The power of press releases and journalists’ decisions : a content analysis of the press releases issued by startups and the news articles using them

(1)

The Power of Press Releases & Journalists’ decisions: A content analysis of the

press releases issued by startups and the news articles using them

Alessio Maggiore

Supervisor: Pytrik Schafraad

Master’s Thesis- Corporate Communication

Graduate School of Communication

(2)

1

Abstract

Over the past decades, the academic literature investigating the media/journalism field and the public relations field have pointed out the fundamental interplay between the journalists working on news media and the public relations professionals working in organizations. Public relations professional need journalists in order to enable their organization to gain media reputation, while journalists need public relations professionals in order to gather news contents without having costs related to the search of news contents. This study uses news factor theory in order to investigate whether the news selection process made by journalists is influenced by certain characteristics of the press release. Moreover, this study utilize also churnalism theory in order to gather insights related to the way journalists use the press releases once they decide that those press releases will be part of the news agenda. The content analysis conducted in the study indicated that the presence of news factor and of the right main topic in the press release can improve its chances to be covered by the news media. Moreover, the findings of the study also highlight that there is a tendency among journalists to present in the news articles the same information presented in corporate press releases. Therefore, this study shows that, by comparing corporate press releases and the news articles using them, it is insightful to use news factor theory in order to measure the agenda-building power of the press release. In addition, this comparison also enabled this study to successfully apply churnalism theory in order to investigate how journalists used the original press releases in their news articles.

(3)

2

Introduction

Over the last few years, the academic literature about the relations between news media and

organization developed many concepts aimed to explain the interplay between these two actors. Past studies have shown that organizations seek media coverage in order to improve their reputation and, consequently, gaining new customers or reinforce their relationship with the ones that are already customers (Deephouse, 2000, Carroll, 2010, Westphal & Deephouse, 2011).

Therefore, organizations need to send to the news media information subsidies that fulfill journalists’ selection criteria. In this way, they can foster the chances of their information subsidies to be used by the news media. The literature analyzing these criteria used by the journalists in order to select the news contents highlights the potential influence associated to the presence of news factors in the original source (Galtung & Ruge, 1965; Harcup & O’Neill, 2001; Eliders, 2006; Maier & Ruhrmann, 2008; Schafraad, van Zoonen & Verhoeven, 2015). Moreover, one of these studies focusing on press releases pointed out the important role covered by the topic of the press release. Indeed, this study indicated that in many cases the focus of the press release can be seen as a news factor by journalists and, therefore, it can incentivize the journalists to cover that press release (Schafraad, van Zoonen & Verhoeven, 2015).

Other important concepts related to the way journalists and news media use the information sent by organizations are disclosed by previous studies. These studies have shown that copy-pasting parts of the press releases in the news articles is a practice widely used by journalists (Lewis, Williams & Franklin, 2008; Kroon & Schafraad, 2013). In addition, other findings support the assumption that journalists do take that information but they also reframe and expand them when developing the news article (Pander Maat & de Jong, 2013; Kroon & Schafraad, 2013).

Journalism and public relations seem to be dependent from each other (Gans, 1979) and the concepts highlighted by the academic literature can be used as indicators of this fundamental relationship.

Accordingly, this study will analyze press releases sent by organizations and the news articles related to them in order to gather useful insights about the elements fostering press release coverage. Moreover, it will also investigate the criteria used by the journalists of 3 different types of news media, when publishing a news article presenting information received from a press release. Thus, the aim of the study is to highlight the

(5)

4 importance of certain characteristics of the press releases and how news media and journalists select and use these press releases.

However, this study has an explorative design because it investigate just the field of startups and their press releases, so the results of this paper are limited to this field. This type of organization is chosen for this paper because the academic research about it is relatively scarce. It can be due to the fact that these are relatively young. Moreover, also the news media specialized in technology are rarely discussed in the academic discussion around media relations. Thus, this paper will investigate the practices used by this type of news media together with business and high-quality general news media.

The following section will discuss the theories utilized in this study, and the consequent research question and hypotheses.

Theoretical framework

Information subsidies and agenda-building.

The role of journalists in news media and PR practitioners in other organizations have been frequently described as very dependent from each other. On one hand, journalists need to select the news subjects that should be covered by the news organizations (Neijens & Smit, 2006; Cornelissen, Carrol & Elving, 2009). This process implies several costs for the news organizations and journalists because they need to find and check the news subject and the news sources. On the other hand, many other organizations hope to be part of the news agenda in a positive light because in this way they can earn media reputation (Deephouse, 2000). A good media reputation is a key element for organizations because it can enable them to create a good image of the organization from the public’s perspective, and also to get the publicity for corporate events or product and corporate-related news (Carroll, 2010; Westphal & Deephouse, 2011). Therefore, PR practitioners, taking into consideration the expenses that faced by the news organizations, try to influence journalists’ news selection process by providing them of information subsidies which describe news subjects related to their organization and are shaped according to journalist’s and newspaper

organization’s needs and requirements (Berkowitz & Adams, 1990).

In other words, PR practitioners in this process cover the role of news sources of journalists and they also reduce the costs for the news media organizations which do not need to research and find news subjects. Past studies have highlighted that the information subsidy can be presented in several ways such as press

(6)

5 conferences, interviews with CEO, telephone calls or press releases (Berkowitz & Adams, 1990). Moreover, press releases are described as one of the main factors influencing news media agenda building process (Seletzky & Lehman-Wilzig, 2010; Kiousis, Popescu & Mitrook, 2007). In other words, press releases represent one of the best ways utilized by organizations in order to obtain positive media attention. Press releases are defined as “printed news announcement with a focused message, disseminated (through Email, messenger, mail etc.) simultaneously to all or some news editorial departments and/ or journalists dealing with the subject area” (Seletzky & Lehman-Wilzig, 2010). Therefore, through the press releases,

organizations not only can publicize themselves but also the news story related to them (Schafraad, van Zoonen & Verhoeven, 2015).

These processes are very insightful when interpreted from the perspective of two important interconnected theories developed in communication literature which are agenda-building theory and agenda-setting theory.

On one hand, agenda setting theory argues that “the salience of elements on the news agenda influences their salience on the public agenda” (Carroll & McCombs, 2003). In other words, it implies that a news story that is prominent on the news agenda will also be prominently discussed in the public sphere. In addition, this theory also indicates the importance of the salience of the news story attributes (Carroll & McCombs, 2003). This represents the second level of agenda setting which takes into account specific elements of the news story that, as in the first level of agenda setting, are salient in both media and public agenda.

