First Monday

The contribution of online news attributes to its diffusion: An empirical exploration based on a proposed theoretical model for the micro-process of online news adoption/use <br>by An Nguyen



Abstract
This paper develops and tests a theoretical model of the online news adoption/use process, based on expectancy–value and innovation–diffusion theories, to examine the predictive power of nine common features of online news in its potential development. Using data from a national survey of online news uses in Australia, the study finds that while some online attributes such as its immediacy and content–richness features have a strong impact on the way online news is adopted, used and integrated into daily life, its unique attributes such as customisation and interactivity still have a limited effect. Also, despite that online news attributes tend to be integrated together in online news packages and thus can be unintentionally used without motivation, there is strong evidence to suggest that each of the nine online news features is substantially used because it is consciously appreciated and evaluated.

Contents

Introduction
The effect of online news attributes on its adoption/use: The need for a continually self–reflexive model
An expectancy–value approach to media use
A suggested model of the online news adoption/use process
Research questions
Methods
Results
Conclusion: Online news attributes and its future diffusion

 


 

Introduction

Empirical research has found the positive correlation between the much–touted “power” of online news and its adoption/use (Abdulla, et al., 2002; Chan and Leung, 2005; Gunter, 2003; Nguyen, et al., 2005; Nozato, 2002; Salwen, et al., 2005; Schweiger, 2000; Washingtonpost.com, 2005; Weir, 1999; Wu and Bechtel, 2002). However, most of this research body tends to be descriptive or based on loose theoretical linkages, leaving untouched the process in which online news attributes affect its adoption/use behaviours. This paper addresses this situation, examining the predictive power of common online news features for different aspects of its adoption/use via advancing a theoretical model of the micro–individual online news adoption/use process. It first reviews and defines a key weakness of Rogers’ (2003) linearly progressive five–stage innovation–decision process (knowledge, persuasion, decision, implementation, and confirmation) to highlight the need for a non–linear model that sees the online news adoption/use process as a continually progressive and self–reflexive process. The paper then draws on expectancy–value theory in uses and gratifications to address this theoretical requirement, with a review of the strengths and weaknesses of this theory, before integrating it with the diffusion perspective to propose a corrective model depicting different pathways of online news adoption/use. This model provides the basis for the author to then use data from an Australian survey to examine the extent to which and the way in which the much–touted features of online news contribute to its diffusion.

 

++++++++++

The effect of online news attributes on its adoption/use: The need for a continually self–reflexive model

It is almost taken for granted that the attributes of an innovation form a group of decisive determinants of its diffusion. However, the process in which innovation characteristics formulate adoption and use behaviours is not universally understood and can vary from one innovation to another (Hall, 2003; Reagan, 2002; Rogers, 2003). Different disciplines have different ways of looking at this micro–individual adoption process, each with distinctive conceptual problems [1]. An oft–mentioned model is Rogers’ (2003) five–stage adoption process, summarised below:

Although receiving some support from previous research, this five–stage model has to be treated with care in studying the diffusion of new media, especially the many free and ubiquitous online services such as online news. One way to see this is to discuss online news adoption in the light of what economists call “sunk” costs — costs that incur after the decision to adopt an innovation. For example, the fixed costs of purchasing a new technology cannot be recovered after it has been put into use — a substantial part of that cost has been “sunk” (Hall, 2003). Thus, from the economist’s point of view, adoption is usually an “absorbing state” — a new product is hardly abandoned in favour of an old one after being adopted. Once deciding to buy a computer, for example, one is likely to implement it (i.e., put it into substantial use) — at least for the original purposes for which it is purchased. This is because the sunk cost after adoption is substantial — if for some reason, the user is disappointed that the new computer does not work as expected and wants to buy a better alternative, she has to sell the former at a lower price (i.e., accepting a cost for nothing). Because of this high level of sunk–cost risks, the adopter usually takes a serious pre–decision evaluation process, constantly weighting the computer’s nontrivial cost against its perceived function and capacity and her related needs until uncertainties are minimised. This progress could be lengthened due to the fact that the machine is trialable only on a limited basis.

In contrast, online news is associated with almost no financial risk, and thus with much less chance for an “absorbing state.” This is because online news is a service with a diversity of sub–services that are ubiquitous and are, for the most part, offered for no charge on existing technological platforms. Those who have access to these technologies can easily try them. With no further purchase of any device and almost no upfront cost (only a negligible Internet service fee paid to ISPs), Internet users could visit as many Web sites as they wish to try online news before deciding whether to make substantial use of it. That is, the sunk cost involved is trivial and not very “visible”. The adoption of online news is, therefore, not necessarily in any absorbing state. As much as it is easy to try and adopt, online news can be as easily abandoned any time. This is especially likely in the context that there is a range of established and highly accessible news sources that can substantially serve the functions of online news. Thus, when a survey respondent reports that he is using online news, it does not necessarily mean that he has substantially integrated it into his daily routines [2]. He might be just in the trial step of the persuasion stage.

In other words, there is no clear distinction between trial, determined adoption, implementation and confirmation in the case of online news: the adoption sequence might well be knowledge–decision–persuasion–implementation–confirmation to some people or even knowledge–decision–implementation–persuasion–confirmation to others. This is an ongoing and intermingled process in which different attitudes and behaviours modify each other and co–vary. Thus, a reliable forecast of the role of online news features in its potential development needs to go beyond the adoption/non–adoption dichotomy to take into account whether, and to what extent, the supposedly exclusive features of online news are used and appreciated, and how this use and this appreciation interact with each other to influence the way online news is integrated into daily life. This requires a continually cycled model in which the behaviours in different adoption stages interact with each other to affect overall use and appreciation, rather than Rogers’ linearly progressive and close–ended five–stage model. One approach fulfilling these requirements is expectancy–value theory in the uses and gratifications tradition. The next sections will review this theory, discussing its strengths and weaknesses before building a theoretical model of the online news adoption/use process, which will help to explore the extent to which online news attributes influence its future adoption, use and attachment levels.

