PART II

EMPIRICAL PAPERS

PAPER 4

The daily gossip: looking at the content of media gossip from an evolutionary perspective

Abstract

Media Gossip is omnipresent in our current societies. Our interest in this modern form of gossip can, in an evolutionary perspective, be explained by the Learning Hypothesis as being a learning strategy to acquire information about behavior strategies others have tested. According to another hypothesis, the Parasocial Hypothesis, we regard celebrities as being part of our social network because we encounter them on a frequent basis through media images. We value gossip about these people’s lives for the same reasons we value gossip about real members of our social network: because it can help us solve adaptive problems.

In this paper I analyze the content of Media Gossip, focusing on print media and more specific gossip magazines. Gossip magazines are easy to access and rich sources of Media Gossip. In co-operation with ten graduate students I conducted a content analysis of 852 stories from three Belgian gossip magazines. Results show that these gossip magazines report about a variety of gossipees, both celebrities and unknown people. Further, all topics in these magazines can be classified in a list of universal occurring adaptive traits/behavior categories. Third, I noticed that celebrities are gossiped about for a broader range of topics, while unknown people are only subject to gossip topics that have clear fitness-relevance. Fourth, clear differences appear for male and female gossipees. Female gossipees are more likely to be subject of gossip topics concerning physical appearance while male gossipees pop up more in gossip stories about their wealth status and job issues. Fifth and last, the three gossip magazines are heavy on pictures (94% presence), which in my opinion is to deal with reliability problems gossip magazines face.

1 Introduction

In this paper I analyze the content of Media Gossip, focusing on print media. I do not focus on changes in the content of Media Gossip, but rather on the opposite. Taking an evolutionary stance, the scope of this paper is to look at Media Gossip topics that can be expected to remain the same over time and that are also expected to occur cross-culturally.

Gossip magazines are easy accessible sources of Media Gossip, and will be used for this analysis. However, Morreall (1994) once said that all gossip is private and occurs behind-the scenes, certainly in the eyes of those who gossip. That is why “gossip-columns” are misname, they are not about gossip, Morreall says. Still, I think Interpersonal Gossip and Media Gossip indeed are different as verbs, but not as nouns, as Post (1994) as well suggested.

1.1 Media Gossip as a rumor-like act

As a verb Interpersonal Gossip is private, because these gossip conversations take place between people who know each other. Moreover, between people who know and trust each other. I trust you to convey you this information that other people don’t know yet. Although some researchers (Klinghoffer, 1994) say that most gossip is untrue, unchecked information, and even when it is true, this truth has been manipulated in one way or another, other researchers (Anderson, 1995; Bergmann, 1993; Gelles, 1989; Smith, et al, 1999) state that Interpersonal Gossip can be regarded as reliable information.

This reliability issue is what gossip distinguishes from rumors. Rumors are spread to a broad public (Bergmann, 1993). So the setting of people who know, and more importantly, trust each other, is not present in the spread of rumors. That is why rumors lack more reliability than gossip, since the latter occurs in trust-relations (Smith, et al, 1999). “The elements that determine whether information is gossip are the subject, the producer, the recipient, and the relationship between them.” (Smith, et al, 1999: 122).

Media Gossip is not as reliable as Interpersonal Gossip. The reason why Media Gossip is less reliable than Interpersonal Gossip draws from the difference in both phenomena as an act, which I just discussed. Gossip senders and receivers do not know each other, the gossip message is spread to a broad audience, and this action lacks the trust-based setting wherein Interpersonal Gossip is shared.

Media Gossip senders (journalists) seem to be aware of this problem and (un)consciously try to offer solutions to the ‘reliability problem’. They present their Media Gossips in (1) interview style, (2) add pictures, and (3) not always hunt for scoops, but cover stories other media sources have already released, because multiple resources increase credibility (see also chapter 1).

Presenting their stories in interview-style, media sources want to present the information ‘as if’ the gossipee tells you the story. It could be argued not to label interviews with celebrities and public unknowns as ‘Media Gossip’. In chapter 1 I clearly claimed that self-talk is related to but different from gossip. Still, although interviews with media gossipees present the information ‘as if’ the gossipee is telling this himself or herself, this is actually not the case. As I stress, this happens ‘as if’ the gossipee is present and tells you the information first-hand. Actually it is the media journalist who shares this information with you about the gossipee, and not the gossipee himself or herself. The reason media journalists do this is to enhance credibility. Presenting the information ‘as if’ the gossipee is present increases the reliability of the message. “See, he or she told me this with these exact words”. They present the information in interview style just to make you believe the message more.

Next, gossip magazines are heavy on pictures (Bird, 1992). Pictures are added as a kind of proof to show you “this really happened”. We do not lie about Jennifer Anniston and Brad Pitt still wearing their wedding rings after they split a month ago, see for yourself on the pictures: they really do. Pictures might be the best weapon to battle the reliability problem. With their interview-styles, journalists can still easily lie or slightly manipulate the information. With pictures it is less easy to lie. It is still possible, they can come up with old pictures or manipulate the images, but this is still not as easy as lying with words. Moreover, since communication has such a long history of many generations, our ancestors were already faced with problems of false information. Facing these problems generation after generation for at least 100,000 to maybe millions of years (as long as language exists see chapter 3), has most probably shaped our brains to be skeptical towards the believability of language messages. Pictures only exist since the 19th century, and manipulation of pictures is even a more recent phenomenon. Such a short time in evolutionary history is not sufficient to adapt our brains to be as skeptical towards pictures as we are skeptical towards communicated messages. We more easily believe what we can see with our own eyes, and gossip magazines use this to increase the credibility of their gossip stories.

Last, as I explained in chapter 1, a gossip story has the highest value when the content is new to us. It has news value the first time we hear it. However, hearing the same piece of gossip several times is sometimes still valuable, since multiple sources add up on credibility. Since Media Gossip as an act takes the form of rumors, characterized by unreliability because of unknown news sources, multiple sources are of great value in the context of Media Gossip. Journalists do not solely battle for scoops, but also present gossip stories that others have already made public. Readers value multiple sources, since if more people say this and that has happened, and especially if the multiple sources are independent from each other, chances rise that the story is actually true.

1.2 Media Gossip’s gossipy content

Post (1994) once said that Media Gossip indeed is gossip as a noun: it transmits a similar content of ‘backyard chatter’, but Media Gossip as an act lacks the gossip specific interactions, and is therefore more threatening than Interpersonal Gossip. As nouns both Interpersonal Gossip and Media Gossip concern information about the traits and behaviors of other people.