On the other hand, agenda-building is the step before agenda-setting and it refers to the way news media interact with other organizations in order to create issues that could stimulate the public opinion. Hence, agenda-building theory also refers to the activities utilized by organizations in order to influence public policy agenda with issues or news related to the organization (Schafraad, van Zoonen & Verhoeven, 2015). Another important concept linked to this theory is the agenda building power, which refers to the influence made by organizations on media agenda through the use of press releases (Berkowitz & Adams, 1990; Pander Maat & de Jong, 2013). It means that the information sent from organizations to journalists are vital for the process of agenda building. In addition, it also means that the success of the information subsidy

(7)

6 in being picked up by news media depends on its power which refers to the information subsidy’s ability to interest the journalist and influence his/her selection of the news.

As mentioned previously, press releases are one of the efficient among the information subsidies, but other studies have shown that just a limited number of the press releases sent by organizations to news media succeed in being selected by the journalists and eventually covered by the news media (Kroon & Schafraad, 2013; Schafraad, van Zoonen & Verhoeven, 2015). Moreover, past studies showed that the information subsidies sent by governmental agencies or non-profit organizations are more utilized by news media than those sent by other types of organizations (Morton, 1986; Berkovitz & Adams, 1990). Taking into account these findings and the insights disclosed by agenda building and setting theory, this study investigate the relationship between start-ups press releases and the relative coverage on different types of newspaper. Therefore, the following is the research question of this study:

RQ: What elements of startups’ press releases influence their coverage on news media specialized in technology, business news media, and high-quality general news media (1) and how these types of news media use the press releases (2)?

News factor theory and its effects

In order to gather insights regarding the news selection process made by journalists, this study will also take into consideration the news factor theory proposed by Galtung and Ruge (1965). In their study entitled “the structure of foreign news”, they investigated how some events happened outside Norwegian borders managed to be picked up by Norwegian press whereas others did not. The central objective of the study was to understand how an event becomes a news story, so they developed 12 news factors that can be important criteria for the selection of the news made by journalists (Galtung & Ruge, 1965; Harcup & O’Neill, 2001). In addition, they argued that the more an event manage to fulfill those criteria, the more it is likely to be picked in the news selection. (Galtung & Ruge, 1965; Harcup & O’Neill, 2001).

From this perspective, the news factors represent a list of certain characteristics that give value as news story for the journalist to an event. Therefore, this list facilitates the news selection process made by journalists because they will choose the events that possess and involve those characteristics rather than other events (Eliders, 2006; Maier & Ruhrmann, 2008; Schafraad, van Zoonen & Verhoeven, 2015). In other words, journalists consider certain events worthy to be inserted in the news agenda because the news factors

(8)

7 attributed to these events coincide with the characteristics searched by journalists in the news selection process (Schulz, 1976; Eilders, 2006).

Consequently, the presence or absence of these news factors in an event influence the degree of

newsworthiness journalists associate to a certain event or information subsidy (Harcup & O’Neill,2001; Eilders, 2006). Therefore, previous studies have described the newsworthiness of the subject as a judgment utilized by journalists on the pertinence of the news factor associated to an event or information subsidy with their selection criteria (Eilders, 2006; Shoemaker, 2006). This concept has also been described as “a set of selection rules which determine journalistic decisions” (Maier & Ruhrmann, 2008).

Over the last few decades, the list of the 12 news factor developed by Galtung and Ruge (1965) changed in many ways with additional news factors added or names and scope of certain news factors have been modified according to the relative study focus and the news media investigated by those studies (Harcup & O’Neill, 2001; Eilders, 2006). The news factors investigated in this study have been used by previous studies which are: surprise; controversy; impact-reach; positive and negative consequences; elite ad prominent persons and dynamics (Maier & Ruhrmann, 2008; Schafraad, van Zoonen & Verhoeven, 2015). These news factors have been chosen for this study because they are the most adaptable to the sector examined in this paper, which is the world of the startups.

Moreover, as pointed out by the literature regarding news factor theory, news factors have to be interpreted as elements of a constructed reality and not as actual characteristics of an event (Schulz, 1976). For this reason, this study will analyze the press releases published by startups which represent a constructed reality created by these startups. News media attribute to these press releases the role of news sources that provide them of news contents related to the organizations that publish them. Hence, this paper will investigate whether the presence or absence of potential news factors in startups’ press releases influence journalists’ news selection judgments. In addition, past research studies have highlighted that news media organizations and journalists are likely to not taking into consideration press releases that do not fulfill their selection criteria. Consequently, these press releases lack in newsworthiness according to journalists’ perspective (Hong, 2008; Aronoff, 1977).

Taking into account the concepts outlined by the literature, the following is the first hypothesis of the study:

(9)

8

H1: The more news factors are present in start-ups’ press release, the more it is likely to be covered

by the news media.

Field theory and press releases as news factor

This study will also try to shed a light on other factors that can affect journalists’ news selection and on the additional role covered by the focus of the press releases. Bourdieu’s field theory (1986) and its perspective will be adopted in this study in order to gather insights about the agenda building power of the press release and on the news selection made by journalists.

According to Bourdieu’s field theory, society is composed by numerous fields that are partially connected between each other and each field has its own rules which are based on the oppositions between the actors with different powers. These powers are the cultural power, the economic power and the social power (Benson, 1999). Moreover, these oppositions mirror the class division in society between dominant and dominated class, but also within the dominant class where the economical and political powers are the dominant powers and the cultural power is the dominated power. Therefore, they show the divisions within a field, where the economic and politic power are likely to be related to the dominant class while the cultural power is usually associated to the dominated class (Benson, 1999). Thus, from Bourdieu’s perspective, society can be described as built through a continuous contrast between economic and cultural capital. Since fields are partially connected, it also means that the fields that have more economic power, which are the economic and political field, exercise their power over other fields that are closer to the cultural power (Benson, 1999). Bourdieu also affirms that this hierarchy of power is present within each field, meaning that organizations with high economic power have higher position in field in comparison with an organization with high cultural capital but low economic capital. Therefore, organizations with high economic capital will take advantage of the power of this type of capital and will use it in order to gain cultural and social capital, which can help them to improve their position on the field. Whereas, organizations with high cultural capital will utilize it to gain economic and social capital and improve their position on the field. Sending to news media press releases is one of the strategies used by organizations to gain or reinforce their economic, cultural or social capital and their position in the field. This process can be defined as the power struggle in the field (Bourdieu, 1986). In addition, as argued by Weber (Benson, 1999), the class division within a field depends on the characteristic and the rules of the field, so it is possible that in some fields the cultural and

