Participation shaping the values of news articles: implications of news as anti-rival information products

ABSTRACT

In the literature, anti-rivalry has been proposed to describe goods that become more valuable when used. Empirical testing of this theory is lacking, and the concept has not yet been applied in the context of news. This study identifies five distinctive value constructs – news value, emotional value, value of reading comments, social value of commenting and social value of sharing – from the statistical analysis of a survey of 2020 online news consumers on Finnish public broadcaster Yle’s website. We then assess changes in these values when the news is consumed and users have contributed to it by commenting on it and sharing it. The results of the analysis and additional in-depth user interviews indicate that news possesses elements of anti-rivalry. Additionally, the results indicate interesting intercorrelations among the values and user groups. When a piece of news has high news value, user contributions do not play a major role, but the value of sharing increases. When a news product is commented on, the act increases the social value of sharing for sharers, decreases the news value for readers and increases the negative emotions felt by readers.

KEYWORDS:

• Anti-rival
• news value
• produsage
• platforms
• user-generated content

Introduction

Journalism, and more precisely, freely distributed digital news, is considered a public good (Benequista, 2022; Walters, 2020), resulting in free-riding (Fontaine, 2014). This feature makes value capturing in news production difficult (Hess & Ostrom, 2011), thereby complicating the design of sustainable digital business models (Achtenhagen et al., 2013) and the acts of sustaining free press (Kaye & Quinn, 2010; Picard, 2014). Producing public goods by private entities is difficult, as everyone hopes someone else pays for the goods (Kaul et al., 1999). Hence, public goods tend to be underprovided in many social settings (Weber, 2009).

The taxonomy of public, private, common and club/toll goods is based on two axes: whether the use of goods can be excluded by users or if they are nonexcludable; and whether the consumption is rival – meaning that if someone uses the good, its value is subtracted for others – or nonrival (see Table 1) (Buchanan, 1967; Musgrave, 1959; Ostrom & Ostrom, 1977; Samuelson, 1954).

Table 1. Basic taxonomy of goods matrix.

	Excludable	Nonexcludable
Rivalrous	Private goods If someone buys a private good, no one else can buy or use it.	Common goods Goods that are rivalrous but nonexcludable (e.g. usable finite natural resources)
Nonrivalrous	Club or toll goods Goods that are excludable but nonrivalrous (e.g. pay-TV)	Public goods Free for everyone

As Nobel-winning economist Arrow (1962) argued decades ago, data and information are nonrivalrous. Houghton (2005), Cooper (2005) and Smichowski (2016) argued that most forms of data are not only non-rivalrous but also anti-rival, which refers to a unique property of certain goods that increase their value for others when they are used. This developing theoretical concept has thus far mostly focused on the accumulation of value in open-source communities (Weber, 2000), science (Houghton, 2005), commons (Cooper, 2005) and new forms of governance (Nikander et al., 2020; Smichowski, 2016). Perhaps the most widely used example of this is the Linux operating system (Smith & Kollock, 2000).

We define anti-rivalry as the property of a good that results in an increase in value for others when someone uses the good. Similar to the value subtraction of rival goods, the definition of anti-rivalry does not require value accumulation during usage (becoming more valuable the more they are used), and anti-rivalry does not require a linear accumulation of value across repetitive use. Discussions of anti-rivalry have been mostly theoretical, and empirical testing is lacking. Additionally, the concept has been suggested to exist in the context of news (Peters, 2019) but has not yet been tested as such. In this study, we aim to fill these gaps and further theorise the concept of anti-rivalry.

The value taxonomy has more than pure theoretical value for the news industry. News can also be classified as experience goods (goods whose value can only be evaluated after use; Huang et al., 2009). However, for news production, it is valuable to be able to predict these experienced values in online platform strategy, news site design and editorial work. If we can determine when the pure usage of a piece of news – or user contributions to it – increases the value experienced by other users, we can identify new opportunities in news production and platform design.

A reasonable definition of a news product as a good depends on the research perspective. Alternative definitions include a subscription to an online newspaper, an algorithmically personalised daily news offering, and a single news article, including dynamically generated advertisements and single sharable headlines and images. In this research, we define a news product as precisely one online news page containing one news article, potential comments and antecedents within that content context. This unit of online news can be identified by a unique website address (URL). Different users repeatedly experience the possibly changing content of such a news product.

Value, a central part of the definition of any business model (Rohn et al., 2021), is a stock (capital) gained by the participants. Money, time and attention, which users spend on an activity, are factors that guide consumer choices. The construct of personal value is less understood (Heinonen et al., 2013). Emphasising customer value is sometimes referred to as “customer-dominant logic”. This perspective brings to light the challenges of modelling media platform businesses to include company involvement, control, value co-creation, value visibility and the scope and character of the customer experience (Heinonen & Strandvik, 2015; Voima et al., 2010). Thus, services and their exchange systems should focus on customer value creation (Archpru et al., 2015; Edvardsson et al., 2005, p. 118).

There is a lack of conceptual clarity regarding the use of a piece of news and the multidimensional nature of news engagement (Park et al., 2021). Here, we define this usage as users’ involvement in news consumption and evaluation processes through their interactions of reading, sharing and commenting on news products.

According to several studies (Blom et al., 2014; McCluskey et al., 2015; Park et al., 2021; Schaedel & Clement, 2015; Trussler & Soroka, 2014; Wahl-Jorgensen et al., 2010), online news commentary and sharing are driven by affordances, motivations, desires and values. However, few studies have investigated both individual and structural factors simultaneously (Sang et al., 2020).