The major difference between Interpersonal and Media Gossip concerns the gossipees. In our Interpersonal Gossip exchange we gossip about both unknown and known people when we exchange Strategy Learning Gossip, and about people we know when we exchange Reputation Gossip (see chapter 4). Media gossipees are most often celebrities that we ‘know’ but still never encounter, and sometimes totally unknown people. In our exchange of Media Gossip we do not talk about gossipees we encounter in daily life.

As I outlined in chapter 4, I make a differentiation for different kinds of gossip according to their functional design. If all these functions that I attribute to Interpersonal Gossip also are applicable to Media Gossip, then all different kinds of gossip that I differentiated for in the context of Interpersonal Gossip are applicable to Media Gossip as a noun.

1.3 Why we like Media Gossip so much

In chapter 7, section 4.3 I put forward two different hypotheses that explain our interest in mass Media Gossip from an evolutionary perspective: the Learning Hypothesis and the Parasocial Hypothesis. Because both are central to explain the specific predictions I test in this study, I will resume them both.

1.3.1 Media Gossip and the Learning Hypothesis

As explained in chapter 7, the Learning Hypothesis says that we are all interested in Media Gossip if and because we can learn from the gossip information. We learn about strategies media gossipees have tested and that help us to survive, attract and guard mates, and we learn about cultural norms, how to behave. Media gossipees are carriers of fitness-relevant information. The Learning Hypothesis explains the presence of Strategy Learning Gossip (SLG) in the context of Media Gossip.

The Learning Hypothesis predicts that both celebrities and public unknowns are gossipees of Media Gossip, but what will be written about them will be different in some aspects. As explained in chapters 4 and 5, SLG can be about strategies with a clear or an unclear outcome. Most relevant is SLG that gives clear information about the outcome of a strategy. In these cases, we can clearly estimate if this strategy can increase or decrease our own fitness. SLG with an unclear outcome (e.g. X bought new boots) gives information about behavior strategies that we can mimic, but what lacks is information about what will happen if we mimic this strategy. As Boyd and Richerson (1985) explained, such unclear-outcome strategies will be mimicked from higher status individuals, and not from lower status individuals. Mimicking higher status individuals can potentially increase your own fitness, and since it is often unclear what exactly contributes to the status of an individual, the so-called General Copying Bias (GCB) explains why we copy the overall behavior pattern of higher status others. Public unknowns will only be gossipees of SLG with a clear fitness-relevant outcome, while celebrities will be gossipees of SLG with a clear or unclear fitness-relevant outcome.

1.3.2 Media Gossip and the Parasocial Hypothesis

Next to the Learning Hypothesis, I put forward the Parasocial Hypothesis in chapter 7. What I call the Parasocial Hypothesis does not explain our interest in all media characters, but only explains our interest in gossip about celebrities: Celebrity Gossip.

Some of us regard celebrities as one-way members of our social network. As Barkow (1989, 1992) explained, celebrities are present in our lives through the media, and are perceived as members of our social network by our mind. The reason for this is because selection acts so very slowly, we do not have a mind adapted to our present, but to an ancestral environment:

“A possible answer is that the mass media may activate the psychological mechanisms that evolved in response to selection for the acquisition of social information. […] We see them in our bedrooms, we here their voices when we dine: If this hypothesis is correct, how are we not to perceive them as our kin, our friends, perhaps even our rivals? As a result, we automatically seek information about their physical health, about changes in their relative standing, and above all about their sexual relationships.” (Barkow, 1992: 629-630)

Celebrities become part of our social network, but our interactions with them are parasocial: celebrities reveal their private lives to us, and we show emotions towards them, but we don’t share our private lives with them, and they don’t show emotions towards us (personally). Interactions in which reciprocity is lacking are called parasocial interactions (Horton and Wohl, 1956; Rubin, Perse, and Powell, 1985). I therefore talk about celebrities as our parasocial, or one-way members, of our social networks.

For those receivers who regard some celebrities as one-way members of their social network the Parasocial Hypothesis says that they gossip about celebrities for the same reasons we gossip about real, two-way members of our social networks. This includes both Strategy Learning Gossip and Reputation Gossip. Our interest in Strategy Learning Gossip is explained by the Learning Hypothesis, the Parasocial Hypothesis adds to this an interest in Reputation Gossip for some receivers.

The Parasocial Hypothesis predicts some will be interested in Mating Reputation Gossip about celebrities to learn who is a good potential one-way mate: who is a good partner, who dates whom and who broke up with whom. We want to learn who our one-way rivals are, and we will slander them. And we will control our one-way lovers and the one-way lovers of our allies. For all this, just like for Mating Reputation Gossip in the context of Interpersonal Gossip clear sex differences can be expected. Mating Reputation Gossip topics will be different for male and female celebrity gossipees.

The Parasocial Hypothesis also predicts that some will be interested in Social Reputation Gossip about celebrities to learn about one-way social network structures (who is related to whom, who is befriended with whom), to learn about good one-way potential allies, to manipulate the reputations of our one-way friends and one-way foes, and to detect and punish one-way cheaters.

Important to keep in mind is that the Learning Hypothesis predicts the interest in Media Gossip for all people, while the Parasocial Hypothesis only explains the interest in Media Gossip for people who regard celebrities as being part of their social network: as parasocial, or one-way members.

1.4 Media Gossip reflects issues relevant in our evolutionary past

Both the Learning Hypothesis, focusing on Strategy Learning Gossip, and the Parasocial Hypothesis, focusing on Reputation Gossip are based on the idea that Media Gossip reflects issues that were relevant in our evolutionary past. Issues that concern adaptive problems our ancestors faced. In chapter 4 I gave an overview of some of the most important adaptive problems our ancestors faced. Generally, when looking at the content table of introduction works on evolutionary psychology (e.g. Barrett, et al, 2001; Buss, 1995, 1997, 1999; Cosmides, Tooby & Barkow, 1992; Gaulin & McBurney, 2001; Tooby & Cosmides, 1990b) these issues can be classified in problems of survival, problems of mating, problems of parenting and problems of group living. Problems of survival concern adaptive problems that directly affect the fitness of an individual, problems such as finding a place to live, finding food, avoiding predators and so on. Problems of mating concern finding a good mate, dealing with rivals, and knowing how to keep your mate. Problems of parenting concern parent-offspring conflicts. Problems of group living concern problems of co-operation and conflict with individuals surrounding you.