(10)

9 social powers are the dominant powers, rather than the economic and political power, because more relevant to those fields (Benson, 1999). On the one hand, in line with Weber's assumption, Davies (2000) argued that cultural and social capital are more important in the media/journalism field than the economic and political capitals. Indeed, this past study indicated that new comers and non-institutional sources in the

media/journalism field gain more symbolic capital than those institutions or actors with more economic and political power (Davis, 2000). On the other hand, in line with the arguments related to the dominant power of the economic and political powers, past studies have shown that the journalistic field, which is positioned as a subfield of the cultural production field, is highly dependent on the economic and political field (Benson, 1999). It means that news media try to obtain power through interactions with the most powerful fields. Moreover, given the particularity of news media, they can also influence the power relations within and between fields through their publications (Benson, 1999). In other words, the news media try to gain power and improve their position on their field through the publication of articles related to powerful actors or organizations positioned in other fields. These publications also influence the power-struggle in the fields in which the actors or the organizations involved are collocated, so news media can improve or worsen their position and their power in their fields.

Furthermore, also the information subsidies such as press releases have an important role in helping organizations to gain power in their fields. Indeed, past studies have indicated the key role covered by public relations activities in terms of assuring and improving the organizations’ position in their field (Ihlen, 2007). Taking into consideration the concepts outlined by field theory and the economic orientation of the “Billion Dollar Startups Club”, the following is the next hypothesis of the study:

H2: The higher position in the ranking, the higher the chances for the press release to be covered by

the news media.

Given that the fields are partially connected and that for news media organizations this characteristic is vital, this study will try to shed a light on another aspect of press releases that can influence journalists news selection.

Past studies have shown that the topic of the press release can cover the role of news factor according to the news media editorial policy (Schafraad, van Zoonen & Verhoeven, 2015). In other words, press releases based on a topic that is in line with news medium focus have more chances to be covered by

(11)

10 that news medium. For this reason and taking into account the innovative and technological aspects of start-ups corporations, the following is the next hypothesis of the study.

H3a: Press releases focusing on technology related information as main topic are more likely to be

covered by news media specialized in technology than other press releases with a different topic.

Furthermore, the literature shows that financial and economic information are one of the most important topic utilized by corporations in their press releases (Morton & Ramsey, 1994). This assumption is very important in order to analyze the press releases published by startups companies because these types of companies always try to attract new investors. More specifically, for startups companies is vital to gather venture capital from investors that are not part of the organization. Venture capital is defined as “investments by specialized venture capital organizations, called venture capital funds, in high-growth, high-risk, often high-technology firms that need capital to finance product development or growth and must, by the nature of their business, obtain this capital largely in the form of equity rather than debt” (Black & Gilson, 1998). Consequently, startups press releases could be a chance for those organizations to attract new investors by showing them information about company’s economic situation. Plus, it is likely that the economic and/or financial topic of those press releases will be considered by business news media as a news factor. The following is consequent hypothesis:

H3b: Press releases focusing on economic and financial information about the start-up as main

topic are more likely to be covered by business news media than other press releases with a different topic.

The literature also shows that in general newspapers, the news related to corporations are often reported in the sections of the newspaper dedicated to the economy and business (Doyle, 2006; Schafraad, van Zoonen & Verhoeven, 2015). It is due to the fact that the journalists assigned to this section by the news media are also specialized in the field related to that section. Therefore, journalists assigned to the economic section pick the news topics that in his view could be relevant for potential investors rather than just for the general public (Doyle, 2006; Schafraad, van Zoonen & Verhoeven, 2015). Despite this, startups’

technological nature pushes journalists to highlight this characteristic. Moreover, startups are often

presented as young technological companies and are often described through a distinction between them and richer and established technological firms such as Facebook, Twitter or Google. Consequently, it is likely that journalists assigned to the technology section of newspapers publish in their section news content related

(12)

11 to the startups, rather than those journalists assigned to the economic and financial section. From Bourdieu’s perspective, startups do not have a high position in the economic field compared to all the actors involved in that field, so their press releases are likely to be less newsworthy in the eyes of the journalists covering the economic/financial section. On the other hand, startups do have a high position in the technology field since the organizations with higher economic capital are also covered by other sections. Therefore, the journalists assigned to the technology section of newspapers are more likely to take into consideration and cover startups’ press releases. Accordingly, the following is the next hypothesis of the study:

H4: The startups’ press releases covered by general newspapers are more likely to be inserted in the

technology section of those newspapers than in the economic and financial section.

Churnalism

Another important concept highlighted by the journalism and PR literature that will be applied in this study is called “churnalism”. This concept refers to the decline of journalists’ activities of news gathering processes due to the great amount of press releases by other organizations to news media organizations (Davis, 2000; Lewis, Williams & Franklin, 2008; Jackson & Moloney, 2015). But it also refers to another tendency developed by many journalists that is deemed the outcome of the PR-isation of the journalistic world (Jackson & Moloney, 2008). This outcome is the copy-paste techniques utilized very often by journalists who present news stories as they were the authors of those news stories whereas they publish press releases, or big parts of them, sent by other organizations without editing or inserting additional information. (Davis, 2000; Jackson & Moloney, 2008; Seletzky & Lehman-Wilzig, 2010; Kroon & Schafraad, 2013). Indeed, churnalism is defined as “news routinely recycled from elsewhere, and the widespread use of other material that is rarely attributed to its source” (Lewis, Williams & Franklin, 2008). It is also defined as “the use of unchecked PR material in the news”(Jackson & Moloney, 2008). Moreover, Lewis, Williams & Franklin (2008) have found that a large proportion of contemporary news are partially copy-pasted. For this reason, this study will investigate the degree of overlapping information between the press releases covered by the news media and the news article related to them. Hence, the following is the fifth hypothesis of the study:

(13)

12

H5: When a press release is covered by the news media, it is likely that the related news article

published on the news media will have a high degree of overlapping information with the press release from which is originated rather than a slight overlap.

