is more intense, as expected.

Our study applies the agenda-setting theory in the cultural environment and particularly in cultural organizations. The literature review indicates that no previous research has adopted this perspective in the context of agenda setting. Therefore, we argue that there is an opportunity to expand the applications of agenda setting into new uncharted venues, such as cultural markets, while generating useful data in the realm of cultural industries. Future studies could build on the findings of this paper and further examine other factors pertaining to characteristics that apply to particular cultural markets. Data derived from different countries analyzed in a comparative fashion can be more revealing about the relationship between media and public salience of museums. As agenda setting theory evolves, gathering data from different national settings provides useful explanations about the transfer of salience. In the current project, media visibility is examined at the second level of agenda setting in terms of affective attributes that provide additional insight into which frames are related to public salience. At a practical level, museums should defend their image and visibility, as positive coverage seems to affect visitation in a positive fashion. On the other hand, negative portrayals are correlated with diminished visitation. Future research can expand the applications of the agenda setting theory into the cultural market, a field that deserves such explorations, as it represents growing markets in the context of service-oriented economies.

Footnotes

Appendix

Table A2.

Correlations Table

	1	2	3	4	5	6	7
1. Visitation	1
2. Visibility	0.3842	1
3. Positive	0.3597	0.9073	1
4. Negative	0.1732	0.206	0.187	1
5. Log_promotion	0.0551	−0.3148	−0.2818	−0.052	1
6. Log_gdp_cap	−0.0388	0.0342	0.0074	0.0102	0.2463	1
7. Tourism (Number of foreigners)	0.1841	−0.0291	−0.0226	0.079	0.3068	0.0935	1

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the authorship and/or publication of this article.

Funding

The authors received no financial support for the research and/or authorship of this article.

Notes

Bios

Stelios Zyglidopoulos is a University Lecturer in Strategy at the Judge Business School of the University of Cambridge, and Fellow of Homerton College. He received his PhD in Strategy and Organization from McGill University in Montreal, Canada. His research interests include, among others, the evolution of corporate social responsibility, the evolution of corporate corruption in organizations, and the role of agenda setting in the commercial success of business and cultural organizations.

Pavlos C. Symeou is a Lecturer in Marketing at the Faculty of Management and Economics of Cyprus University of Technology. He obtained his PhD in Business Economics from Judge Business School of the University of Cambridge. His research interests include, inter alia, the role of technological convergence in corporate strategy, the marketing strategies of cultural organizations, and the interactions of firm performance with aspects of corporate social responsibility.

Philemon Bantimaroudis is an assistant professor in the Department of Cultural Technology and Communication at the University of the Aegean. He received his Ph.D. in communication from the University of Texas at Austin, USA. His research interests include media and cultural organizations, political and international communication.

Eleni Kampanellou is a Phd student in Human Geography at the Geography School of the University of St Andrews. She received an MSc in Social Research at the School of Social and Political Science at the University of Edinburgh and an MSc in Cultural Informatics at the University of the Aegean—Department of Cultural Technology and Communication. Her research interests include cultural spatial patterns and diversities as well as the function of agenda setting in the promotion of cultural organizations and industries.

References

Baltagi

H. B.

(2005). Econometric analysis of panel data. West Sussex, UK: John Wiley.

Bantimaroudis

Zyglidopoulos

Symeou

(2010). Greek museum media visibility and museum visitation: An exploration of cultural agenda setting. Journal of Communication, 60, 743-757.

Behr

R. L.

Iyengar

(1985). Television news, real-world cues, and changes in the public agenda. Public Opinion Quarterly, 49, 38-57.

Bloj

McCombs

(1974). A progress report on agenda-setting research. paper presented at the Association for Education in Journalism, San Diego, CA.

Carroll

McCombs

(2003). Agenda setting effects of business news on the public’s images and opinions of major corporations. Corporate Reputation Review, 16, 36-46.

Deephouse

(2000). Media reputation as a strategic resource: An integration of mass communication and resource-based theories. Journal of Management, 26, 1091-1112.

Dierickx

Cool

(1989). Asset stock accumulation and sustainability of competitive advantage. Management Science, 35, 1504-1510.

Feldman

J. M.

Lynch

J. G.

(1988). Self-generated validity and other effects of measurement on belief, attitude, intention, and behaviour Journal of Applied Psychology, 73, 421-435.

Fombrun

C. J.

(1996). Reputation: Realizing value from the corporate image. Boston, MA: Harvard Business School Press.

10.

Funkhouser

(1973). The issues of the sixties: An exploratory study in the dynamics of public opinion. Public Opinion Quarterly, 37, 62-75.