Strategy Learning Gossip presents these topics as learning devices to increase our knowledge about strategies and therefore increase our experience records. Reputation Gossip reflects these issues because they were relevant when attached to a specific other individual to learn about that individual and manipulate that individual’s reputation. In this paper I will focus on the presence of these universal issues.

1.5 Previous research on Media Gossip

Previous research on Media Gossip is very scarce. I already gave an overview of some of the most important studies (Bird, 1992; Davis & McLeod, 2003; Levin & Kimmel, 1977; Schely-Newman, 2004; Sloan, 2001) that specifically looked at the content of Media Gossip in chapter 7 (section 5). I will not repeat those studies here, but refer to the most relevant results when discussing my own results, to compare similarities and differences.

2 Hypotheses

In what follows I will present my specific predictions I will test. These predictions concern media gossipees (about whom), media topics (what), sex differences in topics about male and female gossipees and the believability of Media Gossip.

2.1 Media gossipees

In chapter 1 I explained that no restrictions should be made that exclude certain people from gossip. We can gossip about all kinds of people, both known and unknown people. As follows from the results of paper 1 (focus groups on gossip), the respondents I interviewed about their gossip behavior also reported to gossip about anyone. In the next paper I will look more into detail which media gossipees are of biggest interest to the public. What I want to investigate in this paper is who is subject of Media Gossip. I explained that the Learning Hypothesis predicts that media gossipees are mere carriers of fitness-relevant information, and therefore media gossipees can be anyone: celebrities and unknowns. Unknown people who appear in the media will be referred to as ‘public unknowns’. Because no restrictions are made about gossipees for the general definition of gossip, a first prediction is:

Hypothesis 1a:

Media gossipees will be all sorts of people, both known (celebrities) and unknown (public unknowns) to an average person.

Celebrities’ status is a form of prestige that rests on the merit in the eyes of others (Henrich & Gil-White, 2001). Public unknowns as average people, unknown to the audience, lack this prestige and will only be subject to SLG with clear information about the outcome of their used strategy. Celebrities are movie stars, television hosts, soap actors, and so on, whose occupation involves media appearance. They are high status others, and can therefore be subject to Strategy Learning Gossip about strategies with either a clear or an unclear outcome. Moreover, the Parasocial Hypothesis also predicts that celebrities will be subject to Reputation Gossip as well. In sum, celebrities can be subject to a much broader spectrum of gossip categories than unknown people, and will therefore be more present in gossip magazines.

Hypothesis 1b:

Because celebrities can be subject of a wide range of SLG and of Reputation Gossip, and public unknowns can only be subject to a small subset of SLG, I expect celebrities to be more present than public unknowns.

2.2 Behavioral information: Media Gossip topics

I explained above that both Strategy Learning Gossip and Reputation Gossip reflect issues that were relevant in our evolutionary past. In order to be able to define if a gossip story is Reputation Gossip or not, information is needed from the receivers’ side. More specifically a piece of gossip can only be Reputation Gossip to a receiver who knows the gossipee. In this analysis I did not investigate whether Media Gossip stories are Strategy Learning Gossip and/ or Reputation Gossip, because audience research is more suitable for this. I assumed that most Media Gossip stories will be both Strategy Learning Gossip and Reputation Gossip. What I did focus on was if these Media Gossip stories reflect issues from our evolutionary past or not.

Both the learning and Parasocial Hypothesis predict that Media Gossip stories will reflect about issues relevant to our evolutionary past, and I therefore predict that:

Hypothesis 2a:

Media Gossip stories reflect issues relevant to our evolutionary past.

Negative information carries more weight than neutral or positive information. Rozin and Royzman (2001) have argued that both innate predispositions and experience give greater impact to negative entities such as events, objects and personal traits. Lupfer, Weeks and Dupuis (2000) for instance found that in the context of moral and immoral behavior, information about immoral behavior, which is negative information, is considered to have greater news value. Again I argue that good-selling gossip magazines reflect the desires of an individual, and link this to the negativity-bias theory. If, on average negative news is preferred over positive news, I therefore predict for good selling gossip magazines that:

Hypothesis 2b:

Media Gossip stories of good-selling gossip magazines will focus on negative traits and events.

Last, for these topics I expect differences between celebrity gossipees and public unknown gossipees. Public unknowns’ appearance in Media Gossip can only be explained with the Learning Hypothesis. If they can transmit fitness-relevant information, and can function as mere carriers, it does not matter who they are. What is more, public unknowns can only be gossipees of Strategy Learning Gossip with a clear practical fitness-relevant outcome: clearly beneficial or clearly harmful outcomes are mentioned. Celebrities can be subject to a broader range of Strategy Learning Gossip: information with either a clear or unclear outcome. Moreover, celebrities’ presence in Media Gossip can be explained by both the learning and the Parasocial Hypothesis. According to the latter we care for Reputation Gossip about celebrities because we regard them as being part of our social network. Because celebrities are one-way-members of our social networks, we are eager to acquire Mating Reputation Gossip and Social Reputation Gossip about them, which also are forms of gossip that do not necessarily embody clear fitness-relevant practical outcomes. Therefore I predict that:

Hypothesis 2c:

Celebrities will be subject to a broader range of gossip topics than public unknowns.

And:

Hypothesis 2d:

Public unknowns will only be subject to gossip with a clear fitness-relevant outcome.

2.3 Male/female differences

For both Mating Strategy Learning Gossip and Mating Reputation Gossip I expect sex differences to occur between male and female gossipees. As I explained in more detail in chapter 4, section 3 the desires of men and women for potential mates are different (Symons, 1979). As Buss (1994) resumes, in general men desire attractive women, because physical attractiveness cues signal fertility and reproductive value. In contrast, women on average have a stronger desire for men who have abilities and are willing to invest.

I have suggested that in their search for potential mates, both men and women benefit from acquiring the two kinds of Mating Gossip I differentiated: Mating RG and Mating SLG. This either to learn about the mating reputation and availability status of specific others and learn about which mating strategies are successful and which aren’t (see chapter 4). Because men and women have different desires in their search for potential mates, I expect sex differences in interest in Mating Reputation Gossip and Mating SLG. For Mating RG I expect women to have greater interest in gossip about the investment abilities and willingness of male gossipees, while I expect men to have greater interest in gossip about the physical appearance of female gossipees. When talking about Mating SLG, I expect women to have greatest interest in information about strategies that concern physical attractiveness (e.g. how to dress and how not to dress), and men to have greatest interest in information about strategies that influence access to resources.