Contrastingly, other scholars have argued that journalists do not just copy or publish the press releases that organizations send to them, rather they select, investigate, expand and modify the information received (Pander Maat & de Jong, 2013; Kroon & Schafraad, 2013; Schafraad, van Zoonen & Verhoeven, 2015). Taking into account this perspective, this paper also try to check the presence of additional sources in the startups-related news stories published on the news media. Moreover, also the journalists’ tone toward the topic of the news story and the startup, and the presence of references to other startups not mentioned in the original press release will be analyzed. These variables could be useful indicators of the expanding and editing activities utilized by journalists when publishing a news story taken from a press release sent by another organization. Therefore, the next is the last hypothesis of the study:

H6: The more presence of additional news contents in the news story, the lower the degree of

overlapping information between the news story and the original press release.

Method section

This section of the study will provide detailed information about the sampling and data collection processes, the design of the study, the variables used and their reliability.

Sample & data collection

Firstly, this research will analyze press releases published by successful startups that are included in "The Billion Dollar Startups Club", a ranking published on February 2015 by The Wall Street journal in collaboration with Dow Jones venture source. Startups such as Airbnb, Uber, Pinterest, and Snapchat are at the top 15 of this ranking which includes all the startups that are valued more than $1 billion. Also several Asian startups are part of the highest part of this ranking but in many cases English news media cannot gather frequent and constant information due to the limited access to these websites or to language barriers, therefore they are excluded from this study. I selected 5 startups which are placed on the top 15 of “The Billion Dollar Startups Club” ranking, meaning that they are the startups with the highest venture-capital, therefore the most successful and valuable. The 5 startups are: Airbnb, Pinterest, Spotify, Snapchat and Uber. These startups were chosen among the other in the top 15 because of their popularity among the

(14)

13 general public and their industry which is the consumer internet. The consumer internet industry is deemed to be less specific than the other industry inserted in this ranking such as aerospace and defense or financial services. Therefore, given the low degree of specificity attributed to the consumer Internet industry, this study will analyze startups that are part of it.

Moreover, the study will analyze a total of 251 press releases published by these startups on their websites or blogs over the years 2012 – 2015. This period of time should be enough given that some of these startups were founded very recently and do not have many press releases published during the previous years. More specifically, 50 press releases for each startup were analyzed and coded, except for Airbnb that had one additional press release so it reached 51 press releases. Furthermore, Airbnb, Snapchat and Spotify had around 50 press releases spread over the 3 years analyzed in this study, whereas Pinterest and Uber published a higher quantity of press releases in their blogs over the last few years. It required a systematic sampling procedure in order to extract a fair sample of 50 press releases from the great quantity of press releases published by these companies over this period of time. The specific dates used to sample the press releases went from 1 January 2012 to 15 November 2015.

Secondly, the study aims to check whether the press releases published by the 5 startups were covered by 3 different types of news media which are technology news media, business news media and high-quality general news media. Technology news media focus on news story related to technology and the online newspapers analyzed for this type of news media are TechCrunch and Venture Beat. Both news media pay a great amount of attention towards the startups world and the news related to it. Business news media focus on news story relevant from a financial and economic point of view and they are represented in this study by Forbes and Reuters. Both newspapers are well-known and an important source for business news story. Lastly, high-quality general news media cover all types of news story and usually have several sections with a different focus such as technology, culture or opinion. The general newspapers analyzed in this study are The Guardian and The New York Times. Both have been selected because of their readership and reputation and because of their availability in English language. In addition, given that a great number of startups listed on “The Billion Dollar Startups Club” are American, it is important to analyze whether the proximity between the start-ups and the general news media influences the press release chances to be used by the news media.

(15)

14 These different types of news media need to be analyzed because of the particular characteristics of startups. Indeed, startups are an interesting subject for technological news media because of their nature, but they can also be relevant for business news media because of their economic growth and for general

newspapers because of the popularity among the general public and the accessibility of the product and/or service provided by these startups.

A total of 294 articles published by these newspapers have been collected and analyzed throughout this study. All the articles were gathered through the use of the search engine present on the news media websites. Moreover, some of these search engine (TechCrunch, VentureBeat, The NewYork Times) were based allowed to filter the search for periods of time. Therefore, it was sufficient to use as keyword the name of the startup and limit the results to the dates close to the day of publication of the press release by the startup. For example, if a press release was published by Uber on the 15th of January, the search keyword on these news media websites will be Uber and it will take all the articles related to Uber from the 12th January to the 18th January. In other words, days before and 3 days after the publication of the press release. The search engine available on the websites of Forbes, Reuters and The Guardian did not provide the possibility to base the search on the period of time the startups published their press release, so it required one or two additional keywords that refer to the main subjects of each press release. Unfortunately, I had to apply this unorthodox approach for the data collection of this study because databases such as LexisNexis did not provide the articles published by TechCrunch, VentureBeat, Forbes and Reuters.

Design

A comparative design is applied to this study since press releases and news media articles are compared in order to gather insights about the degree of overlapping information they share between each other. Moreover, through this comparison it is possible to understand whether elements embedded in the press releases, such as news factors, foster the chances for a press release to be covered by news media.

A second comparison that is present in this study is related to the differences on the use made of the press releases by general newspaper and magazines, business news media and news media specialized in technology. Some difference among these three types of media is expected because specialized media mainly publish news regarding a sector whereas general newspaper try to take the best news from different sectors.

(16)

15 Moreover, this study also uses an explorative design focused on the interplay between journalists of general and specialized news media and the PR professionals of these big startups. Therefore, a content analysis of the press releases published by the 5 startups analyzed in the study and of their relative coverage on two general newspapers and 4 specialized newspapers (technology and business ) over the years 2012– 2015 will be conducted.

Measures

The codebook is divided into two parts, the first is dedicated to the analysis of the press release while the second is aimed to analyze the press release succession on the six news media considered in this study. Furthermore, the part dedicated to the press releases consisted in fifteen variables. The second part consisted in 7 variables repeated for each news media, excepted for the high-quality general news media (The

Guardian and The New York Times) which have an additional variable indicating the section including the news article. In addition, for TechCrunch, Forbes and The New York Times these variables are repeated a second time because in some cases they covered twice the same press release.