In the digital platform society (van Dijck et al., 2019), news media have become dependent on social media intermediaries (Kleis Nielsen & Ganter, 2018). Identifying new value-creation opportunities from usage and user contributions is central to innovating new sustainable business models for news production. Social media platforms employ network effects, providing more valuable services for users and creating “winner-takes-all” situations for platform owners. Recently, interest in how incumbent media companies create platform business models and challenge platform winners has increased (Cozzolino et al., 2018; Jääskeläinen et al., 2021). Platforms are often identified as creators of value at the platform level (Amit & Zott, 2001; Trabucchi et al., 2021). However, there is a striking shortage of theory on what kinds of value platforms create for their customers and how. Creating customer value is a substantial prerequisite for creating platform value for an operator (Cennamo, 2018; Priem et al., 2018).

Our research questions are as follows:

• RQ1: What value constructs do news consumers gain or lose when reading the news or the comments attached to it and when actively commenting on or sharing the news item?

• RQ2: How do the actions of other users affect the value constructs that news users gain or lose?

Materials and methods

According to the phenomenographic approach applied in this study, we examine the variety of ways users experience a phenomenon rather than from the researcher’s perspective. The ontological and epistemological bases of phenomenography rely on understanding knowledge both subjectively and relatively. Knowledge and concepts are formed when people perceive and interact with their external world. Thus, notwithstanding its interpretative approach, phenomenography falls between realism and relativism. This middle ground is necessary for this study, as value constructs and operational models reflect complex social models that comprise many intertwined constructs. Hence, researching these constructs is worthwhile (Svensson, 1997).

Phenomenography is closely aligned with philosophical hermeneutics, in which interpretation and understanding are at the core. Thus, we view knowledge as emergent from constant and repetitive interpretations. Our research design (Figure 1) acknowledges that the whole cannot be understood before the details are understood and that the details may be understood differently after the whole has emerged. Furthermore, central to the hermeneutic process of understanding is the notion that all comprehension is guided by prior knowledge, which is constantly being reformulated (Kakkori & Huttunen, 2014).

Figure 1. Research design.

We used a mixed-methods approach to collect and analyse the data in this study. Detailed information on news consumers’ value perceptions, behaviours and intentions from reading and commenting on articles and from reading comments and sharing news items during a sample session was collected via a self-administered online survey placed on Yle’s news website¹ for two weeks: 4–19 June 2021. Respondents (N = 2020) accessed the survey linked to 260 different news stories.

The Finnish public service media create value for both individuals and the larger society via platform collaboration (Ylen strategia, 2020). Any strategy of this nature exists in tension with legacy public service ideals (Lowe & Maijanen, 2019). The Yle online news service reached 44% of the Finnish population in 2021, but the audience was heavily skewed towards older people. Notably, Yle’s news audience is widely trusted among the general public (85% trust scores [6–10] on a 10-point scale; Newman et al., 2021, Schulz et al., 2019).

The survey design was based on the operationalisation model introduced by Sheth et al. (1991), which grouped value components based on empirical results, resulting in five value constructs: functional, social, emotional, epistemic and conditional value. In our approach, participants were invited to evaluate the value components they experienced on a scale of 1–10 in 33 statements. The statements were designed to reflect the components of these five value constructs in the news context. Additional free-text responses were collected to identify any additional value components. We also designed the survey to collect appropriate data about user behaviours and to measure the relation of the value experienced to time and effort sacrifices. We applied several rounds of exploratory factor analysis to identify the specific value constructs that people experience while consuming news. These constructs were analysed, interpreted and given representative names: news value (NV), value of reading comments (VRC), social value of commenting (SVC), social value of sharing (SVS) and emotional value (EV). We investigated the correlations of the found value constructs (using them as value variables) with newsreader behaviours using additional statistical approaches.

To triangulate the results and further increase the validity of the method, we randomly selected 17 respondents from those who volunteered in the survey for follow-up thematic interviews. The interviews and written responses were coded and interpreted in three rounds. The first round was consciously open-ended to support the identification of value dimensions that would not otherwise have been found if we had relied solely on theory. Second-order concepts were interpreted based on the first-order concepts and survey results. Aggregate constructs were then interpreted using both data sources. Both survey respondents and interviewees gave their informed consent to participate in the study. Figure 1 illustrates the research design and a detailed list of survey questions, while explanations of variables and statistical analyses are presented in the appendices.

Table 2 present the responses and basic descriptive results.

Table 2. (a). Variables and key results on behaviours. (b). Variables and key results on worthiness compared with time and effort. (c). Variables and key results of value dimensions experienced.