If gossip magazines reflect universal gossip interests, these sex differences in Mating RG topics can expect to be present. More specific it can be expected that female media gossipees will be more subject to gossip topics about their physical appearance and male gossipees to be more subject to gossip topics about wealth (status indicator). Similarly sex differences in Mating SLG as well predict female gossipees to be more subject to gossip about physical appearance (how to achieve this), and male subjects to be more subject to gossip about access to wealth (also how to achieve this). Therefore I predict that:

Hypothesis 3a:

Assuming that gossip magazines reflect people’s interest in mating gossip, articles with only female subjects will present more information about physical attractiveness.

Hypothesis 3b:

Assuming that gossip magazines reflect people’s interest in mating gossip, articles with only male subjects will present more information about wealth status.

Because information about physical appearance of female gossipees is more valuable to the reading public than information about the physical appearance of male gossipees, for both kinds of Mating Gossip, I also expect more pictures of women to be present in gossip magazines. Again, for this I assume that good-selling gossip magazines reflect people’s interest in mating gossip topics. Therefore:

Hypothesis 3c:

Assuming that gossip magazines reflect people’s interest in mating gossip, they will present more pictures with female subjects than with male subjects.

Last, because in our modern societies jobs offer access to resources, and male gossipees are expected to be more subject to Mating RG and Mating SLG about wealth status, we can also expect male gossipees to be more subject to gossip information about their job. Therefore:

Hypothesis 3d:

Assuming that gossip magazines reflect people’s interest in mating gossip, articles with only male subjects will present more information about their job, as a source of wealth.

2.4 Gossip magazines’ reliability strategies

I explained above and in chapter 7 that Media Gossip has to deal with reliability problems. As an act Media Gossip resembles rumors, spread to a broad public to which the source is unknown. To resolve this problem, their best strategy to use, is adding pictures to their stories. Assuming that media sources are aware of their reliability problem, and try to solve this, I therefore predict that:

Hypothesis 4:

Assuming media sources are aware of their reliability problem and want to increase credibility, gossip magazines will be heavy on pictures.

3 Research methodology

The results of this study are based on multiple-coder content analysis of the gossip topics in gossip magazines.

3.1 Research population

The research population I selected my research units from, consisted of Belgium’s two Dutch so-called (De Bens, 2001) ‘sensational magazines’ “Zondagsblad” and “Blik” and one ‘television magazine’ “Story”. At the time this empirical study ran, both “Blik” and “Zondagsblad” took the format of a tabloid (a small-sized newspaper). “Story” was (and still is) a glossy gossip magazine.

At the moment of our analyses (2002) “Zondagsblad” was edited by the Editor Group “Cascade”, “Story” was edited by the Editor Group “Sanoma Magazines Belgium”, and “Blik” was edited by the Editor Group “Edibel” (Media Plan 02/03). Our research population therefore covers three magazines from different, independent publishers, which reduces the risk of overlaps in content of these magazines. Both “Cascade” and “Sanoma Magazines Belgium” are independent of the Belgian television channels. Their content is not directly influenced by a broadcasting channel. This cannot be said of “Edibel”. Editor Group “Edibel” is part of the Editor Group “Persgroep”, who publishes daily newspapers and magazines in Flanders, Belgium. “De Persgroep” is stockholder of “VTM”, a Flemish commercial broadcasting channel (6minutes, 2002). Shortly after our analyses “Blik” was merged with another magazine “TV-Familie”, which is directly edited by the Editor Group “De Persgroep”. It has been told (6minutes, 2002) that “Blik” disappeared because they reported too much about the famous reality television program “Big Brother”, broadcasted by VTM. Directors of this commercial broadcasting channel (VTM) wanted to reserve this information for magazines that were directly published by their stockholder Editor Group “De Persgroep”, such as “TV-Familie” and “Dag Allemaal”.

This illustrates how magazines published by Editor Group “De Persgroep” are influenced by the directors of the commercial television broadcaster “VTM”. That is why I did not opt to take any of the magazines edited directly by “De Persgroep” in my research populations. Magazines such as “TV-Familie” and ‘Dag Allemaal” are gossipy magazines that even reach a greater audience than “Blik” and “Zondagsblad”, and “Dag Allemaal” even is the best selling television magazine in Flanders, reaching an audience of 1,381,000 readers in 2002-2003 (Media Plan 02/03). Still, despite their popularity, I did not incorporate those in my research population, because their content is too much influenced by the directors of the commercial broadcasting company “VTM”. I wanted to analyze magazines that reflect the interest of an average individual. A magazine such as “Dag Allemaal” reflects the interest of a regular consumer of “VTM” television programs, and not the interest of an average individual.

Even though the three magazines I selected are not the top-selling magazines, they also reach a broad audience. Especially “Story” that reached 741,000 readers on average in 2002-2003 (Media Plan 02/03). “Blik” reached an average audience of 245,000 readers, and “Zondagsblad” reached 136,000 readers (Media Plan 02/03).

In a time period of 12 weeks between July 2002 and September 2002 I analyzed three best-selling Flemish gossip magazines, resulting in a total analysis of 36 gossip magazine issues. According to Hüttner et al (1995) 12 issues are sufficient to sample a research population for content analysis. No special events occurred during that period that might have influenced our analysis.

Out of this population I selected 852 articles, using the selection criterion that the main subject of the gossip story must be a human being, since gossip is always person-related. I followed the rule that the person(s) mentioned in the title, or subtitle; and being the main character(s) of the article, were coded as the subjects.

3.2 Coders

In total eleven coders (3 males, 8 females), aged between 22 and 27 co-operated in this study. One main female coder (author of this paper) was in charge of the research and trained the other ten. Coder subjectivity bias is considered as an important methodological obstacle in human coder content analysis. To diminish the scope for subjectivity bias I first of all developed coding questions that were as uncomplicated as possible.

Further, training sessions were organized and continued until all eleven coders obtained sufficient inter-coder reliability. For the training session I always used different materials that also differed from the material used for the main research, to make sure that no bias of familiarity with the coding was possible.