The research literature about news factors theory provides a variety of tools that are appropriate for the purposes of this paper. For this reason, the codebook developed in order to analyze the press release is based on the codebook created by Maier & Ruhrmann (2008) but with some modification required by the subjects of this study. In addition, other concepts uncovered by previous studies (Eliders, 1997; Schafraad, van Zoonen & Verhoeven, 2015) are part of the codebook. The variables used to measure the presence of news factors in the press release originate from these past studies. Only 8 news factors have been selected for this study because they were most appropriate in order to analyze startups’ press releases. These 8 news factors are: 1) controversy; 2) surprise; 3) range, 4) positive consequences, 5) negative consequences, 6) reference to elite people, 7) reference to prominent people and 8) dynamics. All these news factors were measured through a scale that ranged from 0 (not present) to 1 (present), therefore they were all binary variables.

Another independent variable important for the purposes of the study is the main focus of the press release. This variable includes 5 response categories which are: 0) Other; 1) product and service information (focusing on technological aspects); 2) news about the organization (focusing on financial and economical content); 3) events (focusing on events involving the organization); 4) Response to crisis (focusing on

(17)

16 providing an explication and/or apologies in response to an organizational crisis). Consequently, this variable has a nominal measurement level.

In addition, another independent variable is the presence of additional content in the news articles which is a binary variable, so it ranged from 0 (not present) to 1 (present).

Furthermore, also the variables related to the succession of the press release on the news media are very important because cover both the roles of independent and dependent variables in this study. All the variables focusing on the succession of the press releases on the news media analyzed were binary variables, therefore they ranged from 0 (not present) to 1 (present).

Another dependent variable used in order to the hypothesis of the study is the degree overlapping information between the press release and the relative news article when the first is covered by the news media. This variable has 5 response categories which are: 0) No overlap (0% of the press release used); 1) slight overlap (25% of the press release used); 2) moderate overlap (50% of the press release used); 3) high overlap (75% of the press release used); 4) complete overlap (100% of the press release used). Consequently, this variable has an interval measurement level.

Coder training and reliability

The codebook was tested by the coder through the analysis of 10 press releases and their related news articles before starting the actual coding process. In order to check the intra-coder reliability, the coder analyzed a second time 25 press releases and their follow-ups on the news media selected randomly after completing the actual coding. The Cohen’s Kappa and the percentage agreement were calculated, but the some of the scores related to the first measure were not precise due to the low number of cases used to calculate it. Despite this, the scores related to the Cohen’s Kappa indicate that the variable are reliable expect for those measuring the degree of overlapping information for the Guardian and the New York Times. Indeed, all the variables, except for the variables related to the degree of overlapping information for the high-quality news media, have a Cohen’s Kappa higher than .5. In addition, the percentage agreement detected by the reliability analysis ranged from 88 to 100 percent, which can be considered as another indicator of the good reliability. Although the Cohen’s Kappa was not calculated for some variables, this analysis together with the percentage agreement provide show that the variables have a sufficient reliability

(18)

17 given the explorative aim of this paper (Neuendorf, 2002, p.145; Schafraad, 2009, p.45). For more

information about the results of the reliability analyses have a look at the Table 1 in the appendix A.

Results

Hypothesis Testing (H1 & H2)

The first hypothesis of the study argues that the more news factors are present in start-ups’ press release, the more it is likely to be covered by the news media. Moreover, the second hypothesis of this paper posits that the higher is the position of the startup in the ranking , the higher are the chances for the press release to be covered by the news media. In order to test these hypotheses, a multiple regression analysis is conducted by using the succession of the press release on all the news media analyzed as dependent variable. However, a variable computed through the sum of the news factors covered the role of independent variable together with another variable mirroring the position occupied by the startups on the “Billion Dollar Startups Club” ranking. The latter variable was created by recoding the variable “company mainly related to the press release” and, consequently, by assigning a decreasing order that mirrors the ranking position to the response categories of that variable. So, the first category to the startup with the lowest position on the ranking while the last category is attributed to the startup with the highest position on the ranking. The regression model is significant, F(2,244)=62,776, p<.001. Therefore, this regression model effectively predicts whether the presence of news factors in a press release and the ranking position attributed to the startup can influence its chances to be covered by the news media. This regression model, based on the presence of news factors on the press release and the startup position on the ranking, predicts 34% of the variation on the succession on the news media (R2=.34). News factors total, b*(0,48), t=8.94, p<.001,

95%[0.39,0.62], and position on the ranking, b*(-0,34), t=-5.93, p<.001, 95%[-0.45,-0.23], have a significant moderate association with the succession on the news media.

On one hand, for every news factor present in a press release while holding the variable referring to the ranking constant, the chances of the press release to be covered by the news media increase by 0.507. Consequently, the first hypothesis arguing that the more news factors are present in a press release issued, the higher are its chances to be covered by news media is supported. On the other hand, for every unit increases on the ranking variable while holding the presence of news factors constant, the succession on the news media decreases by .338. According to the second hypothesis, a higher position on “The Billion Dollar

(19)

18 Startups Club” ranking will predict a higher succession on the news media, but the findings did not support this hypothesis. Indeed, these findings significantly showed that the lower is the position on this ranking, the higher are the chances of the press release to be picked up by news media.

Hypothesis testing (H3a)

In order to test the hypothesis 3a, another regression analysis is required. Hypothesis 3a argues that startups’ press releases presenting technology-related information as the main topic are more likely to be covered by the news media specialized in technology-related news content. Thus, a variable consisting in the succession on all the technology news media was computed and used as dependent variable. Whereas another variable representing the presence of a main focus associated to technology-related information was created by recoding the variable main focus of the press release into a dichotomous variable. This variable covers the role of independent variable in the regression analysis.

Also in this case, the regression model is significant, F(1,249)=68,357, p<.001. Therefore, the model predicts whether the presence of a main focus associated to technology-related information in the press release can influence its succession on the news media specialized in technology-related news. The regression model predicts the 21.5% of the variation on the succession of the press releases on the technology-related news media (R2=.21). Technology-related main focus, b*(0,82), t=8,27, p<.001,

95%[0.62, 1.01], has a significant moderate association with the succession on the technology-related news media. When a press release presents a technology-related main focus, it increases the chances of the press release to be picked up by technology-related news media by .464. Consequently, these findings support the hypothesis 3a of this study since it posited that press releases with a main focus related to technology information would have higher chances to be covered by technology news media than those press releases with a different main focus.