Code	Label	Question			Yes						No
dshare	Shared	Did you share the story or are you going to share the story via social media or email?			361						1,352
dcomm	Commented	Did you comment on the story?			380						1,186
drcomm	Read comments	Did you read the comments?			1,058						837
Code	Label	Question			Obs			Mean			SD		% 7–10		% 1–4
cwt	C Comment-worthy	Commenting this story was worth my time and effort.			1,364			5.09			3.02		33.4		39.5
rcwmt	RC Readworthy comments	Reading the comments was worth my time and effort.			1,039			5.39			2.89		39.6		37.6
wmt	A/C Readworthy article and comments	Reading the story and its possible comments was worth my time and effort.			1,955			7.36			3.05		71.1		19.5
rwt	A Readworthy article	The story was worth reading.			1,954			8.14			2.69		80.7		13.9
Usefulness
rcuse	C Usefulness	The comments in the story were useful to me.			1,041			4.91			2.67		30.7		40.2
use	A Usefulness	I got useful practical information or advice.			1,951			5.93			3.04		49.8		30.8
Informativeness
caq
rcinf	C Curiosity-fulfilling	By commenting, I was able to ask questions that came to mind from this story or the attached comments.			373			6.07			2.63		49.5		23.3
scr	RC Informativeness	Those who comment on the story know more about the story than the journalist who wrote it.			1,035			3.89			2.65		16.1		59.1
qst	A/C Social curiosity	I would like to get more information about what other people think of this story or the comments posted to it.			1,907			4.96			2.95		33.8		41.9
rccnf	A Confusion	After reading the story, I have more questions than answers.			1,952			5.00			2.97		33.3		43.8
cms	RC Coherence from comments	My thoughts on the news item were confirmed in the comments.			1,036			5.49			2.82		36.7		30.6
dsc	C Story improvement	I commented to point out a mistake or missing information in the story.			372			5.59			3.24		42.6		35.1
inf	A/C Social usefulness	I may use information from this story or its comments when discussing with people who don’t already know the story.			1,954			6.73			3.07		62.5		22.8
cnf	A Informative	I got valuable information for myself about what is happening right now.			1,951			6.81			2.98		63.6		23.0
													66.5		14.0
Emotional Value
rlx	A Relaxing	I feel more relaxed now than before reading the story or comments.			1,947			4.25			2.64		18.7		50.2
irr	A Irritating	I feel more irritated now than before I read the story.			1,949			4.53			3.28		30.8		51.2
swf	S Emotion expression	I want to share the emotional experience I got from this story with others.			1,919			4.72			2.96		30.8		46.4
rcirr	RC Irritating comments	I got irritated at the comments.			1,034			4.88			2.90		31.0		44.2
bth	A Bothering	The story or comments might bother me for a while.			1,954			4.92			3.14		36.4		44.6
mod	A/C Uplifting	The content of the story or its comments made me feel good.			1,946			5.12			3.09		36.4		43.1
prd	A Personal proudness	I’m proud of myself because I use my time to read stories like this.			1,929			6.19			3.01		50.7		25.4
cwf	C Emotion-sharing	By commenting, I tell others what kind of feelings this story woke in me.			373			7.00			2.70		65.8		15.8
cim	C Personal relevance	I commented because this topic is important to me.			372			7.71			2.61		74.8		11.5
Social Value
unf	A/C Social justice	Reading the story or its comments helped me to understand the unfairness in society.			1,952			6.09			3.25		51.8		29.7
swt	S Expression of opinion	I want to talk publicly and to as many people as possible about what I think of this story or the comments attached to it.			1,921			4.94			2.91		33.3		42.7
rcexp	RC Empathy	Reading the comments helped me understand the experiences of other people in relation to the topic of the story.			1,031			5.78			2.77		46.0		28.3
cid	C Sharing idea	I commented because I had an idea.			373			5.82			2.80		47.3		27.3
cdf	C Advocative	By commenting, I can defend the interests of individuals and groups.			372			7.36			2.68		68.4		13.7
cpw	C Adding perspective	By commenting, I make a forgotten point of view visible.			372			7.81			2.26		78.3		7.8
Social Coherence
stn	S Social statement	I want my friends and other people close to me to notice that I read this story and made comments.			1,918			3.98			2.85		20.4		55.6
ftf	C Relatable	It would be a pleasure to talk face to face with people who comment on stories like this.			1,927			5.30			2.89		36.0		35.9
rcimp	RC Duty to read comments	It is important that readers read the comments attached to this story.			1,036			5.78			2.87		42.9		30.1
chb	C Habitual commenting	I usually comment on stories like this.			375			6.14			2.59		50.0		24.5
dt2	RC Social duty	It is important to the common good that people read others’ comments when they exist.			1,929			6.59			2.78		57.8		22.1
plm	A Identity-building	People like me read stories like this.			1,914			6.90			2.44		60.6		12.6
dt1	A Personal duty	It is important for the common good that people read stories like this.			1,940			7.74			2.85		75.7		14.4
opp	C Opportunity	It is important that a story like this be open for comments.			1,934			7.74			2.56		74.1		11.0

A = reading article, A/C = reading article or comments, C = commenting, RC = reading comments, S = Sharing.

What values do news consumers experience?

Generally, according to the descriptive statistics reported in Tables 2, the news articles were considered to be worth reading and worth the reader’s time and effort. Of the respondents, 47.6% assigned a score of 10 to questions about whether the story was worth reading. When correlated with sacrificed time and effort, reading the story was considered substantially more valuable than commenting and reading comments. Nevertheless, a third of the respondents considered it worth the time and effort to comment (responses 7–10). Two value dimensions that stand out from the results concerning the story pertain to obtaining confirmation of the reader’s previous thoughts and the public interest in stories, such as those read.

On the value of commenting, several results reflected high-value perceptions among the commenters. The highest responses were given to the claims that the commenter made a forgotten view visible and that the topic was important to the commenter. Claims that commenting allowed them to defend the interests of a group of people and that sharing emotions awakened by the article were also of high value to the commenters. All the respondents considered the opportunity to comment on a story vital, reflecting the value of the principle of openness in the form of commenting.

Overall, readers did not report high values connected to reading the comments. The highest personal value experienced was understanding other people’s experiences (average value of responses 5.78), whereas other high values were connected to the general importance of reading the comments or of reading comments specific to the story about which they responded in the questionnaire.

Regarding sharing the story, the highest reported value was related to sharing opinions and emotional experiences, but both were below the theoretical mean of 5.5. The reported values varied substantially. The expression of opinion was important to a third of the respondents (N = 1,921, 33.4% responded 7–10), and the transfer of emotion was important to 30.8% of the respondents (N = 1,911).

Descriptive results indicate that personal experiences of value vary significantly among respondents, and the general mean-based analysis does not reveal the features of the phenomenon thoroughly. Owing to issues with the normality of several variables, we analysed the dataset using exploratory factor analysis with a principal component method using oblique rotation, as we assumed the factors would be associated with each other. Because some questions were posed only to those who reported having commented on or read the comments of others, we analysed the dataset in three phases, depending on the number of variables and observations in each. The phases were used to triangulate and confirm the results and interpretations.