 

++++++++++

An expectancy–value approach to media use

Founded by Martin Fishbein, expectancy–value theory is a social psychology approach widely applied to explaining behavioural outcomes in a range of disciplines. According to this theory, an object (e.g., a computer, an online news site, or a philosophy) has many attributes, to each of which an individual develops an evaluative response called an attitude. This attitude is a multiplicative product of the individual’s belief about the object’s possession of this attribute and her affective evaluation of it (Fishbein, 1967; Fishbein, 1968; Fishbein and Ajzen, 1975). As this expectancy–value approach develops into the theory of reasoned action, attitudes are seen as the direct driver of an individual’s intention, which in turn is the direct driver of behaviours (Ajzen, 1991; Fishbein and Ajzen, 1975; Malhotra and Galetta, 1999). In other words, individuals behave in a specific manner because they have had the intention to do so; and this intention is formed on their attitude to that behaviour, which in turn derives from their subjective belief in and evaluation of the potential outcomes of the behaviour. When a range of possible alternative behaviours are available, the one chosen will be the largest multiplicative product of expected success and value (Fishbein and Ajzen, 1975).

Applied to media uses, expectancy–value theory helps clarify the concept of expectation, a central element of U&G research since its inception. Galloway and Meek (1981) used expectancy–value theory to propose a process–oriented and path–goal perspective on media choice. At the heart of this proposal is that one seeks to satisfy some felt needs via exposure to a medium or content type, whose resulting gratifications will then modify later expectations of the potential outcomes of exposure to that medium/content and thus affect subsequent exposure levels within and between media. More specifically, Palmgreen (1983), Palmgreen, et al. (1980), Palmgreen and Rayburn (1985) and Rayburn and Palmgreen (1984) proposed a process depicted in Figure 1. Accordingly, any media experience (including use behaviour, behavioural intention or attitudes) is a function of expectancy — or belief, formally defined as “the perceived probability that (a media) object possesses a particular attribute or that a behaviour will have a particular consequence”, and evaluation, “the degree of affect, positive or negative, toward an attribute or behavioural outcome” [3]. This is an ongoing process: the multiplicative product of beliefs and evaluations generates some to–be–sought gratifications (GSs), leading to media consumption. The resulting gratifications from this consumption (gratifications obtained — GOs) then feedback to (reinforce or degrade) the initial beliefs about the consumed medium or content type. In formula terms:

 GSi = biei
  
where 
  
 GSi = the ith gratification sought from media object X
bi = the belief that X possesses some attribute or an X–related behaviour
ei = the affective evaluation of that particular attribute or outcome.

 

Figure 1: Expectancy-value model of media consumption

Figure 1: Expectancy–value model of media consumption.
Source: Palmgreen and Rayburn (1985).

 

Unlike the linear adoption process in diffusion theory, expectancy–value provides a continual and circular mechanism to explore how the persuasion and implementation of new media like online news services intermingle and affect each other in the progress toward full adoption and substantial use. This is partly because expectancy–value is better and more specific in explaining belief formation — the constitution of “the primary information components determining the seeking of gratifications”. In particular, Palmgreen (1983) and Palmgreen and Rayburn (1985) recognised three ways of belief formation:

As will be seen shortly, these pathways of belief formation are quite important in the case of online news. For now, from the concept of descriptive beliefs, we could expand expectancy–value theory to include what is called the exposure–learning model (McLeod, et al., 1982), which posits that one can learn unexpected valuable attributes while seeking some other attributes through exposure to a media object. This happens in circumstances where what is originally sought from a source is not what is received but the unexpected or unplanned outcome is another valuable attribute. Put simply, an individual uses a media object to seek a highly valued X attribute but during exposure, he does not find X but Y, an attribute that he also highly values but is not aware of initially; and thus instead of abandoning the content or medium because its lack of X–related gratifications, the individual still continues to use it for Y–related gratifications. For example, during an election campaign, an individual logs on the Internet with the motive to enhance or reinforce his judgement of candidates’ stand on social issues but then finds that the medium is not as helpful as expected in this regard. This, however, does not necessarily stop him from continuing to use the medium because throughout his contact with it, he increasingly enjoys, for example, its immediacy–enabled horse–racing quality in covering the election. Obviously, this exposure–learning process is very important for the success of a newly introduced media service like online news.

Although it has been successfully applied in numerous studies of media uses, especially television use, expectancy–value theory is not without limitations. One fundamental problem is Palmgreen and Rayburn’s conceptual treatment of GSs as equivalent to attitudes. They argued (without any explanation) that both the seeking of gratifications and attitudes — originally defined by Fishbein and Ajzen (1975) as “generalised predisposition(s) to act in a consistently favourable or unfavourable manner toward an object” — could be predicted independently by the biei product. This “is a logical impossibility”, according to Stanford [4], who viewed GSs as not equivalent to attitudes but as something “approximating Fishbein and Ajzen’s behaviour intentions”, i.e., the model should go more like this: attitudes (sum of biei) GSs exposure. In this light, the sum of biei (attitudes) is more like the concept of perceived attributes in innovation–diffusion theory than GSs in expectancy–value theory. This crucial distinction between attitudes and GSs will be considered later in building a theoretical model of the online news adoption/use process.

Another potential problem is that expectancy–value theory ignores the fact that exposure is not always a logical outcome of gratifications seeking. Media use can be unintentional in many circumstances. This might take the form of passive use due to the structure of media provision. For instance, there is a high chance for being exposed to unwanted content — e.g., when following the TV news bulletin. Unintentional use could also take the form of ritualised, convenience–based use of media services that are placed a neutral value on. An individual who neither likes nor dislikes a talk show might well watch it if it is convenient to use, although she would not care much if the show is not on at the time she is available for consumption. Expectancy–value theory addresses values in a negative/positive dichotomy with deliberate attitudinal and behavioural outcomes and largely ignores these neutral values. In sum, actual media use is a combination of three use forms — active, passive and ritualised — i.e., it is not only a product of active gratification–seeking as expectancy–value theory posits. For the same individual, media use might be active (motivated), passive or convenience–based, depending on particular circumstances. Thus, media exposure and gratifications sought are not always complementary; sometimes they can even form a negative relationship — e.g., when one becomes dissatisfied with being exposed to too many unwanted items in a TV news bulletin. These issues will be integrated in the model of online news adoption/use below.