The main coder coded all 852 articles, the student coders each coded part of the articles, and I made sure that in the end each article was coded by three independent coders at different places and times.

3.3 Codebook

The codebook consisted of personal information about the subject, such as age, gender and celebrity-status. Next I set up a list of variables to categorize the subjects’ traits and behavior, and three context variables to situate the behavior: in the subjects’ job context, their leisure time or within their daily routine.

I gave simple coding instructions, asking the coders to answer following questions: What is the sex of the subject? Is the age of the subject mentioned? If mentioned, encode the age. And, encode if following traits/behavior categories are mentioned about the gossiped subject in the analyzed story: survival, long term mating, short term mating, status (wealth), beauty, parental care, kin investment, friendships, reciprocal investment and competition. For each category the coders needed to answer if this trait/ behavior was present in the gossip story in either a positive or negative sense. These behavior categories are based on an overview of some basic works in the field of evolutionary psychology (e.g. Barrett, et al, 2001; Buss, 1995, 1997, 1999; Cosmides, Tooby & Barkow, 1992; Gaulin & McBurney, 2001; Tooby & Cosmides, 1990b). Each of these strategies is presumably selected by natural or sexual selection, because they offered solutions to adaptive problems occurring in the environment of evolutionary Adaptedness. A ‘rest’ category was added to cover topics that could not be coded in any of the other categories. I refer to table 4.1 for an overview of used definitions, and their relation to adaptive problems.

Table 4.1. An overview of the universal traits/behaviors categories used in this content analysis of three Belgian gossip magazines

The category ‘survival’ enlists topics about the health status of the gossipee, whether he or she is in good health condition, or bad health condition. The negative subcategory of this variable (bad health condition) also enlists the articles that mentioned that a gossipee has died. This variable has the clearest fitness-indication, and will be used as an indicator of ‘clear fitness-relevant outcome’ information present or not. If this variable is present in an article, the audience gets clear fitness-relevant outcome information about the strategies of gossipees, as I predicted in hypothesis 3d, I expect this variable to be dominantly present in articles with public unknown gossipees.

For codings about the sex and age of the gossiped subjects, I had reliability rates of respectively 97% and 95%. Reliability rates of 80% or better are generally considered to be the minimum level of adequacy (Krippendorff, 1980). However, for the different traits and behavior categories, I could not always reach this level of reliability (see electronic annexes for detailed information). Therefore I opted to sum up the codings of the three independent coders for each article. This resulted in a 4-point scale, with 0= not at all present, 1= indication of presence, 2=probably present, 3 = definitely present. For this scale (0) means the trait is absent according to all three coders; (1) means the trait is present according to one coder; (2) means the trait is present according to two independent coders and (3) means the trait is present according to all three independent coders. For my results I excluded ‘indications of presence’, therefore ensuring that for all our data at least two out of three coders agreed on the presence of the trait/ behavior.

4 Results

4.1 Media gossipees: celebrities and public unknowns

To test my first hypotheses I looked at the variable celebrity-status, which divides media gossipees into categories according to their occupation. Categories are entertainer, athlete, royalty, politician, other media occupation and non-media-occupation. Gossipees that can be classified in the first five categories will be referred to as ‘celebrities’, because of the public character of their occupation, which leads to media exposure, and therefore potential pseudo-encounters with an average person. Subjects classified in the last category will be referred to as ‘unknowns’.

An overview of the results for this variable can be found in table 4.2. I have to remark that a celebrity can be classified into different categories, e.g., an athlete having a television show scores both on ‘athlete’ and ‘entertainer’, so that the sum of the valid percents outscores the total of 100%.

Table 4.2. Celebrity status of media gossipees of research population

My results confirm the first hypothesis (1a) that both known and unknown people appear in gossip magazines. A result also Bird (1992) found in her study of American gossip magazines. Both celebrities, with different media occupations are represented and unknowns as well. These latter take up a total percentage of 19.1% of all 852 articles, which means in 1 out of 5 Media Gossip stories an unknown person is gossiped about. This seems quite a lot. The explanation for this is that the major part of these ‘unknowns’ are unknown people who are connected to a celebrity. This is, they do not have a media occupation, but are connected to a celebrity because they are family (kin), or friends (allies). From the 162 gossip stories about ‘unknowns’ only 50 gossip stories (5.87% of the 852 articles) are solely about real unknown people, who have no connection to a celebrity or whatever. When I will talk about gossip stories with ‘public unknowns’ as gossipees, I will only refer to these 50 articles.

As predicted (hypothesis 1b) the majority of the Media Gossip subjects are celebrities, which appear in the media often and have high social status. Most often appearing in gossip stories are Belgian entertainers (35.1% of all gossip stories), followed by foreign entertainers (25.6% of all gossip stories). Levin and Kimmel (1977) found the same result in their research of gossip columns. A second major part of the media gossipees (29.9% of all gossip stories) consisted of royalties (26.8% foreign and 3.1% Belgian of all gossip stories). Royals are both high status and symbolize power, which makes them interesting Media Gossip subjects. Of all articles, 6.8% of the gossip stories reported about athletes, appearing in the media because of their public occupation as a national athlete.

Only 0.8% of all gossip stories had politicians as media gossipees. This last finding is surprising when compared to Levin and Kimmel’s (1970) finding that from the 1950’s to the 1970’s politicians seemed to increase in number as media gossipees in American gossip columns. And Bird (1992) as well reported about the presence of politicians in American gossip magazines. Differences might be due to cultural differences, future cross-cultural research is necessary to support this however. The 1.4% gossip stories that had ‘other’ celebrity gossipees were mostly about artists, who appeared in the media because of their occupation as being an artist.

4.2 Traits/behaviors categories information

The trait/behavior categories I set up reflect a variety of issues that represent adaptive problems, occurring in our evolutionary past. In a first hypothesis (2a) about these trait/behavior categories I predicted that all gossip topics will be classifiable in these categories. Hence, when approaching gossip from an evolutionary perspective it can be expected that gossip in the most general sense is about fitness-relevant issues, as Barkow (1989) explained. If gossip topics in modern Media Gossip do not reflect adaptive problems of our evolutionary past, both my hypotheses (learning and mismatch) might not be true. If all Media Gossip topics can be coded in this list of categories, this indicates that modern Media Gossip reflects issues that were relevant to our ancestors in the context of Interpersonal Gossip. Both the learning and Parasocial Hypothesis find support if all gossip topics can be listed in this category-system.