11.

Ghanem

(1997). Filling in the tapestry: The second level of agenda setting. In McCombs

Shaw

Weaver

(Ed.), Communication and democracy (pp. 3-14). Mahwah, NJ: Lawrence Erlbaum.

12.

Graber

(2001). Processing politics: Learning from television in the Internet age. Chicago, IL: University of Chicago Press.

13.

Herr

P. M.

Kadres

F. R.

Kim

(1991). Effects of word-of-mouth and product-attribute information on persuasion: An accessibility-diagnosticity perspective. Journal of Consumer Research, 17, 454-462.

14.

Higgins

E. T.

King

(1981). Category accessibility and information-processing: Consequences of individual and contextual variability. In Cantor

Kihlstrom

(Eds.), Personality, cognition and social interaction (pp. 69-122). Hillsdale, NJ: Erlbaum.

15.

Iyengar

Kinder

D. R.

(1987). News that matters: Television and American opinion. Chicago, IL: University of Chicago Press.

16.

Iyengar

Kinder

D. R.

Peters

M. D.

Krosnick

J. A.

(1984). The evening news and presidential evaluations, Journal of Personality and Social Psychology, 46, 778-787.

17.

Iyengar

Ottati

(1994). Cognitive perspective in political psychology. In Wyer

R. S.

Jr. Srull

T. K.

(Eds.), Handbook of social cognition, Volume 2: Applications (pp. 143-187). Mahwah, NJ: Lawrence Erlbaum.

18.

Kadres

F. R.

(1994). Consumer judgement and decision processes. In Wyer

R. S.

Jr. Srull

T. K.

Handbook of social cognition, Volume 2: Applications (pp. 399-466). Mahwah, NJ: Lawrence Erlbaum.

19.

Kiousis

(2004). Explicating media salience: A factor analysis of New York Times issue coverage during the 2000 U.S. presidential election. Journal of Communication, 54, 71-87.

20.

Kiousis

McDevitt

(2008). Agenda setting in civic development: Effects of curricula and issue importance on youth voter turnout. Communication Research, 35, 481-502.

21.

Lombard

Snyder-Duch

Bracken

C. C.

(2002). Content analysis in mass communication: Assessment and reporting of intercoder reliability. Human Communication Research, 28, 587-604.

22.

Lynch

J. G.

Marmorstein

Weigold

M. F.

(1988). Choices from sets including remembered brands: Use recalled attributes and prior evaluations. Journal of Consumer Research, 15, 169-184.

23.

McCombs

(2004). Setting the agenda: The mass media and public opinion. Cambridge, UK: Polity.

24.

McCombs

Shaw

(1972). The agenda setting function of mass media. Public Opinion Quarterly, 36, 176-187.

25.

Meijer

Kleinnijenhuis

(2006). Issue news and corporate reputation: Applying the theories of agenda setting and issue ownership in the field of business communication. Journal of Communication, 56, 543-559.

26.

Moon

S. J.

(2008). Agenda-setting effects as a mediator of media use and civic engagement: From what the public thinks about to what the public does. Unpublished Doctoral Dissertation, the University of Texas at Austin.

27.

Roberts

P. W.

Dowling

G. R.

(2002). Corporate reputation and sustained superior financial performance. Strategic Management Journal, 23, 1077-1093.

28.

Shoemaker

P. J.

Wanta

Leggett

(1989). Drug coverage and public opinion. In Shoemaker

P. J.

(Ed.), Communication campaigns about drugs: Government, media, and the public (pp. 67-80). Hillsdale, NJ: Lawrence Erlbaum.

29.

Stone

G. C.

McCombs

M. E.

(1981). Tracing the time lag in agenda-setting. Journalism Quarterly, 58, 51-55.

30.

Van Riel

C. B. M.

Fombrun

(2002). Which company is most visible in your country? An introduction to the special issue on the global RQ-Project Nominations. Corporate Reputation Review, 4, 296-302.

31.

Wanta

Y.-W.

(1994). Time-lag differences in the agenda-setting process: An examination of five news media. International Journal of Public Opinion Research, 6, 225-240.

32.

Winter

J. P.

Eyal

C. H.

(1981). Agenda-setting for the civil rights issue. Public Opinion Quarterly, 45, 376-383.

33.

Wyer

R. S.

Jr. Srull . (1986). Human cognition in its social context. Psychological Review, 93, 322-359.

34.

Zyglidopoulos

Phillips

(1999). Responding to reputational crises: A stakeholder perspective. Corporate Reputation Review, 2, 333-350.