 

++++++++++

A suggested model of the online news adoption/use process

Based on the above discussion, a more specific model for explaining the whole process of online news adoption/use [5] and the effect of online news attributes throughout this process has been developed for this study. This model takes into account the discussed strengths and weaknesses of expectancy–value theory and adopts the concepts of rejection and delayed adoption from innovation diffusion. In particular, the starting point here is that online news as a whole is a multiple service with diverse components, each representing one — or a combination of several — of its innovative features. Because users have a high level of control over online content, not all of them will use all or most online features. Some might adopt and use the Internet to mainly check updated news, some look for in–depth analysis, some surf for general news of the day and some might combine some or all of these news–seeking purposes. In any way, users’ subjective image of online news as a whole is based primarily on which of its attributes they have used and appreciated. Thus to understand online news adoption/use, we could start from the adoption/use process of individual attributes. This is depicted in Figure 2 and could be understood in plain language as follows:

In the first instance, online news is believed or known to possess a particular attribute or content type (X). This belief or knowledge can come from all the three sources mentioned by Palmgreen and Rayburn (1985), with each having a distinctive potential to boost the uptake of online news. First, an individual might hear about X from someone in his/her social networks (friends, families, work colleagues etc.) or from the mass media (step 1a). Online news has an enormous advantage in relation to this belief formation, thanks both to the traditional news media’s widespread self–promotion of their online operations in their offline versions and to the way peer networks can communicate about online news via online news delivery platforms. Second, one might infer something about online news from other non–news Internet uses (1b). For instance, after a sufficient time of exposure to hypertextual links in using non–news information, a potential user might assume that online news could be provided with great depth via similar links. Given the Internet’s multipurpose nature, this sort of inferential information — although requiring a certain level of cognitive intelligence — is a considerable potential driver of online news adoption.

Third, a substantial proportion of knowledge about X can be gained from unexpected exposure to it when using another online feature or service called Y (1c). Y can be a non–news service of the Internet: for example, while a user is searching for some specific information related to work or study, a news item of interest comes up, compelling her to click through. After this unexpected consumption, she discovers that this news item was posted on the Web within only, say, two hours of the reported event. Y, however, is more likely to be another online news feature. That is, descriptive beliefs in X are very likely to be formed during the implementation of another online news feature. This is thanks largely to the convergent nature of online news attributes — the trend to which different online news features are integrated in the same package. An example of this attribute convergence might be that when reading a story for updated details (Y), a user might be urged to click on the links to related stories in the body or at the end of the story — and becomes aware of the indepth coverage of online news (X). That is, her initial belief in online news’s depth is descriptive in nature, deriving from unexpectedly using it.

After the initial belief/knowledge formation stage, there are three possible outcomes:

First, if the individual has long been placing a negative value on X (step 2a), the outcome is likely true avoidance or rejection of X. However, due to structural problems beyond his control, he might sometimes have to passively use it. For instance, an e–mail bulletin with a story on a public figure’s suicide attempt is sent to a user’s mailbox at 9 AM. Because of being busy at work, the user cannot click on this headline from his e–mail alert until lunch time, when this link takes him not to the original story about what happened initially (supposed to be consumed at 9 AM) but to another story about new developments of the event (supposed to be consumed at lunch time). This might cause some discomfort because what the individual is interested in might be the background information of the story itself rather than the updated details. Indeed, it might well be because the user is too busy that he wants to save time surfing for important news of the day by subscribing to the e–mail alert service.

 

Figure 2: A suggested model for the process of online news adoption/use

 

Second, if the existing value system of the individual neither favours nor disfavours X, the outcome is more a matter of convenience and availability (step 2c). If X is readily available for use at a time the individual does not have any situational restraint, then she uses it in a ritualised way — maybe to pass time or to avoid boredom. If X is not convenient for use OR if the individual is not available, then she does not use X without any real dissonance. For example, when a person visits a news site to check the main news of the day, it is likely that the most updated news items are available for consumption, with breaking news and developing stories being placed in a separate section or on the top of the home page. Although breaking news is not really what the user wants, she would not mind clicking through to read updated stories if she has some free time to do this after checking the main issues of the day — her main conscious purpose. In this case, the user might be substantially exposed to the immediacy of the Web without appreciating it very much.

Third, and most importantly, if X is attractive to the potential adopter, the individual will develop a positive appreciation of the Internet in terms of X–related gratifications (step 2b), which can be seen as roughly equivalent to attitudes in Fishbein and Ajzen’s (1972) original expectancy–value theory or perception of relative advantages in Rogers’ (2003) innovation diffusion theory. Appreciation here is not the direct driver of exposure to X like GSs but takes the form of a general reason for which intentional use of X takes place on a daily basis More precisely, as a result of this general appreciation, when an individual feels the need for X–related gratifications, he will develop a use intention that leads to exposure to X (step 3). For example, an individual with a belief in the immediacy of the Internet will think of the Internet as a potential gratification source whenever he needs breaking news, which then leads to exposure to online breaking news. This has two alternative products:

It must be noted that the adoption/use “mini–processes” of different features need not be separated in time — they can and are likely to be temporally intermingled. Over a sufficient period, these mini–processes result in a certain set of online news appreciation — and thus a set of well–articulated needs for the whole medium. While some initially sought gratifications might not be provided, there are others well supplied during the process. When these gratifying factors are taken together over time, they form a distinctive image of online news in the user’s recognition, leading to a certain level of attachment to it, which in turn leads the user to integrate it into daily news consumption. Again, it is believed that while overall attachment to and overall use of the online news medium co–vary, use will become the outcome of attachment at some point in time.