In the analysis of 33 variables, identified from respondents who had both commented and read comments, and 213 observations, we interpreted five retained factors abbreviated as NV, VRC, SVC, SVS and EV. In other words, we modified the original division of values developed by Sheth et al. (1991) based on the exploratory factor analysis of five news-specific distinct clusters (constructs), which we named NV, VRC, SVC, SVS and EV (see Table 3 for detailed description on how these were indentified) . The cumulative percentage of variance explained by these five factors was 92.6%. We conducted the same procedure using the same statistical methods for 26 variables and 758 observations, omitting responses from those who reported that they had not commented on a story. This process revealed the same constructs, apart from SVC, confirming the results of the first analysis with a larger dataset. Finally, we conducted a factor analysis with 19 variables and 1,759 observations to compare the results using the widest sample size but without responses from those who reported not writing or reading comments. The analysis and interpretation resulted in the confirmation of three factors (i.e. NV, SVS and EV). When analysing the results, we observed that the EV factor consisted of positive loadings of negative feelings and negative loadings of positive feelings, resulting in the interpretation that the factor reflected the negative EV experienced by the respondents.

Table 3. Value factors and highest variable loadings.

Factor Name	Code	Variables with Highest Loadings
News Value	NV	A Usefulness*** A Informative*** A/C Social justice*** A/C Social usefulness*** RC Coherence from comments*** A Personal proudness*** A Personal duty*** A Identity-building**
Social Value of Commenting	SVC	C Habitual commenting* C Curiosity fulfilling* C Emotion sharing* C Advocative* C Adding perspective* C Personal relevance*
Social Value of Sharing	SVS	S Emotion expression*** S Social statement*** A/C Social curiosity***
Value of Reading Comments	VRC	RC Social duty** C Usefulness** RC Informativeness** C Personal relevance** C Opportunity* A Coherence* RC Empathy*
Emotional Value	EV	A Irritating*** A Bothering*** A Confusion* RC Irritating comments* A/C Uplifting*** (negative loading) A Relaxing** (negative loading)

The number of stars indicates whether the results were repeated in several analyses or in only some of them. Each star represents one analysis. For commenting, only one analysis was available, and for reading comments, only two analyses were available. A = reading article, A/C = reading article or comments, C = commenting, RC = reading comments, S = Sharing.

In analysing the correlations between these five factor-based variables, there was a weak but statistically significant (p < .05) negative correlation between NV and EV. We found a strong positive correlation between NV and SVS, a weak-to-moderate correlation between NV and VRC, and a moderate-to-strong correlation between VRC and SVC among those who participated. Based on the analysis, we interpreted that commenting, sharing and reading comments had distinctive value combinations compared with NV or EV, which were awakened by the story. Based on the analysis of factor loadings, we conclude that EV is a distinctive value construct from NV.

According to the quantitative findings, reading the article was perceived as the most value-inducing behaviour. Commenting was perceived as somewhat valuable on average and very valuable by those who commented. About one-third of the respondents considered commenting worth their time and effort, compared with nearly 40% who considered reading comments worthwhile. In summary, personal experiences varied significantly.

After the quantitative analysis, we began the qualitative analysis by inviting participants to respond with free-text answers to the statement, “If you wish, you may elaborate on what you thought was valuable about this news story” (N = 1,047).

During consecutive coding rounds, it became evident that many respondents had a sense of responsibility to keep themselves up to date on current events. This was evident in the way they pointed out their appreciation for high-quality journalism, which helped them meet this responsibility. This appreciation was most apparent in investigative journalism, but it was also present in other types of journalism. The respondents specifically valued the ability of journalism to address societal issues.

Through the qualitative analysis of the free-text responses, it became evident that a distinction was needed between positive and negative emotional values, supporting our quantitative results. The respondents found their ability to be touched by a news story valuable. Often, the emotions evoked were sadness or anger. Both positive and negative EVs were often linked to the respondents’ previous personal experiences and confirmed their previous thoughts. An element of nostalgia was often present; this was most evident in an article on the 1990s television series “Kummeli”, of which many Finnish people have fond memories. The respondents also highlighted the importance of having the opportunity to comment, owing to its democratising effect.

In summary, the free-text answers allowed us to confirm the two-sided (positive and negative) nature of EV and the importance of the opportunity and duty to make and read comments, which we interpreted as part of the social dimension of value already identified in the quantitative phase.

The free-text responses (N = 188), designed to control the survey design and reveal possible new value dimensions of commenting, provided three categories: socialisation, self-expression and sense of responsibility. However, as these were all present in the quantitative questions, this finding does not reveal a need to substantially alter our approach in the quantitative part of the study.

When asked to elaborate on the reasons participants had for sharing news content, three key motivations were present (N = 166): sharing one’s emotional experience with others, informing others and self-expression. These motivations were all included in the questions of the quantitative part of the study.

Finally, we analysed the interviews to understand and explain our value constructs and dimensions. The results indicated that the news-reading behaviours of the interviewees were reflective of their identity. For some interviewees, the element of carrying out a responsibility when reading the news was present. One interviewee stated that they consumed news stories on a certain topic so that they could use the knowledge in their conversations with family members. The interviewees also highlighted the choice of news source as important. Some felt the need to distance themselves from certain types of news sources, most often mentioning tabloids as less favourable.

The interviewees indicated several motives for news commenting: the news story felt personal, a desire to voice one’s opinion, a sense of responsibility and to show appreciation for the journalist or the subject of the story. The desire to voice one’s opinion was particularly strong if no comment containing a similar argument had been posted by someone else. In some cases, this urge was linked to a sense of responsibility if the commenter felt that either the news story or its comments contributed to some form of injustice, making the commenter feel obligated to comment. One interviewee said that they had commented to make their view visible, as they had not seen other comments reflective of their opinion.

When asked specifically if it felt important that they commented, some interviewees said “yes”. They also felt compelled to write something when no one else had commented on the news story. Some reported feeling a sense of responsibility because they believed they had information relevant to the story. One interviewee said that they decided to comment because they were entitled to do so. Some expressed that their opinions were as valuable as the opinions of others and were thus worthy of presentation.