To test this I looked at the ‘Rest’ category which was added to list those gossip topics that could not be coded in any of the other categories. When looking at the results (see table 4.3), I get clear confirmation of my prediction; only one gossip topic could not be coded in any of the other categories. The article was about two Belgian celebrities who had been subject to a candid camera program, and reported about the joke. The coders all agreed that this ‘joke’ report did not fit in any of the other categories. This as good as empty rest-category indicates that Media Gossip topics reflect issues that were relevant in our evolutionary past, which support my hypothesis 2a.

Table 4.3. Frequencies of different gossip topics reflecting issues of our evolutionary past and ‘rest’ category, specifying for negative/positive news value

* Using Statistica Other Significance Tests: Difference between Two Proportions

Next, I predicted that negative issues would be more present, since independent good-selling gossip magazines reflect a general interest of their audience in Media Gossip, and generally negative news is preferred over positive news. To test this I looked at the different frequencies and valid percentages of presence of the different traits/behaviors categories, differentiating for negative and positive subcategories. Using Statistica Other Significance Tests: Difference between Two Proportions I calculated the significance of the differences between positive and negative subcategories for each trait/behavior category. An overview can be found in table 4.3. Note that for the categories ‘Competition Physical’ and ‘Competition Verbal’ there is no subdivision, since both are automatically ‘negative news’. They are listed in table 4.3 to give an indication of their presence in the overall research population.

Results show that significant differences exist between positive and negative subcategories. However, only for the category ‘Survival’ negative gossip outscored positive gossip in present frequencies. For all other categories positive news outscores negative news (see table 4.3 and graph 4.1).

Graph 4.1. Comparing rates of positive/ negative gossip stories in three Belgian gossip magazines

Next I looked at the traits/behaviors categories and compared the presence of the topics for celebrity gossipees and public unknown gossipees. For this analysis I selected only those articles with celebrity-only gossipees and public-unknown-only gossipees. Articles with unknowns connected to a celebrity and articles with both celebrities and unknowns were left out of these analyses. Out of the 852 articles, 739 were selected, from which 50 articles with public unknowns as gossipees and 689 articles with celebrity gossipees.

I first compared the frequencies of presence of all traits/behaviors categories for celebrity-gossipees and public unknown-gossipees (see table 4.4). My results show that Media Gossip about public unknowns reports almost always (90% of all articles with only unknowns as gossipees encoded as ‘survival negative’) about life threatening events. Negative information about the health condition of the media gossipee signals clear fitness-relevant outcome information of their strategies. This supports the hypothesis that stories about unknown people report about fitness-relevant behavior with a clear outcome (hypothesis 2d).

Table 4.4. Comparing frequencies of different traits/behaviors categories for celebrity-gossipees and public unknown-gossipees.

* Using Statistica Other Significance Tests: Difference between Two Proportions

I looked more into detail at the five stories about public unknowns where survival negative was not encoded by at least 2 out of three coders. For all the gossipees of these stories, results indicate that their health (fitness) was not threatened directly. One story was about a father whose abused daughter needed help and was refused help, which reports about a threat to the inclusive fitness of the gossipee (the father). Another story was about a robbery. The article focused on the misbehavior of the robbers, clearly labeling them as gross cheaters (they had robbed an institute for handicapped people). These two stories still carry over negative information that warn the readers about threats; threats of cheaters acting in the society. The robbers were labeled clearly as cheaters, as well as the people in the first story that refused to help the abused daughter.

Next to these two warning forms of gossip, other stories reported about how to clearly improve your status. There was one story that reported about the successful strategy of a guy who won an important contest. This resulted in the fact that his wife was really proud of him. This is a clear form of Strategy Learning Gossip, from which the readers can learn how they can increase their own status, by simply mimicking what this public unknown-gossipee did. Another story reported about a couple who declared that no-one would be able to tear them apart. Their family and friends did not support their relationship and the gossip story reported on how these people dealt with this problem and managed to maintain their relationship. This is a form of what I would call Mating Strategy Learning Gossip, from which the readers can learn how to solve problems of mating. More specific, in this case readers can learn about a successful strategy to cope with an environment that is hostile towards a relationship (in this case the solution of the gossipees was ignore those who disapproved and seek new friends).

A last story reported about two people who were heavy addicts of Media Gossip, buying and storing every issue of the tabloid “Blik”. This information does not carry over any strategy-relevant information to the readers. The gossipees were no high status individuals, thus I would not expect this unclear strategy information to be reported about them. The simple reason why this was published, in my opinion, was because “Blik” was soon to be merged with another gossip magazine (“TV Familie”) and I think they wanted to boost their own status as tabloid for the last time, by presenting people who clearly valued their product. This gossip story was more a hidden advertisement for the tabloid “Blik”.

Next, as follows from the overview given in table 4.4, fewer empty and nearly-empty categories are present for celebrity gossipees. A broader range of traits/behaviors is discussed about celebrities, which is in line with my prediction 3c.

When exploring a little more on what is written most about celebrities and what is most present for public unknowns, my results indicate that articles about celebrities significantly (p<0.001) report more about their access to resources and their physical attractiveness (see table 4.4). Media Gossip about public unknowns significantly (p<0.001) reports more about both physical and verbal conflicts (38.0% & 44.0% vs. 0.6% & 11.4% in the celebrity-gossipees gossip), and negative relations with partner, kin and non-kin (see table 4.4).

Next to comparing the frequencies of all categories, I also ranked the categories for celebrity-gossipees and public unknown-gossipees (see table 4.5). Top ranked gossip topic for celebrity-gossipees is ‘status positive’. Media Gossip reports most about their wealth status and how they achieved this. The wealth status of public unknowns is also enlisted in the top ten ranked gossip topics about them, but it does not concern positive information, but negative: how poor they are and what caused this. It is gossip information that informs readers about what to avoid doing themselves if they want to avoid a decrease in their own wealth status. Celebrities carry information on what to mimic, public unknowns teach us what to avoid.