Hypothesis testing (H3b)

The hypothesis 3b posits that startups’ press release presenting an economic/financial-related

information as main focus are more likely to be covered by news media specialized in business news content. Therefore, a variable formed by the succession on all the business news media was computed and then used as dependent variable in this regression model. Whereas, a variable representing the presence of an

(20)

19 the press release was recoded into a dichotomous variable including in one category all the press releases focusing on news related to the organization and on responding to organizational crisis whereas the second category includes all the other press releases. This variable, which represents the presence of an

economic/financial focus in the press releases, is used as independent variable in the regression model. The regression model testing the hypothesis 3b is significant, F(1,249)=4,046, p<.05. Thus, the regression model predicts whether the presence of an economic/financial-related focus on the press release can influence its succession on business news media. The regression model predicts only the 1.6% of the variation on press releases’ succession on business news media (R2

=.01). Economic/financial main focus, b*(0,12), t=4,65, p<.05, 95%[.003,0.29], has a very weak association with the succession of the press releases on the news media specialized in business news content. When a press release presents an economic/financial-related main focus, it increases the chances of the press release to be covered by business-related news media by .126.

According to these findings, the hypothesis 3b is accepted since it proposed that press releases with a main focus related to economic/financial information would have higher chances to be picked up by business news media than the other press releases with a different main focus.

Hypothesis testing (H4)

Hypothesis 4 posits that the press releases covered by high-quality general newspapers are more likely to be inserted in the technology section of those newspapers than in the economic and financial

section. Thus, a paired sample t-test is conducted in order to compare the number of news articles covered by the high-quality general news media that were inserted in the technology section with those inserted in the economic/finance section. Firstly, all the variables section of the general news media were recoded into two dichotomous variables respectively related to technology section and economic/finance section. Then, only the press releases with follow-ups on general news media were selected and, consequently those variables were compared through the paired t-test analysis.

This analysis shows that there is a significant difference in the scores for the section technology in the Guardian (M=.94, SD=.23) and the section economy/finance in the Guardian (M=.06, SD=.23)

(21)

20 In addition, the analysis shows that there is a not significant difference in the scores for the section technology in the first article of the New York Times (M=.54, SD=.51) and the section economy/business in the first article of the New York Times (M=.27, SD=.45) conditions; t(21)=1,45, p=162.

The analysis related to the sections attributed by the New York Times when covering with a second article a press release issued by a startup did not meet the requirements for a t-test. It is due to the low number of observations (N=3).

On one hand, the fourth hypothesis of this study, which expects a higher number of news articles inserted in the technology section rather than in the economy/finance section by high-quality general news media when covering startups’ press releases, is accepted regarding the Guardian. On the other hand, this hypothesis is not supported by the findings regarding the New York Times, so it cannot be accepted for this news medium. Consequently, the fourth hypothesis of the study is partially confirmed.

The fifth hypothesis of the study argues that when a press release is covered by the news media, it is likely that the relative news story published on the news media will have a high degree of overlapping information with the press release from which it is originated rather than a low degree of overlap. Thus, several crosstabs are conducted in order to compare the number of news articles with a high degree of overlapping information associated to the original press release with those news articles that have a slight/moderate degree of overlapping information with the original press release. In order to run these analyses, the variables measuring the degree of overlapping information between the press release per news outlet, and the succession on each news outlet are utilized in the cross-tabulations. The following table will outline the results indicated by these analyses.

Table 2. Overview of the degree of overlapping information between press releases and news articles.

Overlap TC1 TC2 VB Forbes1 Forbes2 Reuters Guardian NYT1 NYT2

< 25 4 (4,3%) 3 (37,5%) 4 (5,4%) 3 (8,6%) 1 (50%) 1 (4,5%) 2 (5,8%) 1 (4,9%) -

26 – 50 18 (19,4%) 2 (25%) 12 (16%) 7 (20%) - 3 (13,6%) 9 (26,5%) 5 (22,9%) 1 (33,3%)

51 - 75 31 (33,3%) 1 (12,5%) 28 (37,3%) 14 (40%) 1 (50%) 10 (45,5%) 14 (41,2%) 13 (59,1%) 1 (33,3%)

76 - 100 40 (43%) 2 (25%) 31 (41,3%) 11 (31,4%) - 8 (36,4%) 9 (26,5%) 3 (13,6%) 1 (33,3%)

n 93 8 75 35 2 22 34 22 3

(22)

21 As it is shown by table 2, the majority of the news articles have a high or complete overlap of

information with the original press release. Indeed, the 76,3% of news articles published by TechCrunch, which are based on startups’ press releases, have a high or total overlap of information with the original press release while just 23,7% have a slight or moderate degree of overlapping information with the press release. The same tendency is indicated by the news articles published by Venture Beat, where 78,6% of the news articles have a high or total overlap of information with the press release whereas just 21,4% have a slight or moderate overlap with the press release. In addition, also the 71,4% of the news articles published by Forbes have a high or total overlap of information with the original press release whereas just 28,6% of them have a slight or moderate degree of overlapping information with the press release. Moreover, 81,9% of the news articles published by Reuters have a high degree of overlapping information with the press release whereas only 18,1% of them have a slight or moderate overlap with the original press release. This tendency was also repeated in the high-quality general news media. Indeed, the 67,7% of the articles published by the Guardian have a high or complete overlap of information with the press release while only 32,3% of them have a slight or moderate overlap with the press release. The same applies for the news articles published by the New York Times, indeed 72,2% of the news articles published by the New York Times have a high or complete overlap with the original press release while 27,8% of them have a slight or moderate overlap of information with the original press release. According to these findings, news articles with an high or complete overlap with the press release are significantly more frequent than those having a slight or moderate overlap of information with the press release, given that the difference between the news articles with high/complete overlap and those with slight/moderate overlap is higher than 20%. Therefore, these findings support the hypothesis 5 of this study.

Contrastingly, the analyses related to the news articles published by TechCrunch when covering for the second time the same press release showed that 62,5% of them have a slight or moderate overlap with the press release whereas 37,5% of them have a high or total overlap with the press release. Also in this case the difference between the degrees of overlap is higher than 20% and, therefore, significant. Moreover, the results related to the news articles published by Forbes and the New York Times when covering for the second time the same press release did not provide a significant difference between the high/complete degree of overlapping information and the slight/moderate degree of overlapping information. Thus, the findings

(23)

22 related to these news outlets do not support hypothesis 5. Consequently, the fifth hypothesis of the study is partially confirmed.