We asked the interviewees who reported having commented on whether they thought their comments would have been useful to others. Some said that they hoped so or that they believed in the potential while others had not thought about it from any particular perspective or did not believe their comment would benefit others. Most who reported engaging in commentary recognised that they had benefited from the act, describing feelings of relief and satisfaction. Some interviewees stressed that the act of commenting allowed them to respond to comments they disliked, which brought them satisfaction. One interviewee explained that they had gained a clear conscience through commenting, while another claimed that commenting helped them structure their thinking. Social gratification was also present in the interviews.

The findings from both the open questions and the interviews suggest that NV, SVC, SVS, VRC and EV are distinct categories of perceived value that resonate with a wide variety of intentions, some of which are beneficial not only from the perspective of individuals but also groups of people and the wider society. Commenting requires more effort than reading and is especially emphasised from a societal perspective.

How do the contributions of other users affect experienced values?

After identifying the value categories associated with news articles and collaborating activities enabled by the platforms, we explored how these activities affected the values perceived by other users. We performed statistical analyses on the differences in values experienced, depending on whether the respondents had read the comments attached to the news story and how and when they had done so.

First, we analysed the dataset of 1,730 observations regarding the correlation of reading comments to NV, SVS and EV experiences For NV, reading comments had a negative effect; for SVS, no significant positive relation with reading comments was found; and for EV, the relation was positive and weak, reflecting the low predictive power of negative emotions. In another analysis, participants who did not read the comments experienced slightly higher NV but lower SVS and (negative) EV compared with participants who read them (See Table 4 for a matrix of correlation analyses). Furthermore, we compared replies that occurred outside and inside the given commenting period and found that reading comments did not increase, on average, the total value of the news article.

Table 4. Pearson’s Correlation Among Factors

Factors	NV	SVC	SVS	VRC	EV (negative)
NV	1
SVC	(a) 0.46*	1
SVS	(a) 0.61* (b) 0.66* (c) 0.59*	(a) 0.50*	1
VRC	(a) 0.28* (b) 0.32*	(a) 0.51*	(a) 0.31* (b) 0.55*	1
EV	(a) -0.25* (b) -0.22* (c) -0.24*	(a) 0.03	(a) 0.02 (b) 0.01 (c) 0.05*	(a) -0.18* (b) -0.06	1

Note. The table above shows the strength of the correlations between factors. Letters a (N = 213), b (N = 758) and c (N = 1759) indicate the three different factor analysis results. Star (*) indicates statistically significant correlation at the level of p < 0.05.

According to the survey results, those who claimed that they did not read the comments found the news content more satisfying, reporting that they were likelier to discuss the content with others when they did not read other people’s comments. Likewise, those who read the comments were less relaxed and felt that the content was less worth their time than those who did not.

We also performed a two-way analysis of variance on the respondents who came early to the story and responded to our survey compared with those who came later when more comments had been written and the story was older. The analysis was conducted by counting the average time in seconds from story publication to the survey responses and dividing the data into two groups: earlier and later than the calculated average. We found no significant difference among the groups in NV, but we found that those who came to the story later experienced a higher SVS.

Finally, we conducted a series of multivariate analyses on the relations between value constructs the users experienced and (1) the story type (i.e. soft news, hard news or opinion), (2) the time of the response compared with publication time, (3) the respondent’s age and (4) the respondent’s gender. We found no significant relationships with the response time, indicating that the values did not increase when other users’ contributions accumulated. In the story type analysis, we found only an indicative result that opinionated articles were related to higher VRC (p = .062) and lower negative EV (p = .036) than hard news.

Of the behaviours, sharing had a strong and significant positive correlation with NV (p = .000), SVC (p = .008), VRC (p = .000) and SVS (p = .000). Commenting had a moderately significant positive coefficient with SVS (p = .001) and VRC (p = .000). Reading comments had a moderately negative coefficient for NV (p = .000). For (negative) EV, both reading comments (p = .000) and commenting (p = .034) had positive coefficients, indicating an increase in negative emotional value perceptions (See Table 5).

Table 5. T-Test of differences in value constructs by behaviour.

	Did not share	Shared or will share	Did not comment	Commented	Did not read comments	Read comments
Obs	146	45	538	220	748	982
NV Mean SD Var	−0.305* 0.963 0.927	0.753* 0.590 0.347	0.020 0.955 0.911	−0.049 1.000 1.000	0.166* 0.926 0.858	−0.141* 0.976 0.954
Pr(|T|>t|)	0.000		0.381		0.000
SVS Mean SD Var	−0.291* 0.803 0.664	0.811* 0.798 0.636	−0.096* 0.908 0.824	0.234* 0.899 0.808	−0.064* 0.952 0.906	0.028* 0.897 0.804
Pr(|T|>t|)	0.000		0.000		0.041
EV Mean SD Var	0.057 0.883 0.780	−0.073 0.998 0.996	0.007 0.881 0.777	−0.018 0.909 0.826	−0.123* 0.934 0.873	0.091* 0.842 0.708
Pr(|T|>t|)	0.435		0.721		0.000
VRC Mean SD Var	−0.132* 0.933 0.869	0.386* 0.664 0.441	−0.130* 0.967 0.935	0.319* 0.834 0.695	N/A	N/A
Pr(|T|>t|)	0.000		0.000
SVC Mean SD Var	−0.118* 0.988 0.975	0.334* 0.771 0.594	N/A	N/A	N/A	N/A
Pr(|T|>t|)	0.002

However, in the post-estimate analysis of multivariate linear models, several tests indicated that the relations were nonlinear and that the models had only moderate explanatory power, leaving room for other, perhaps deeper, explanations. Another concern is the direction of causality, which statistical tests cannot determine. Whether the experienced value affects behaviours or the behaviour affects the experienced value remains an open question.

We then performed a qualitative analysis of the open-text responses and interviews to find possible explanations and contextual clues. Based on the open-ended question (N = 354), we observed that comments affected the perceived value of the reading experience. Many participants appreciated how the comments highlighted perspectives that were not presented in the story. One respondent emphasised that comments demonstrate how people perceive the same news story differently, thus promoting empathy. Those who claimed that they did not gain added value from reading the comments reported that this was due to their poor quality.