In general, celebrities are gossipees of positive gossip topics. Top ranked topics are their good parental behavior, good relations with family and friends, and successful relationships. Three negative topics that appear in their top ten ranked topics are survival negative, which concerns their negative health condition. Their disputes with others (verbal conflicts) and their unsuccessful love affaires: troubles in relationships. Relationships are a popular topic in Media Gossip. This topic also pops up in the top ten ranked topics for public unknown-gossipees, both in a positive and negative sense. Levin and Kimmel (1977) as well reported about the dominant presence of love issues in Media Gossip columns. Relationships are situations almost any person gets confronted with in his or her life. Gossip about such a universal often appearing situation, that even involves many problematic situations, appeals to almost anyone. Therefore it is no surprise gossip magazines publish so much on this, because their will always be an audience for this gossip topic.

The top ranked issue for public unknown-gossipees is survival negative. The negative health status (fitness) of public unknowns is almost a necessity for them to become a media gossipee. Other negative aspects, such as conflict situations, both verbal and physical, as well appear in their top ten ranked gossip topics. However, it is not all negative information about public unknowns that reaches Media Gossip, even though almost all articles report about their negative fitness condition, these articles do report about positive aspects such as their good parental behavior, their altruistic behavior towards kin, friends and totally unrelated others. Media Gossip reports how they offer and find support in their social network, even though they are in a life-threatening situation.

Table 4.5. Top ten ranked traits/behaviors categories for celebrity-gossipees
 and public unknown-gossipees of three Belgian gossip magazines.

In sum both celebrities and public unknowns are media gossipees on information about their well functioning social network. For celebrity-gossipees this information is added up with information about their wealth status, while for public unknowns life-threatening events are a necessity to become a media gossipee.

4.3 Sex differences in gossip magazine topics about media gossipees

As I predicted in hypotheses 3a, 3b and 3d I expected male and female gossipees to be subject to different gossip topics. To test these predictions I left out all articles about mixed sex gossipees. In total 585 articles were used for these analyses, of which 304 were about male gossipees and 281 about female gossipees.

Again I looked at the frequency of articles were topics where coded to be present by 2:3 or 3:3 coders. Using Cross-tabulations and looking at the Contingency Coefficient for nominal data I looked at the significance of sex differences in topics about media gossipees.

Of all 585 articles I looked at, 253 articles mentioned something about positive status aspects of the gossipee (he or she is wealthy, and how he or she acquired resources). Out of these 253 articles 141 had one or more male gossipees as subject and 112 one or more female gossipees. When looking at all articles about male gossipees, 46.6% mentioned something about status, which is slightly, and not significant (p=.112) more than the 39.9% of all articles about female gossipees that mention something about their positive wealth status (see table 4.6). These results cannot support my hypothesis 3b.

Still, I did find support for hypothesis 3b when looking at negative status aspects of the gossipees (he or she is not wealthy and/ or how he or she lost resources). In general, the number of articles mentioning negative status aspects is much lower (58; 9.9% of the total of 585) than the amount of articles mentioning positive status aspects (253; 43.2% of the total of 585). Still, of all 304 articles about male gossipees 13.5% mentioned negative status aspects, which is significantly (p<.05) more than the 6.0% of all 281 female gossipee articles that mentioned about status aspects (see table 4.6), and in support of what I predicted; that male gossipees would be more often subject of gossip about their wealth status than female gossipees.

Table 4.6. Differences between male and female gossipees for gossip topics about mating cues

* Percentages within sex of media gossipee

When looking at the results for sex differences in gossipees of Media Gossip stories about beauty aspects, I found support for hypothesis 3a, which predicted that female gossipees would be more subject to gossip stories about physical appearance than male gossipees. Gossip about physical appearance has a low occurrence in general. Only 67 (11.5%) of all 585 articles mentioned positive physical aspects (he or she is pretty and how he or she came to look so good) and only as little as 11 articles (1.9%) or all 585 articles presented information about negative physical aspects of the gossipees (he or she looks bad and what causes this). Still for both variables I found sex differences. Of all 281 articles about female gossipees 21.7% mentioned positive beauty aspects, which is highly significantly (p<.001) more than the 2% of all 304 articles with male gossipees that talked about positive beauty aspects. And of the 281 articles with female gossipees only 3.2% mentioned negative physical aspects, but still this is significantly (p<0.5) more than the 0.7% of the articles with male gossipees that talked about this (see table 4.9). In general these results support my prediction that articles with female gossipees would focus more on beauty aspects than articles about male gossipees.

When looking at the 337 articles that mentioned something about the job of the gossipee, as a cue for access to resources, I noticed that these are strongly significantly (p<.01) more present in the 304 articles of male gossipees (63.5%) than in the 281 articles about female gossipees (51.2%; see table 4.6). This supports my hypothesis 3d that predicted that articles about male gossipees would talk more about their job as a cue for access to resources than articles about female gossipees.

Table 4.7. Representations of male and female gossipees on pictures on three Belgian gossip magazines.

Last, I also looked at who is represented on pictures in gossip magazines. I expected female media characters to be more present than male media characters (hypothesis 3c) on pictures. When looking at the variable ‘Who is pictured’, I found no sex differences. To test this I looked at the male/female rates of all depicted media gossipees of all 852 stories. Of all pictures 40.1% pictured both men and women, 27.7 percent pictured only men, and 26.4% pictured only women (see table 4.7). These results do not support my prediction.

4.4 Gossip magazines and the reliability problem

In a fourth and last hypothesis I predicted that all three gossip magazines would be heavy on pictures, to solve the reliability problem they all have to deal with. Bird (1992) already had shown that gossip magazines rely heavy on pictures, and my results can only confirm this again. In general 801 (94%) of all 852 articles were illustrated with one or more pictures about the gossipee(s)’ actions.

When comparing the three gossip magazines I analyzed, I noticed that “Zondagsblad” is least heavy on pictures. In 16.3% off all 263 articles selected from this gossip magazine even no picture was added, which is significantly (p<0.001) higher than both other gossip magazines’ articles that had no pictures (Story: 1.1% out of 435 articles; Blik: 1.9% out of 154 articles). The pictures that were added to the articles in “Zondagsblad” were most of the times (63.9%) smaller than the text they illustrated. Only in 10.6% of their articles the picture added was bigger than the gossip story. For both “Story” and “Blik” there is an opposite trend, their pictures are in half of the articles (Story: 50.3%; Blik: 52.6%) bigger than the text they illustrate (see table 4.8).

Table 4.8. Comparison of picture and text size for three Belgian gossip magazines.