 

++++++++++

Research questions

It is, however, not the purpose — and beyond the scope — of this study to test this whole proposed model. In conceptualising the whole process in that way, I aim at designing a theoretically sound model to explore the way in which online news attributes play a role in shaping its adoption/use and thereby to explore the extent to which the much–touted features of online news contribute to its diffusion. To arrive at a conclusion about this, the model suggests three questions concerning three specific relationships:

At the least abstract level, the model posits that the overall use of the Internet as a news medium is the summation of individual feature uses. While this sounds trivial, the complex plethora of online news features suggests that this is an important basis to explore what is actually generating traffic on online news sites today. It might well be that the use of only one or a few features contribute substantially to overall usage. At one extreme, for example, an individual might mostly receive breaking news online but devotes so much time to it that she displays a higher level of general online news use than one who uses many features in a less substantial way. This raises the first research question:

RQ1: Which used features of online news predict its overall use level?

At a more abstract level, the model posits that while an appreciated online news feature (X) is likely to be used, a used feature is not necessarily an appreciated one. There is a high chance that some people use X substantially without appreciating it as a reason for their online news use because it is just an unintentional by–product of some other intentional use. For instance, while clicking on links for in–depth coverage, one might be led to updated details of the needed information but might not appreciate the immediacy of the medium as a reason for using it because it is depth of coverage and not immediacy that he consciously seeks. The second research question derives from this:

RQ2: To what extent are online news features being used because they are appreciated?

Beyond the relationship between the appreciation of a particular feature X and its corresponding use is one between the appreciation of X and the use of and attachment to online news as a whole. The model implies that because the convergent nature of online news features might lead to ritualised or passive uses, the use–use relationships underlining the first research question explain what drives traffic at online news sites but might say little about why online news is used. In other words, the most important features of online news might not be those that drive traffic to news sites but those whose appreciation serves as a statistically significant driver of its overall attachment and use. Even when one appreciated feature has a significant bivariate relationship with overall attachment and use, it is still inconclusive about its predictive power until this appreciation is controlled for the effect of other appreciated features. This is because the appreciation of one feature might underline that of some, or maybe all, other features. It is quite likely, for example, that an individual cites in–depth coverage and/or immediacy as their reasons for using online news only in the context that she has taken it primarily as being convenient to use. In other words, convenience might be the primary reason and immediacy and indepth coverage are, although important, only secondary reasons. This can be likened to the fact that a car user with a priority for safety appreciates speed and comfort as reasons for using a car brand on the underlying evaluation that it has satisfactory safety–related features. This forms the basis for the third research question:

RQ3: Which appreciated features of online news drive its overall attachment and use levels?

 

++++++++++

Methods

This paper utilises data from a national news usage survey in Australia. Questionnaire forms, with cover or reminder letters, were sent out twice to 2,500 randomly selected Australians in July and August 2004. A total of 790 valid responses were received and nearly 400 were returned with notices of wrong addresses, deaths and other invalid circumstances, resulting in a response rate of 38 percent. Of the 790 respondents, 528 used the Internet and 218 used the Internet for news. In accordance with research from other countries (see Nguyen, 2003 for an overview), the online news users in this sample are more likely to be people with socio–economic advantages: 72 percent lived in a metropolitan area; 67 percent worked full–time; 67 percent were professionals/managers or white–collar workers (those in clerical/sales and services occupations); 72 percent were holding at least a TAFE/trade certificate; and, over 58 percent had a household income of AU$50,000 or more. 66 percent of the online news users were males and four out of five were under 50 years of age.

Dependent variables

Online news use levels are represented by (1) its general use frequency; and (2) the amount of time spent on online news in the most recent session. The most recent online news session was treated as a typical one because the 2004 survey was conducted during a period in which there was no event of widespread interest to attract an unusual audience size and time spending (e.g., a national election or the Olympic Games) and thus can reflect the effect of general news orientation to online news. However, care was taken during the data interpretation process because an individual’s day–to–day media use can be strongly determined on many situational factors (McQuail, 1997; Weibull, 1985). Indicators of attachment to online news include (1) perceived importance of online news in Internet usage; (2) self–identification as a fan of online news (yes/no); and, (3) perception of the Internet as “the best news medium to serve your news need” (yes/no).

Independent variables

There are two main sets of independent variables. The first includes 16 use behaviours corresponding to 16 popular news services on the Web, measured in terms of either yes/no questions or frequency questions (very often/often/not very often/never). These are used as predictors of overall online news use to answer RQ1. These variables will manifest themselves in Table 1. The second set of independent variables consists of the appreciation of nine common attributes of online news, operationalised as perceived reasons for using online news. For the purpose of table presentation, these are condensed as follows: no cost (“because I don’t pay for it”); more news choices (“because I have more news choices on the Internet”); multitasking (“because I can combine getting news with other purposes online”); in–depth/background information (“because I can look for in–depth and background information whenever I want”); 24/7 updates (“because I can check for updated news whenever I want”); customised news (“because I can get news tailored to my interest only”); “have my say” (“because I can have my say to the news media”); discussing news with peers (“because I can discuss news and current affairs with my peers”); and different viewpoints (“because I can find different viewpoints on the Internet”).

 

Table 1: Logistic regression analysis for the effect of the uses of different online news features on general online news use (n = 194)a.
(a) For frequency of online news use with 1 = Frequent use (“every day” or “several times a week”) and 0 = Infrequent use (“several times a month” or “less often”);
from (b) to before (c): dummy variables with 1 = “Yes” and 0 = “No”;
from (b) to before (d): all variables with the same categories: 0 = “Never”; 1 = “Not very often”; 2 = “Often”; 3 = “Very often”;
(d) Likelihood Ratio Chi–squared Test with 16 degrees of freedom.
Source: Australian News Usage Survey — 2004.
 Coefficientρ
Subscribe to free e–mail alerts of general newsb 1.75.02
Subscribe to free e–mail news alerts tailored to your own interests only-1.21.03
Set up a personalised page offered by Internet services and online news providers-0.72.13
Set your favourite news home page as the default front page of your Web browser-0.32.46
Use search tools to find news of your interestc-0.31.19
Get up–to–the–minute news several times a day1.41< .001
Visit a number of sites for the same news item-0.15.65
Get audio news in addition to reading0.21.49
Get video news in addition to reading0.19.56
Get video news in addition to reading0.19.56
Click on links to related stories for in–depth coverage (including background information) 0.32.28
Participate in online news polls0.69.08
Find other perspectives from sources outside the news mainstream media0.12.69
Go to an information exchange site to express your opinions-1.05.06
Receive links to news stories from your peers-0.32.48
Read weblogs0.22.71
   
LR chi2 (df 16)d57.90< .001

 

Statistical procedures

For RQ2, pairwise correlations were calculated with their statistical significance. This allows us to see the strength of relationship between the appreciation of one online news feature and its corresponding use (e.g., the appreciation of online immediacy and the use of online breaking news) and to compare this association with the cross–feature use–appreciation associations (e.g., the appreciation of online immediacy with the use of, say, hypertextual links for in–depth information; or the use of immediacy with the appreciation of depth of coverage). Some composite variables constructed during this process will be introduced when they first appear. It must be noted that while survey data do not normally reveal a cause–effect relationship as raised in RQ2, the wording of the relevant questionnaire items (see above) makes them very operatable in this study.