The responses highlighted the respondents’ emotional reactions to the comments, which can be divided into positive and negative types. Positive emotional responses include feeling touched and supported by comments. Negative emotional reactions stem from readers perceiving the content as shocking or otherwise heavy. Respondents either felt irritated by the comments or were forced to stop reading after a while.

The responses to the open-ended statement revealed the respondents’ behavioural patterns while reading the news story comments. Three respondents stated that they read the comments despite not expecting them to bring any positive value. Two respondents believed that it was always worthwhile to read the comments.

Most interviewees claimed that they received some benefit from reading the comments. Many felt that doing so diversified their perspectives. The additional information provided by the comments was also valued. None of the interviewees claimed that they did not benefit from reading the comments, and one interviewee said that reading comments is important if one plans to add one’s own comments.

We also explored the interviewees’ emotional reactions to the comments. Most reported experiencing feelings of disgust, fascination, irritability, sadness, fear and worry when reading the comments, while two reported feeling positive after reading like-minded comments. Some interviewees stated that they read the comments to see if their own opinions matched those of other commentors.

A few interviewees felt that it was especially harrowing to read comments on news stories with a particularly polarising topic. Comments on news items covering lighter topics, such as culture, were deemed easier to read. We also found that reading the comments provided a safe place to judge other people. In this case, excessive comment reading was deemed negative.

The interviewees’ responses helped broaden our understanding of the behavioural aspects linked to reading comments. Some interviewees elaborated on their news-sharing behaviours, placing the emphasis on accounting for the audience and, more specifically, on the personal relationship with the person to whom the content was shared. Furthermore, their sharing decisions were often made with a desire to inform others. Additionally, the aspect of seeking emotional support from others was present, owing to comments that were deemed shocking.

The following lists summarise the key results to our research questions:

RQ1:

What value constructs do news consumers gain or lose when reading the news or the comments attached to it and when actively commenting on or sharing the news item?

• Most people who read the article (N = 2,020) did not share (N = 361) or comment (N = 380) on the news product, but more than half of the respondents read the comments.

• Reading the story was considered substantially more valuable than commenting and reading comments. Nevertheless, a third of the respondents considered commenting worth the time and effort (responses 7–10).

• On average, readers did not report high values connected to reading the comments.

• Several results reflected high-value perceptions of commenting among the commenters.

• The opportunities to comment and read comments were considered valuable, even by those who did not participate in these activities.

• High (negative) EV of the story predicts lower NV of the story, and high NV predicts lower negative EV experienced by readers.

RQ2:

How do the actions of other users affect the value constructs that news users gain or lose?

• Commenting, sharing and reading comments have distinctive value combinations compared with NV or EV, which were awakened by the story. EV is a distinctive value construct from NV.

• Reading other users’ comments did not increase, on average, the total value of the news article. For NV, reading comments had a negative effect. Those who read comments were less relaxed and felt that the content was less worth their time compared to those who did not read comments.

• Reading comments and commenting predicted, on average, a higher (negative) EV, indicating that either negative feelings spark commenting or commenting sparks negative feelings.

• Personal experiences about the value of other users’ contributions varied substantially. Among the respondents, there was a substantial subgroup that highly valued user contributions, but others considered these to be value-subtractors. Of the respondents, 39.6% said reading the comments was worth their time and 30.7% found the comments useful.

• A moderate-to-strong correlation was found between the value of commenting and the value of reading comments, indicating that the values experienced by these activities are intercorrelated among those who actively participate in these actions.

• We found a strong positive correlation between NV and SVS, indicating that stories with high news values were considered shareworthy. The positive relationship between EV and SVS reflected the predictive power of negative emotions in sharing. Sharers get slightly more value from news articles that have comments.

• Those who came to the story later experienced a higher SVS. However, in general, the values experienced did not increase when other users’ contributions accumulated.

Discussion and theory building

After statistically assessing whether the value of news products increased when they were used by others, the results were mixed regarding the rivalry or anti-rivalry of news as goods. The value gained from reading only the story was substantially higher than the value gained from reading the comments. Although many of the interviewees claimed to have benefitted from reading the comments, the statistical analyses revealed that only a fraction of the users considered they had benefitted. Of those who read the comments, only 42.9% said that it was important to do so, whereas 57.8% of all respondents said that it was important to read comments if they existed.

Although the interviewees vividly described various negative emotions and frustrations related to the comments, the social value of sharing showed a slight increase among those who commented when the comments were read.

These mixed results were caused by several groups of participants using the news product differently while creating various value shifts differentiable by the group. This outcome is in line with previous research. Seven such groupings of user profiles were identified in the Yle comment sections by Kangaspunta (2021). The key finding that some news product values were anti-rival for some groups is masked if we look solely at the average values in different cases.

We can identify distinct user groups based on what users do with the news story, which can be consumed by reading it, by reading its comments and by sharing it or commenting on it. Commenting consumes the news product in ways that only some readers and sharers notice, owing to the structure and format of the product, wherein the results of produsage (i.e. the practice of being both a user and producer of online content; Bruns, 2014) of the commenters are presented separately below the professionally produced news piece. News consumption has been extensively studied through the lenses of uses and gratifications theories (Katz et al., 1973). For example, news consumption is driven by consumers’ sense of anticipated gratification, including entertainment, information, escapism and socialisation (Chung & Yoo, 2008; Diddi & LaRose, 2006).