In order to test the sixth and final hypothesis of the study another regression analysis is conducted. The hypothesis 6 posits that a news article that presents additional information is likely to have a slight degree of overlapping information with the original press release. Thus, a variable including the degree of overlapping information of each news media was computed and used as dependent variable. Whereas, a variable created by summing all the variables representing the presence of additional information in the articles for each news media was computed and utilized as independent variable in the regression model. The regression model is significant, F(1,249)=3697,53, p<.001, and it predicts the 93.6% of the variation on the degree of overlapping information (R2=.93). Additional information total, b*(0,97), t=60,80, p<.001, 95% [2.01,2.14] has a very strong association with the degree of overlapping information. Thus, when a news article presents additional information, the news article increases its degree of overlapping information with the original press release.

The sixth hypothesis of the study argued that the presence of additional information on the news articles will predict a slight degree of overlapping information between the press releases and the news articles, but the findings indicate the opposite. Consequently, the last hypothesis of the study is rejected.

Conclusion and discussion

All the hypothesis were aimed to answer the research question: What elements of startups’ press

releases influence their coverage on news media specialized in technology, business news media, and high-quality general news media and how these types of news media use the press releases?

The first part of the research question is answered by testing the first 3 hypotheses of the study while the second part of the research question is tackled by the last 3 hypotheses.

Firstly, the assumption that the presence of news factors in the press release can influence journalists’ news selection processes and, consequently, the press releases succession on news media is supported by the findings related to the hypothesis 1. Thus, journalists do use the of presence of news factors as selection criteria. It means that press releases that do not present any news factor lack in newsworthiness according to journalists’ perspective, so they do not satisfy journalists’ selection criteria. These findings are

(24)

23 in line with those outlined by previous studies (Aronoff, 1977; Harcup & O’Neill,2001; Eilders, 2006; Hong, 2008; Schafraad, van Zoonen & Verhoeven, 2015). According to these findings, more news factors are present in a press release, higher are its chances to be picked up by the news media. Therefore, organizations that want to use their press releases in order to gain media coverage, which can result in media reputation, need to include as many news factors as possible in the press release in order to increase its chances to be picked up by news media.

Secondly, the findings related to the second hypothesis do not support this hypothesis but they showed that the position of the startups on “The Billion Dollar Startups Club” ranking does have an influence on the succession of startups’ press releases on all the types of news media analyzed in the study. On one hand, the findings indicated that press releases issued by the startups with a lower position on the ranking are more likely to be covered by the news media than those with higher position on the ranking. Thus, these findings highlight the fundamental role covered by press releases in those strategies aimed to assure and improve organizations’ position in the field. Therefore, the findings confirm the assumptions on the decisive role covered by press releases proposed by Ihlen (2007). On the other hand, the assumption that startups with a higher position on the ranking are more likely to be picked up by the news media is not supported by the findings. It could be due the fact that the startups with a higher position on the ranking, and therefore with more economic capital, published many types of news contents on their blogs and with higher frequency than those with lower economic capital. Therefore, the wealthier organizations published more news content but the majority of these news content is not relevant by journalists. From field theory

perspective, these news contents not considered by the journalists might be aimed to improve the cultural and social capital through publishing news contents more relevant for customers, since these startups are those with greater economic capital. Contrastingly, the organizations with a lower position on the ranking published news contents less frequently and they were more relevant from journalists’ perspective. These findings confirm the ones outlined by Davis (2009), which indicated that in the media/journalism field the cultural and social powers are more relevant than the economic and political power. Therefore, according to field theory, these startups by publishing these press releases that are relevant to journalists try to

(25)

24 Thirdly, the hypothesis 3a, which argues that press releases with the main focus related to

oriented information are more likely to be covered by news media specialized in technology-oriented news content than those press releases with a different main topic, is supported by the findings. Similarly, also the hypothesis 3b which affirmed that press releases with the main focus related to

financial/economic information are more likely to be covered by news media specialized in business news stories than those press releases with a different main topic, is supported by these findings. Thus, these results confirm the findings highlighted by Schafraad, van Zoonen and Verhoeven (2015), indicating that the focus of the press release covers the role of news factor if it is aligned with the editorial policy of the news media. Consequently, the focus of the press release is a very important element of this type of information subsidy and it can be directed accordingly to the coverage needed by the startups. For example, startups that need to gain new customers or to affirm themselves in the technology industry should apply to their press releases a main focus oriented to technological information. In this way, the press releases will have more chances to be picked up by the news media specialized in technology news stories and, therefore, to gain new customers and reputation. In addition, startups that need to attract venture capital funds, in order to gain economic power and to help their growth, should attribute to their press releases a main focus oriented to economic information aimed to capture the interest of the venture capital organizations. By doing so, the press releases will have higher chances to be covered by news media specialized in business news content, which are the most appropriate outlets to showcase the economic potentialities of the startup. In addition, they are also an important source of information for venture capital funds.

The first three hypotheses pointed out the importance of the presence of news factors and of the right topic in startups press releases in order to gain media coverage and the consequent media reputation. These elements of the press release can really influence the selection process used by journalists because these elements are used as selection criteria by the journalists themselves. The last three hypotheses answer the second part of the research question.

Firstly, the hypothesis 4 is partially supported by the findings of this study. It affirms that, when a press release is picked up by high-quality general news media, it is more likely to be published on the technology section rather than on the economic/financial section of those news media. This assumption is supported by the results related to the Guardian but it is not by those related to the New York Times. On one

(26)

25 hand, it is likely that the Guardian does select the section to include the news stories related to these startups accordingly to their identity of young, technological companies. On the other hand, this characteristic treats of the startups do not seem to correspond with the criteria used by the editors and journalists of the New York Times when choosing the section to include the news articles related to these types of organization. A reason for these findings might be that the New York Times applies different criteria, when selecting the right section for the articles related to these organizations in comparison with the Guardian. These criteria could be more based on the focus of the news article than on the identity or the industry of the startups. Three cross tabulations were conducted in order to check this possibility. These crosstabs checked whether there is an association between the topic of the press releases and the section in which news article based on that press release is inserted on the high-quality general news media (The Guardian, and first and second news articles by The New York Times). The crosstabs indicated that there is no association between the section selected by the Guardian and the main topic of the press release, r =-.01, p=.925. In addition, the analyses showed that there is a weak not significant association between the section selected by the New York Times, when covering the press release for the first time, and the main topic of the press release, r = .21, p =.353. Furthermore, there is a very strong not significant association between the section selected by the New York Times, when covering the press release for the second time, and the main topic of the press release, r = .94, p =.212. Therefore, these results show that there are not significant associations between the section selected by these news media and the main topic of the press releases. Despite this, these results highlight the fact that the Guardian do not consider at all the topic as criteria in order to select the section where to include the news articles based on startups’ press releases, given that there is no association between these two variables. Contrastingly, the results show that the New York Times uses criteria more based on the topic of the press release when selecting the section where to include the first or second article using a press release, than the Guardian. Future studies should further investigate the relationship between topic of the press release and the section of the news media in which the news article based on the press release is included.