5 Conclusion

Media Gossip stories from the three gossip magazines I analyzed present gossip information about a broad range of gossipees. Both well-known celebrities and totally unknown people are subject of Media Gossip. Still, celebrities are dominantly present, and when focusing in detail on the unknown gossipees it must be remarked that those consist of real public unknowns and unknowns who are in some way connected to a celebrity. The real public unknowns only take up a marginal part of (5.87%) of all 852 articles I analyzed.

About these media gossipees a broad range of traits/behaviors topics is discussed and all topics can be classified in one or more issues that reflect adaptive problems that appeared in our evolutionary past. Media Gossip reflects fitness-relevant information.

Because of the negativity bias, I expect Media Gossip from good-selling gossip magazines reflects more negative information than positive information. My results however indicate that this is only true for the variable ‘survival’ that covers gossip topics about the health condition of gossipees. For this gossip topic negative information (about illnesses, life-threats and death) outscores positive information (about wellness, good health). For all other issues positive news clearly outscores negative news. Media Gossip reports more about the successful relations, successful mating strategies, successful parenting behavior, and successful other social relations of media gossipees. Reason for this might be that media gossipees are dominantly high status celebrities. It has been shown that mimicking the successful behavior of others can increase your own status (Boyd & Richerson, 1985; Henrich & Gil-White, 2001). High status celebrities might therefore be ideal individuals to copy successful strategies from.

When looking at the differences in gossip topics appearing for celebrities and unknowns, I indeed found that celebrities are more often subject to positive gossip information, while unknown gossipees seem to function as carriers of negative gossip information. Information about the negative health status (fitness) is a top ranked story for public unknown-gossipees. The top ranked gossip topic for celebrity-gossipees is their positive wealth status. Still, public unknowns are subject to some positive gossip as well. For both celebrities and public unknowns, the Media Gossip stories I analyzed report about their successful parental behavior, their successful relations with family and friends, and their successful interaction with their spouse. Relations are top ranked for both kinds of media gossipees and this for both positive as negative sub categories. This is in line with the results of previous research on Media Gossip (Levin & Kimmel, 1977).

When focusing on the differences in frequencies of topics appearing for male and female gossipees, my results indicate that male gossipees are more subject to negative information about their wealth status than female gossipees are. Together with the fact that more is gossiped about the job (as access to resources) of male gossipees than female gossipees, this supports my prediction that male gossipees would be more subject to gossip about wealth status than female gossipees. Female gossipees however outscore the male gossipees on presence in articles about both positive and negative aspects of physical appearance. This also confirms the predictions I made.

Last, I explained how Media Gossip as a noun might be similar to Interpersonal Gossip as a noun, but both forms of gossip are different as verbs. As an act Media Gossip comes closer to rumors, being spread to a broad and unknown audience. Media Gossip has to deal with reliability problems, because their audience does not know the source of information that well, and the typical trust-based relations context of Interpersonal Gossip is lacking. An ideal way to increase their credibility is adding pictures as ‘proof’ of their gossip stories. Bird (1992) already found that American gossip magazines are heavy on pictures, and my results show that Belgian gossip magazines as well present a lot of visional evidence: 94% of all 852 articles I analyzed were illustrated with pictures.

6 Discussion

Bird (1992) already noticed a change from negative loaded Media Gossip to more positive toned Media Gossip in the 1980s. My results also show how nowadays in Belgian gossip magazines positive Media Gossip outscores the negative news. Only for health information (the clearest fitness indicator) negative news seems to occur more than positive news. For all other topics positive gossip gets higher frequency rates. Negative gossip is valuable because it can warn us about dangers and strategies that threaten our fitness. Theory on negativity bias predicts higher interest in negative news, and still this is not reflected in Media Gossip. I think more research on how Media Gossip evolved over time in different countries would be useful to see if this is a general trend in different societies. And if so, additional audience research might be useful as well to investigate why Media Gossip seems to be subject to a positive bias rather than a negative bias. I think the fact that media gossipees are dominantly higher status celebrities might be an answer to this puzzle. The General Copying Bias (Boyd & Richerson, 1985) predicts that we tend to copy the overall behavior of successful others. However, if we can choose between information about the failures of higher status others and information about their successful strategies, the latter makes more sense. Future content analyses should focus more on the specific outcome of gossiped about strategies of media gossipees.

Another suggestion for future research concerns the use of pictures by Media Gossip stories. Since my results show that “Zondagsblad” is less heavy on pictures than “Story” and “Blik” it would be nice for future research to test whether the believability of “Zondagsblad” is lower than the credibility of both other gossip magazines. However, “Blik” does not exist as a separate magazines any longer. In 2003 this tabloid merged with the magazine “TV Familie”. Still, future analyses on picture size and credibility of the different gossip magazines could shed more light on the influence of pictures on believability

Next, I did explain already that in order to classify a gossip story as Reputation Gossip or not, information is needed about the relation between gossipee and receiver. Audience research is therefore most suitable to investigate whether a piece of gossip is Reputation Gossip or not. However, a simple rule can already predict if a piece of information is pure Strategy Learning Gossip, pure Reputation Gossip or both; by replacing the gossipee with any other subject can give this information. If the value of the gossip story does not change at all, it can be classified as pure Strategy Learning Gossip. If the value changes and looses most information it is pure Reputation Gossip. And if the value changes, but still the information remains valuable, it is a mixture of both. Pure Strategy Learning Gossip does almost actually not exist, since a gossipee will always be known by someone, which makes each piece of Strategy Learning Gossip be Reputation Gossip to at least one person (who knows the gossipee and initially spreads the information). I did not investigate whether our analyzed stories are pure Reputation Gossip or not, but I do encourage future researchers to look at this, by simply using the trick of changing the media gossipee with ‘X’ and looking if the information still has value (Strategy Learning Gossip present) or not (pure Reputation Gossip).

As a last suggestion for future research I want to comment that at the moment of our analyses, my classification system of gossip was not yet fully elaborated. I therefore did not encode the presence of all the different kinds of gossip I differentiated in chapter 4 in this study. I merely looked at the presence of universal issues relevant to our evolutionary past. The fact that for instance mating cues and information about relations is present in Media Gossip indicates that forms of gossip such as Mating SLG, Mates Detection RG, and Mating Structure RG and so on, are present in Media Gossip. However, I think a content analysis of Media Gossip, using the classification system I presented in chapter 4 would be relevant and give more details about what exactly is represented in Media Gossip. Instead of using articles as coding units, I would also suggest to take the media gossipees as coding units.