In this article, we have identified the social values of sharing and commenting as distinctive constructs compared to the value of the news story. Although the motive of socialisation is a weaker indicator of behaviour than informational value among newsreaders, its role is more significant when looking at those who write news commentary and share content within their social circles. Reading news satisfies the need for information and entertainment, while facilitating a conversation about the news responds to our need for socialisation. The search for social gratification leads to news items being shared in social circles via digital platforms, and commenting on the news mediates the effect. We also investigated the role of emotions as a value of the news construct. The emotional nature of journalism has long been acknowledged, as news organisations have historically exploited scandals and subversive topics with a shock value that sells copies (Beckett & Deuze, 2016). The role of emotions in consumer behaviour has been studied extensively. However, the value consumers find in emotional provocation has not received a great deal of attention until recently (Wahl-Jorgensen, 2020). Thus, emotional factors are often discussed as company perspectives with regard to how tactics can be used to attract readers and promote engagement.

Our results show that news value remains a dominant factor in its production. When a piece of news has high news value, user contributions do not play a major role. Commenting and sharing are separate value constructs that mainly consist of social value dimensions. Furthermore, the opportunity to comment is a strong value dimension by itself.

The results also indicate that news value changes as other users use the news product in different ways. Online news products are not always nonrivalrous goods; instead, they behave in complex ways – both subtracting and accumulating values – when different users anticipate, prefer and emphasise their changing values differently. This effect was demonstrated in this study via the commentary phenomenon. Additionally, news value subtraction and other value-shifting activities related to commenting are nonlinear. We were unable to demonstrate whether an increase in the number of comments shifted the values further. One possible explanation for this being non-present is that the value of commenting lies in its social signalling, approval, importance and other values. From this perspective, the results demonstrate how social factors influence consumption behaviours and decisions differently. Our consumption behaviour is affected by what we believe is being signalled to other people. Additionally, some products possess interactive features that accommodate social interactions. Both factors are present in news consumption.

Consumers view the popularity gained by a product as social proof of its quality and value (Goedegebure, 2019). This value is therefore directly linked to the extent to which people recognise the product. The visibility of online information products, such as news content, increases as they are shared. Thus, the social value of different news sites is determined significantly by their popularity (Nelson, 2021).

According to our results, a high news value is connected to a willingness to share. People will not share news articles if they believe they will be of no interest to people on their social networks. News consumers tend to share content that morally and emotionally matches their group identity (Brady et al., 2020; Dafonte-Gómez, 2018; Johannesson & Knudsen, 2021).

According to some previous studies (Kangaspunta, 2021), the motives behind writing comments range from avoiding boredom (escapism) to gaining social recognition. The opportunity to comment is viewed as a socially significant function, as it enables the expression of opinions and self-affirmation. The exchange of viewpoints is central to the commenting function, as it allows readers to compare opinions and engage in deliberations with fellow readers. Despite these shortcomings, news consumers find the existence of the comment field necessary (Engelke, 2020).

Those who write news comments are often driven by social motives (Wu & Atkin, 2017), whereas reading them seems to fulfil entertainment and cognitive needs. However, cognitive needs are often not fulfilled to the desired extent if the perceived quality of comments is deemed to be low. Even comments perceived to be of low quality may receive attention but perhaps only for entertainment (Springer et al., 2015).

The interviewees emphasised their right or opportunity to comment, as it allowed them to practice their freedom of expression, often anonymously. While news comments are valued for their democratising effect and authenticity, they are also viewed with scepticism because of the absence of quality controls (Wahl-Jorgensen et al., 2010). People may perceive complementary information regarding a news story by reading the comments, and some are motivated by showcasing unique perspectives. Anonymous commenting is an interesting form of performance art because readers inevitably form an image of the commentator in their minds. Indeed, commenting can provide an opportunity to express oneself without the pressures or expectations created by in-person social circles (McCluskey et al., 2015).

Finally, our results indicate several important shifts in value regarding the news product as it is used. When a news product is consumed by commenting on it, the act increases the social value of sharing for sharers, raising the likelihood of sharing; decreases the news value for readers, meaning that the users considered some of the aspects included in the news value – such as the practical information or advice, the value of the information, the understanding of the injustice in society and the usability of information – to be less than what they got from the piece of news without the comments; and may decrease the negative emotions felt by readers.

Thus, consuming a news product with commentary is rivalrous in terms of news value, but it is anti-rivalrous in terms of the social value of sharing. By that, we mean that the news product’s news value, experienced by the users who follow the commenter, decreases when the news product is commented on and that this act simultaneously increases the social value of sharing for those who consume the product after someone has commented on it. However, the value does not increase when other users’ contributions accumulate. Regarding emotional value, a commented news product can be either rival or anti-rival, depending on the emotional preferences of the reader and their background. Before determining whether perceived emotional value increases with other people’s engagement, we must first identify how emotional value manifests in the context of news content. We hypothesised, but could not demonstrate, that emotional values would differ between different types of users (Kangaspunta, 2021).

The ability of news content to provoke emotional responses affects the attention it receives and the time readers spend on a given platform. Thus, there are economic incentives for news companies to produce emotionally provoking content and a “fellow-feeling” (Bakir & McStay, 2018) experience for readers. Even if negative emotions are felt, many readers intentionally seek out such content. Previous research shows that news consumers claim to favour positive content, even when their actual behaviour suggests otherwise (Trussler & Soroka, 2014). Negative news content appears to be the most dominant, owing to demand-side factors (McCluskey et al., 2015). The high demand for negative news may indicate that readers perceive the value of such content as high. However, just because negative news content can attract attention does not necessarily mean that readers perceive it to be emotionally valuable. Despite its attractiveness, negative news may strain people instead of enriching their lives, even when it is read for entertainment. Excessive consumption of negative news content can, in the long run, distort the reader’s worldview, nullifying the intended objectives and informative purposes of the news. However, when measuring a reader’s perceived value of news content, the long-term value is undetectable.

In summary, using a news product can both consume and contribute value, and contributing to a news product can consume value from other users because articles provide several categories of value based on reading, commenting and sharing. Fulfilling these value desires, especially those related to commenting, changes the value composition of the news article and impacts the value it provides for other needs.