For RQ1 and RQ3, logistic regression models were employed for all five dependent variables. Two of these variables (e.g., online news fans and perception of the Internet as the best news medium) were originally dichotomous (yes/no). The other three dependent variables were recoded from their original categories as follows: use frequency was binary–recoded as frequent (every day or several times a week) versus infrequent (several times a month or less often); time spending as “above average” (spending more than the mean time, 14.9 minutes, in that session) versus “below average” (less than 14.9 minutes); and, perceived importance as “an important part” and “an essential part” versus “not an important part” of Internet use. All analyses were weighted according to the sex–by–age joint distribution in Australia’s 2001 Census data.

Since logistic regression analysis is possibly an unfamiliar technique, I will present the first logistic model with all essential information about coefficients and p-values both in Table 1 and the analysis text. For the purpose of condensed data presentation, later models will be displayed using the familiar star–system (*) to indicate statistical significance and be analysed with regard only to key information about statistical significance. The actual ρ–values associated with coefficients will not be presented in tables but will be mentioned in the analysis text. The result section will be split into three parts, each devoted to one of the research questions espoused earlier.

 

++++++++++

Results

Relationship between the uses of individual online news features and its overall uses

Table 1 shows the logistic regression model that answers RQ1 on the relationship between the uses of different online news services and its overall use. It provides the following information:

Some plausible explanations could be found for this pattern of effects. The positive effect of continuous use of updates on general use frequency is easy to make sense of: getting up–to–the–minute news several times a day apparently contributes substantially to the number of times news is received from the Internet in an average week. Similarly, the positive effect of subscription to general e–mail news alerts is not surprising because e–mail is the most prevailing application of the Internet so far. Meanwhile, the negative coefficient of subscription to customised news alerts does not mean that this use has a negative effect on general use. Rather, it suggests that there might be people who use online news only infrequently because this use is tied with seeking some specialised news categories that are delivered to the e–mail box on an infrequent basis. For example, one might subscribe to customised news services on a weekly or monthly basis and only (or mostly) visits an online news site when these specialised news alerts arrive at his mailbox.

Relationships between the appreciation of individual online news features and their corresponding uses

As demonstrated before RQ2 was raised, the use and appreciation of an online news feature can be unassociated or even negatively associated because, due to the convergent nature of online news features, it can be used in a passive or ritualised manner. Thus a feature can only be seen as a driver of its use when it satisfies two statistical conditions: (1) its use and appreciation must be at least moderately correlated; and (2) this correlation must be higher than that between the appreciation of this feature and the use of any other feature. In order to explore this, pair–wise correlations between perceived reasons for online news use and the use frequencies of different online news features were constructed in Table 2. Several methodological points need to be mentioned before analysing this Table. First, of the nine reasons introduced above, the cost factor was eliminated from this table because there is no real corresponding behaviour to explore. Second, the Table contains a few composite variables that were created from original use variables to simplify the data:

 

Table 2: Pair–wise correlations between the appreciation of different online news features and their corresponding uses.
Coefficients between appreciations of online features and their corresponding uses are bolded; — Not explored because there are no prima facie relationships between these variables; ns = not significant.
Source: Australian News Usage Survey — 2004.
 Use online non–news servicesUse customised news servicesGet updates several times/dayVisit many sites for same news itemClick on links to related storiesFind views from non–mainstream sitesUse multimedia contentExpress opinions on a news–trading siteExchange news with peers
Multitasking.36Ns.31.33.25.17.19.18.16
More news choices .15.44.48.26.20.27.18.23
In–depth and background information Ns.23.33.23.19NsNsNs
Different viewpoints Ns.16.29.17.35.23.23.21
24/7 updates Ns.42.43.35.17.29Ns.19
Customised news .15.26.29.24.20Ns.24Ns
“Have my say” to the news media .12.17.32.22.21.20.39Ns
Discussing news with peers .22Ns.14Ns.28.16.21.28

 

Now, looking at Table 2, it can be first noted that every correlation coefficient between the use of an online news feature and its corresponding appreciation (bolded in Table 2) is statistically significantly positive. Although none of the coefficients exceeds .48, most showed moderately positive relationships. More importantly, there is quite consistent evidence to infer that the appreciation of a specific feature does explain its corresponding use. Reading the table horizontally, it can be seen that the highest correlation coefficient in each row is almost always found between the appreciation of one feature and its corresponding use behaviour(s). In particular, the strongest association across the rows was found between:

A slight deviation from this pattern is that the appreciated immediacy of online news (“because I can check for updated news whenever I want”) has its strongest correlation with visiting many sites for the same news items (.43), rather than with getting updated news several times a day (.42). However, these correlation coefficients are almost identical, allowing us to tentatively conclude that online news updates are used because they are appreciated. Indeed, the data show that more news choices and immediacy tend to covary with each other. This is confirmed by two other facts: (1) the appreciated ability to have more news choices was found to form its second strongest correlation with getting updated news several times a day (.44); and, (2) the appreciated ability to find in–depth and background information had the second strongest association with the same behaviour (.23). This is not difficult to understand because when people look for immediacy, they might tend to (and sometimes have to) visit many news sites because (a) not all sites are continuously updated and (b) developing stories are usually short of details that stimulate further exploration.