Secondly, also the results related to the hypothesis 5 partially support it. This hypothesis argues that when a press release is covered by the news media, it is likely that the news story published on the news media will have a high degree of overlapping information with the press release from which it is originated

(27)

26 rather than a low degree of overlap. The results for the majority of the news media analyzed in this study (TechCrunch, Venture Beat, Forbes, Reuters, the Guardian and the New York Times) support this

assumption. Therefore, the findings confirm the assumption outlined by previous studies (Lewis, Williams & Franklin, 2008) indicating that a large proportion of contemporary news are partly copy-pasted.

Contrastingly, the findings related to the news media presenting a second additional article for the same press release (TechCrunch, Forbes and The New York Times) do not support this hypothesis. Indeed, the findings related to Forbes and The New York Times, when covering with a second article the same press release, highlight there is not a significant difference between the high/total degree of overlapping

information and slight/moderate degree of overlapping information for these news outlets. It may be due to the limited amount of observations for these news media covering for the second time a press release. Moreover, the findings related to TechCrunch, when covering with a second additional article the same press release, indicate that there is a significant difference between the high/total degree of overlapping

information and slight/moderate degree of overlapping information for this news media but it is the opposite of the one expect by the hypothesis. Therefore, the majority of the articles published by TechCrunch, when covering with a second news article the same press release, have a slight/moderate degree of overlapping information with the original press release. Thus, this news outlet applies an high or total overlap with the original press release when covering it with just one news article, while it is possible that, when this news medium covers the same press release with a second news article, it applies a slight/moderate overlap with the original press release in order to provide a different information from the first news article published. In other words, this news outlet could prefer to use a slight/moderate degree of overlapping information with the press release when covering it for the second time in order to provide different information about the same news content of the first news article to the readers. In this way, the news article provide a different angle when publishing an additional second article about the news contents used in the first article and in the original press release.

Thirdly, hypothesis 6 is not supported by the results. This hypothesis argues that news articles that include additional information are likely to have a slight degree of overlapping information with the original press releases. The findings indicate that the higher is the presence of additional information on the news articles, the higher is their degree of overlapping information with the original press release. Although these

(28)

27 findings do not support the last hypothesis, they do support the assumption of previous studies. More

specifically, these results are in line with the assumption that journalists do not just copy and publish the press releases sent by organizations, rather they select, expand and modify that information (Pander Maat & de Jong, 2013; Kroon & Schafraad, 2013; Schafraad, van Zoonen & Verhoeven, 2015). Indeed, according to the findings of this study, journalists use to insert additional information in their press releases even if they have already inserted the majority or the whole press release in the news article.

The last three hypotheses shed a light on some of the criteria and techniques used by the journalists when creating and publishing news articles using press releases sent by organizations as a news source. On one hand, the findings indicate that copy-pasting techniques are used by journalists are very common among all the types of news media analyzed in this study. On the hand, reframing and expanding techniques used by the journalists are as common as the previous techniques. Moreover, the results also partly confirmed the possibility that journalists use criteria based on the characteristic of the company related to the news article when choosing the section of the news media to include it.

In conclusion, elements of the press releases, like the presence of news factors and the main focus of the press releases together with the copy-pasting and reframing techniques used by the journalists, need to be investigated because they are important factors that can help the discussion in the media relations literature. More specifically, they could be introduced on the sense-making model presented by Corneilissen, Carrol and Elving (2009) as factors influencing news media in the negotiation process negotiation process. In addition, the relationship between news factor theory and agenda-building theory is very insightful and needs to be further investigated. Indeed, this study highlighted that elements such as the presence of news factors in the press releases and the main topic of the press releases can influence journalists news selection process. In other words, these elements of the press release increase its agenda-building power, which is the ability to influence the public policy agenda with news contents related to the organization. According to agenda-setting theory, the news contents that are prominent in the news agenda are will also be discussed in the public sphere (Carroll & McCombs, 2003). This study pointed out that press releases including news factors and a right main topic have a good agenda-building power and effectively help organizations to gain more reputation.

(29)

28 In addition, the findings related to churnalism and reframing techniques also have to be taken into consideration in future studies because these techniques are widespread among journalists and can shed a light on the relationships between companies and news media.

This study presents some limitations related to the sample used. Indeed, the lack of news media such as TechCrunch, Venture Beat, Reuters and Forbes on databases like LexisNexis and the time-consuming processes of data collection from the websites of those news media affected the number of startups analyzed in the study. Initially, this paper aimed to analyze 8 startups positioned on the top 15 of “The Billion Dollar Startups Club” ranking but due to the long times needed by the data collection process, the study focused on just 5 startups instead of 8 of them.

Other limitations are related to the reliability of the variables utilized in this study. Indeed, an insufficient number of cases was used in order to provide a precise measure of reliability. Therefore, future studies using the codebook developed for this study, which is based on the one created by Maier &

Ruhrmann (2008) and expanded with concepts outlined by other studies (Eliders, 1997; Schafraad, van Zoonen & Verhoeven, 2015), should further test the validity of this codebook by using a bigger sample. Despite this, the codebook found relevant findings in this study but variables such the degree of overlapping information between press releases and news articles, and the presence of additional information need to be further tested due to the explorative design of this study.

Finally, this study outlines relevant findings related to the elements of press releases that influence journalists’ selection processes, and to the way journalists use these press releases when writing a news article about the related the information. Nonetheless, the results of this study cannot be generalized since they are limited to the context of startups.