Designing news business models for anti-rival consumption

Online news products are not always nonrivalrous goods. In the case where a news consumer comments on a news story and reduces its news value, the news product behaves rivalrously; that is, its value is subtracted by use. Similarly, commenting on the news product may increase its social or emotional value for some users. From this perspective, the news product behaves anti-rivalrously. On the other hand, the excludability of a news product depends on the producer’s role. The context researched here was a public broadcasting company aiming to produce nonexcludable public goods.

Understanding that the product type changes based on different consumer roles is significant and provides an important link to media platform business modelling. When such models are applied, various consumer value constructs can be viewed through the eyes of participants on different “sides” of the platform market, which may enable new kinds of business models for media corporations. Contemporary models support free-riding and are largely unable to allow equitable capturing and sharing of value. Sustainable media platform business models can be created by purposefully manipulating the value exchange between user groups. This requires clever strategies to govern the interactions and externalities among participants, depending on the product category (see Table 6 e.g. common good, private good, network good or symbiotic good; the latter two category names introduced by Nikander et al. (2020)).

Table 6. News product categories and governance approaches.

Case	Governance
A nonexcludable news product used by commenters is a common good for those reading the comments.	Common goods have a free-rider problem and need rules to limit their use or transfer value from the consumers (in this case: commenters) to comment readers.
An excludable news product used by commenters is a private good for those reading the comments.	Markets can be established for private goods, setting up cost structures to capture value from consumption.
Excludable news products behave as network goods when an optimal number of users generate social (sharing or emotional) value. This kind of behaviour is likely to exist in the commenting section, where a small number of commenters can have meaningful discussions. Too many comments lead to information overload and other undesired outcomes.	Intelligent pricing strategies can support optimising network goods. For example, auctions can be used to find the optimal price that simultaneously maximises the optimal number of users.
Symbiotic goods are nonexcludable and anti-rival: they gain value with use. In this research, additional comments did not seem to increase value.	Similar to a path that gets easier to walk upon the more people travel on it, when a news product operates as a symbiotic good, to maximise its value, additional captured value should be directed to reduce excludability towards anti-excludability. Hence, it becomes more valuable the more users it reaches.

As a practical example of a platform business model analysis regarding the anti-rivalry of news products and the yield for media corporations, from the results of this study, Yle should limit comment visibility to users who want to comment or share. Doing so would help maintain the product as a public good, as it would not reduce social value for those who prefer not to use it for social interactions. Furthermore, Yle should consider performing sentiment analysis on the comments and categorising the results based on their controversy so that users searching for negative emotions or conflicting views can easily find them (and those wishing to avoid them can act accordingly).

Implications

Our study contributes substantially to the emerging theoretical discussion of anti-rivalry by introducing and testing it in the context of news and media business model design. It also contributes to platform theory by adding a layer to the value-creation logic regarding how produsage affects the value of the product itself instead of increasing the value of the platform. Furthermore, empirical testing of the concept of anti-rivalry has been lacking, and this study partly fills this gap. However, there are significant areas that should be covered regarding how user-perceived value behaves when news products are used and how this behaviour should impact the business model designs of news providers.

From the news user perspective, further evidence is needed to understand how users gain different emotional and social values from news products. For example, a future study should answer, “What different kinds of user groups exist in news consumption, and how do these user groups interact in value creation?”

From a media firm perspective, improved insights into how users of media products transfer, increase or decrease product value as the products are consumed are expected to provide crucial tools for business model designs. The separation and filtering of user contributions for specific groups based on varying preferences are possible and may be profitable in the digital news ecosystem.

Validity, reliability and future research

As this research used a mixed-methods approach, the validity of the methodological approaches should also be considered separately. While the logic of the research was found to be sound, the results do not conflict internally, and initial contradictions between the results and the method have been explained, there are important limitations to the internal validity of this research. First, the sample of news articles and respondents was necessarily biased. The research period of two weeks was insufficient to cover all kinds of topics or provide equal weight to the results. In addition, only a few of the news articles had comments enabled and were thus included in the survey. Furthermore, the sample of respondents is based on self-selection because the survey link was published on a news website and people took the survey voluntarily. Second, Yle users are not a representative sample of general Finnish news readers. The online reach of the Yle website is 44% of the Finnish population (Newman et al., 2021), but survey respondents were leaning towards older people, with 40.4% self-reporting to be over 60 years old. Respectively, 29.4% of the Finnish population is over 60 years old (Tilastokeskus, 2021). These issues were accounted for in the external validity.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/16522354.2023.2189426

Additional information

Funding

The work was supported by the Media-alan tutkimussäätiö [20210044].

Notes on contributors

Atte Jääskeläinen

Atte Jääskeläinen is a doctoral candidate at LUT University in Finland. He was a professor of practice at LUT University from 2018 to 2020 and conducted visiting fellowships at LSE London (2017–2020) and the Reuters Institute at the University of Oxford (2017–2018). He has held the position of Director General of Higher Education and Research Policy at the Ministry of Education and Culture in Finland since 2020. He was director of news and current affairs at the Finnish Broadcasting Company from 2007 to 2017 and managing director of the Finnish News Agency from 2004 to 2006.

Johannes Koponen

Johannes Koponen is a doctoral candidate at the University of Helsinki. He has advised the public and private sectors on platform business models, media business and strategy for over 10 years. He is a member of the Council for Mass Media, a self-regulating committee that interprets good professional practices in journalism in Finland. In the past, Johannes served on the Finnish Prime Minister's Office's Strategic Coronavirus Communication Steering Group, worked with the renowned think tank Demos Helsinki, and founded three startups. Notably, two of these startups have received prestigious media innovation awards.

Vera Djakonoff

Vera Djakonoff is a graduate student at the University of Helsinki (UH). Her academic background is in political science and communication studies at UH and at University College London. She is a consultant in a prominent think tank, where she specialises in the analysis and evaluation of the interface between policy and communication.

Notes

1. https://yle.fi/uutiset.