Customisation is the only feature that deviated from the above pattern of correlation strengths. The appreciation of the ability to “get news tailored to my interest only” had a fairly weak, although significant, association with its corresponding use of customised news services (.15). It was also the weakest significant correlation across its row.

Relationships between the appreciation of different online news features and its overall attachment and use levels

As bolstered for RQ3, the defining features of online news — those that serve as a benchmark to compare different online news use and attachment levels and to explain the primary reason for online news to be chosen in a diversified media landscape — are not necessarily those that are significant in the use–use relationships underlining RQ1 but are those whose appreciation survives as significant predictors of online news attachment and use after all appreciated features are controlled for each other. Tables 3 and 4 provide insights into this issue.

Table 3 contains three logistic regression models for the effects of appreciated online news features on three aspects of online news attachment: perceived importance of online news in Internet use, perception of the Internet as “the best medium to serve your news needs”, and self–identification as an online news fan. As for the perceived importance of online news in Internet usage, the logistic model was a good fit with LR chi2 = 84.8 (ρ < .001) and yielded three significant positive predictors: 24/7 updates (coefficient = .95, ρ < .01), indepth/background information (coefficient = .66, ρ = .02) and more news choices (coefficient = .47, ρ = .05). That is to say, as the whole, online news was perceived as an important or essential part of daily Internet use thanks more to its appreciated provision of some traditionally upheld news qualities — its immediacy as well as its content depth and diversity — than to any innovative utilities that are unique to online news (such as customisation, interaction with peers or feedback to the news providers).

 

Table 3: Logistic regression analyses for the effect of the appreciation of online news features on different aspects of online news attachment (regression coefficients).
(a) Perceived importance of online news in Internet use (1 = “Essential” or “Important”; 0 = “Not Important”);
(b) Choosing the Internet as “the best medium to serve your news needs” (1 = Yes; 0 = No);
(c) Self–identifying as an online news fan (1 = Yes; 0 = No);
(d) Likelihood Ratio Chi–squared Tests with nine degrees of freedom; * ρ ≤ .05; ** ρ ≤ .01; *** ρ ≤ .001
Source: Australian News Usage Survey — 2004.
 Importancea
(n = 199)
Best mediumb
(n = 164)
Fandomc
(n = 200)
No cost0.170.280.34*
More news choices0.47*0.340.24
Multitasking-0.040.47-0.21
In–depth/background information0.66*-0.10-0.05
24/7 updates0.95**1.66***0.76**
Customised news0.29-0.170.48**
“Have my say” to the news media0.04-0.040.21
Discussing news with peers-0.060.03-0.09
Different viewpoints-0.010.310.12
    
LR chi2 (df 16)d84.81***62.22***53.83***

 

As for the perception of the Internet as “the best medium to serve your news needs” (LR chi2 = 62.2, ρ < .001), immediacy was the only significant predictor (coefficient = 1.66, ρ = .001). Meanwhile, online news fandom (LR chi2 = 53.8, ρ < .001) was the outcome primarily of the appreciation of its immediacy (coefficient =.76, ρ = .005), its ability to get news tailored to one’s interests (coefficient = .48, ρ = .01) and its lack of cost (coefficient = .34, ρ = .03). Compared across all three dimensions of online news attachment, the common predictor is the appreciated immediacy of online news.

Table 4 moves beyond indicators of online news attachment, displaying the effects of the appreciation of online news features on its use levels. As for use frequency (LR chi2 = 46.9, ρ < .001), three of the nine appreciated features serve as significant positive predictors: the non–cost factor (coefficient = .41, ρ = .01), the ability to do to multitask (coefficient = .62, ρ = .01) and immediacy (coefficient = .74, ρ < .01). The first two elements are explainable: the more activities people conduct online, the more frequently they use the Internet, which in turns leads to more chance to use online news. This might be a product of intentional use or of serendipity (being accidentally informed when being online). And of course, this use would be much more limited if online news is available with some charge. The third significant contributor is also easy to understand: the more people expect the Internet to be a good source of breaking news, the more likely people are to go online for it frequently.

 

Table 4: Logistic regression analyses for the effect of the appreciation of online news features on general online news use (regression coefficients).
(a) 1 = Frequent use (“every day” or “several times a week”) and 0 = Infrequent use (“several times a month” or “less often”);
(b) 1 = Above average and 0 = Below average;
(a) Likelihood Ratio Chi–squared Tests with nine degrees of freedom; * ρ ≤ .05; ** ρ ≤ .01; *** ρ ≤ .001
Source: Australian News Usage Survey — 2004.
 Use frequency
(n = 200)a
Time spending in most recent session
(n = 191)b
No cost0.41**0.15
More news choices-0.120.41*
Multitasking0.62**-0.50*
In–depth/background information0.08-0.12
24/7 updates0.74**0.64*
Customised news-0.090.02
“Have my say” to the news media-0.19-0.05
Discussing news with peers0.210.48*
Different viewpoints0.06-0.11
   
LR chi2 (df 16)c46.94***28.15***

 

As for the amount of time spending in news in the most recent session (LR chi2 = 28.2, ρ < .001), four variables remain as significant predictors: more news choices (coefficient = .41, ρ = .05), multitasking (coefficient = -.50, ρ < .05), 24/7 updates (coefficient = .64, ρ < .05) and the ability to discuss news with peers (coefficient = .48, ρ = .02). The three significant positive predictors in the model can be explained as “qualitative” use reasons that tend to consume time. In other words, people spend more time on news in a typical session because (1) they receive more news from more sources; or (2) they consume more breaking stories; or, (3) passing news of interest to peers is often accompanied with expressing ideas and writing comments, which certainly take more time than merely reading the news.

Meanwhile, the negative contributor – the Web’s multipurpose nature – can be seen as a “quantitative” element for several possible reasons. First, it might be the case that while the ability to do multitask contributes significantly to the density of online news usage (i.e., use frequency), it also reduces the time spending on news in any specific session. In other words, as people do many things frequently on the Web and get news according to this frequent use, there is less news suiting their interests each time they log on news sites, resulting in less time spending than that of less frequent visitors to these sites. Second, it could be because those who combine news with their Internet uses might be too busy and cannot spend much time on news in each section. The reason for these people to get news on the Internet might well be just to save time. The third possibility is the same as what Jacob Nielsen (1997) said about why people often scan/skim online news rather than read it word by word: because the Web is a “compulsive” environment, people usually do something online with the feeling that they need to move on because there are millions of other things out there. That is, as users get news because it is combinable with other activities, their news use might be restrained by the drive to do other non–news activities.

Before reaching a conclusion about the effect of online news features on online news use, there remains another task. The model in Figure 2 suggests that over a certain amount of time, the appreciation of online news features does not directly affect its use but indirectly affects online news attachment that leads to its use. Is it supported by the 2004 survey data? In order to explore this, two more logistic models were regressed for online news use frequency and time spending on the nine appreciation variables, controlling for the above three dimensions of attachment. The results are shown in Table 5. As can be seen, in both models, none of the appreciated attributes was a significant predictor and neither was online news fandom. The significant effects of perceived importance of news in Internet use and the view on the Internet as the best news source suggest two alternative explanations: (1) the appreciation of online news features affects its use level indirectly through directly determining its attachment levels; or (2) only more or less cognitive attachment to online news (not totally affective attachment like fandom) influences its use levels. Which of these holds across time and space and why it happens are subjects for future enquiries.

 

Table 5: Logistic regression analyses for the effect of the appreciation of online news attributes on online news use, controlling for attachment levels (regression coefficients).
(a) 1 = Frequent use (“every day” or “several times a week”) and 0 = Infrequent use (“several times a month” or “less often”);
(b) 1 = Above average and 0 = Below average;
(c) Likelihood Ratio Chi–squared Tests with 11 degrees of freedom;
* ρ ≤ .05; ** ρ ≤ .01; *** ρ ≤ .001
Source: Australian News Usage Survey — 2004.
 Use frequency
(n = 196)a
Time spending in most recent session
(n = 196)b
No cost0.270.11
More news choices-0.380.47
Multitasking0.39-0.41
In–depth/background information-0.16-0.52
24/7 updates0.430.61
Customised news-0.24-0.03
“Have my say” to the news media-0.15-0.17
Discussing news with peers-0.520.39
Different viewpoints0.010.00
Being an online news fan0.60-0.11
News important or essential in Internet use1.63**1.69***
Internet as best medium to serve news needs2.50*-1.06*
   
LR chi2(11)c65.36***36.65***

 

It must be noted, however, that while perceived importance was a significant positive contributor in both cases (coefficient = 1.63 with ρ = .002 in the case of use frequency and coefficient = 1.69 with ρ = .001 in the case of time spending), the best–medium perception was only positive for use frequency (coefficient = 2.50, ρ = .03) and is significantly negative in the case of time spending (coefficient = -1.06, ρ = .03). While this seems contradictory to common sense, it might happen because those who see the Internet as the best news medium are compelled to use it much more frequently and thus spend less time in each specific session than those who do not. But, again, this tentative explanation needs to be further explored before a conclusion can be made.

 

++++++++++

Conclusion: Online news attributes and its future diffusion

Based on the suggested theoretical model of the online news adoption/use process, the empirical data in this paper provide in–depth information to conclude that taken as a whole, the much–touted features of online news have a strong effect on its use and attachment. First, although online news packages are often delivered with an integration of several features, there is convincing evidence that all the online news features explored in this study (except customisation) are not used passively or unintentionally as a by–product of using another (Table 2). Second, the data reveal that most of the examined online news features are crucial contributors to the overall online news experience in one way or another. In particular:

The general conclusion is that although online news has many unique attributes such as customisation and interactive participation opportunities, it is still the same established attributes such as immediacy, content richness and use convenience that are likely to be the major drivers of its development in the future. It must be noted, however, that the appreciation of multimedia — a unique attribute of online news — was not included in the survey.

Beyond the above, this research has proposed a theoretical model describing the specific processes in which online news features are encountered, accepted, adopted, used and integrated into daily life by individual users. If the model holds over time and across space, it will not only make a fairly important theoretical contribution to diffusion theory and the uses and gratifications research tradition but also will be a good basis for online journalism practitioners to design and develop strongly users–oriented news sites and online news businesses. It is noteworthy, however, that the findings in this study can confirm the overall principle and major assumptions of this model but not all of the specific processes depicted in the model. For instance, what is the relative importance of outside sources, self–inferences and the uses of other features in the adoption/use of an online news feature? How is use intention reflected in reality? Or how can we measure the influence of positive, negative and neutral values on online news adoption/use? How do these support or challenge well–established and closely related theories in the mass communication literature such as media dependency? These are questions that deserve further attention.

Future research might also need to further examine the effect of online news attributes within the socio–psychological context of its users. How do, for example, an individual’s general news orientation and behaviours (e.g., news needs, attachment to the news, attitudes to the importance of news, news use habits) affect each of the stages in the proposed model of online news adoption/use? How does experience with the Internet (e.g., Internet accessibility, Internet habits, Internet skills, bandwidth) improve the way online news is encountered, appreciated, adopted, used and integrated into daily life? Or even more fundamentally, to what extent do social locators (e.g., age, sex, education, income, occupations etc.) determine the process? These — and other variables (e.g., personality) — are important factors that are likely to have an effect on every step in the proposed model but, for practical reasons, were not explored in this paper. Studying them will also help researchers gain a more thorough understanding of whether, and to what extent, this process is changed from one cultural context to another. Are the above Australian findings applicable to, for instance, a British, an American or an Asian context? Why or why not? End of article

 

About the author

An Nguyen is a lecturer in journalism studies at the University of Stirling in Scotland. He has published in the areas of online journalism practices, online news audiences, science journalism, and professionalism in journalism education in Australia, the U.K. and the U.S.

 

Notes

1. See Hall (2003) for an economic perspective and Schiffman, et al. (2001) for a consumer–behaviour one.

2. In reality, the nexus between the adoption decision and the actual use level is applied to any innovation, including hardware innovation. As Steinfield, et al. (1989, p. 61) argued in the case of the computer: “Despite continually optimistic forecasts of adoption and evidence of steady computer sales, data are sparse concerning what actually happens after a family acquires a personal computer. Does it entertain in the living room, facilitate work in the study room, or gather dust in the closet?” The point here is that online news, as a free and ubiquitously available product online, has much more chance than a computer to be adopted without being substantially used.

3. Palmgreen and Rayburn, 1985, p. 62.

4. Stanford, 1983, p. 248.

5. Because of the blur line between adoption and use, from this point onward, I tend to use “adoption/use” rather than “adoption and use”, except when there is a need to separate the two terms.

 

References

R. Abdulla, B. Garrison, M. Salwen, P. Driscoll, and D. Casey, 2002. “The credibility of newspapers, television news and online news,” paper presented at the annual convention of the Association from Education in Journalism and Mass Communication. Miami Beach (9 August).

I. Ajzen, 1991. “The theory of planned behaviour,” Organizational Behavior and Human Decision Processes, volume 50, number 2, pp. 179–211.http://dx.doi.org/10.1016/0749-5978(91)90020-T

J. Chan and L. Leung, 2005. “Lifestyles, reliance on traditional news media and online news adoption,” New Media and Society, volume 7, number 3, pp. 357–382.http://dx.doi.org/10.1177/1461444805052281

M. Fishbein, 1968. “An investigation of relationships between beliefs about an object and the attitude towards that object,” Human Relationships, number 16, pp 233–240.

M. Fishbein, 1967. “Attitude and the prediction of behavior,” In: M. Fishbein (editor). Readings in attitude theory and measurement. New York: Wiley.

M. Fishbein and I. Ajzen, 1975. Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, Mass.: Addison–Wesley.

J. Galloway and F. Meek, 1981. “Audience uses and gratifications: An expectancy model,” Communication Research, volume 8, number 4, pp. 435–449.

B. Gunter, 2003. News and the Net. Mahwah, N.J.: L. Erlbaum.

B. Hall, 2003. “Innovation and diffusion,” at http://elsa.berkeley.edu/~bhhall/papers/Diffusion_Ch18_BHHfinal.pdf, accessed 30 March 2004.

J. McLeod, C. Bybee, and J. Durall, 1982. “Evaluating media performance by gratifications sought and received,” Journalism Quarterly, volume 59, number 1, pp. 3–12, 59.http://dx.doi.org/10.1177/107769908205900101

D. McQuail, 1997. Audience analysis. Sydney: Sage.

A. Nguyen, 2003. “The current status and potential development of online news consumption: A structural approach,” First Monday, volume 8, number 9, at http://journals.uic.edu/fm/article/view/1075/995, accessed 4 September 2003.

A. Nguyen, L. Ferrier, M. Western, and S. McKay, 2005. “Online news in Australia: patterns of use and gratification,” Australian Studies in Journalism volume 15, pp. 5–34.

Y. Nozato, 2002. “Credibility of online newspapers,” paper presented at the annual convention of the Association from Education in Journalism and Mass Communication, Miami Beach (9 August), at http://www.inma.org/subscribers/papers/2002-nozato.pdf, accessed 23 March 2004.

P. Palmgreen, 1983. “The uses and gratifications approach: A theoretical perspective,” Media Panel Report, number 30. Lunds Universitet & Högskolan I Vxäjö.

P. Palmgreen and J. Rayburn, 1985. “An expectancy–value approach to media gratifications,” In: K. Rosengren and L. Wenner (editors). Media gratifications research: Current perspectives. Beverly Hills, Calif.: Sage.

P. Palmgreen, L. Wenner, and J. Rayburn, 1980. “Relations between gratifications sought and obtained: A study of television news,” Communication Research, volume 7, number 2, pp. 161–192.http://dx.doi.org/10.1177/009365028000700202

J. Rayburn and P. Palmgreen, 1984. “Merging uses and gratifications and expectancy–value theory,” Communication Research, volume 11, number 4, pp. 537–562.http://dx.doi.org/10.1177/009365084011004005

E. Rogers, 2003. Diffusion of Innovations. Fifth edition. New York: Free Press.

M. Salwen, B. Garrison, and P. Driscoll, 2005. Online news and the public. Mahwah, N.J.: L. Erlbaum.

L. Schiffman, D. Bednall, E. Cowley, A. O’Cass, J. Watson, and L. Kanuk, 2001. Consumer behaviour. Second edition. Sydney: Prentice Hall.

S. Stanford, 1983. “Comment on Palmgreen and Rayburn, ‘gratifications sought and media exposure’,” Communication Research, volume 10, number 2, pp. 247–251.http://dx.doi.org/10.1177/009365083010002005

C. Steinfield, W. Dutton, and P. Kovaric, 1989. “A framework and agenda for research on computing in the home,” In: J. Salvaggio and J. Bryant (editors). Media use in the information age: Emerging patterns of adoption and consumer use. London: L. Erlbaum, pp. 61–86.

Washingtonpost.com, 2005. “The online news and information user: Understanding media choices,” at http://www.washingtonpost.com/wp-srv/marketing/presentation/nielsen/content/frame.htm, accessed 23 July 2005.

L. Weibull, 1985. “Structural factors in gratifications research,” In: K. Rosengren and L. Wenner (editors). Media gratifications research: Current perspectives. Beverly Hills, Calif.: Sage.

T. Weir, 1999. “Innovators or news hounds: A study of early adopters of the electronic newspaper,” Newspaper Research Journal, volume 20, number 4, pp. 62–81.

H. Wu and A. Bechtel, 2002. “Website use and news topic and type,” Journalism and Mass Communication Quarterly, volume 79, number 1, pp. 73–86.http://dx.doi.org/10.1177/107769900207900106

 


Editorial history

Paper received 18 February 2008; accepted 4 March 2008.


Copyright © 2008, First Monday.

Copyright © 2008, An Nguyen.

The contribution of online news attributes to its diffusion: An empirical exploration based on a proposed theoretical model for the micro–process of online news adoption/use
by An Nguyen
First Monday, Volume 13, Number 4 - 7 April 2008
https://firstmonday.org/ojs/index.php/fm/article/download/2127/1952