Choosing and reading online news: how available choice affects cognitive processing.

Giving people the ability to choose from a wide variety of content when they want to see it is a hallmark of today's interactive media landscape. News/information portals, blogs, video-on-demand, and file-sharing galleries all give computer users the freedom to choose content from vast arrays of options, then acquire and view that content with the click of a mouse. To date, researchers have not thoroughly examined the mental processes that occur when people choose and receive information from Web sites that offer varying amounts of options.

The ritual of scanning, choosing, and receiving hyperlinked content can be repeated multiple times during a person's visit to a Web site. This ritual and the underlying mental processes involved fundamentally shape the experience of receiving news online. Therefore, it is important to understand how various features involved in presenting online news affect cognitive processes involved in selecting, reading, and remembering an online news story. This study serves as an initial investigation of how one feature, the number of hyperlinked stories presented, affects cognitive processing of a selected story. Results obtained can advance theoretical understanding of information processing of mediated content in an unexplored area of news consumption. Findings may also provide news producers with some insight into Web site design that will maximize the ability of their audience to be informed by attending to and remembering content of online news stories.

It is proposed here that the mental work a person does in scanning and choosing a hyperlinked news story has consequences for cognitive and emotional processing of information contained in the story. Thus, any feature of a news Web site that could affect mental effort invested in selecting stories, such as the number of hyperlinks from which an individual chooses stories to read, could significantly affect how the content of a story is attended to and remembered.

There is precedence for theorizing that the number of hyperlinked news stories on a Web site could influence responses to received information. Scholars studying decision-making have demonstrated that the number of options available for an individual to choose from affects responses to their choice (Iyengar & Lepper, 2000; Wise & Pepple, in press). This finding has been obtained in the context of choosing among different brands of a product as well as selecting pictures to view from varying arrays of photographs. Receiving news from online sources provides another interesting decision-making context in which individuals are presented with numerous options. Google News, for example, links Web surfers with every available story on a particular event. For prominent events, the number of stories from which a person can choose may reach into the hundreds. Scholars studying the effects of available options on responses to an individual's choice have not yet examined it in the context of choosing stories from an online news interface.

This study is designed to fill a gap in media processes and effects research by examining cognitive processing in the unexplored context of choosing, reading, and remembering online news. Such research is necessary because the mental processes used in selecting and processing online news could be somewhat different from processes engaged by previously studied decision-making tasks. This research may also advance theoretical understanding of the relationship between choosing media content and cognitively processing the selected content. Media processes and effects scholars have not thoroughly explored how media features presented to an individual in the act of selecting content influences cognitive processes engaged during exposure to the received content.

Studying the relationship between choosing and processing selected online news stories requires analysis of two mental tasks: (1) scanning pictures and headlines in order to choose a story, and (2) reading the text of the story once it has been chosen. These tasks are important to consider because the precise demands placed on cognitive resources are likely to differ between them.

A theoretical model that has received widespread support for explaining how individuals allocate cognitive resources to processing mediated messages is Lang's Limited Capacity Model of Motivated Media Message Processing (A. Lang, 2006). First described by A. Lang (2000), the fundamental assumption of the model is that processing a mediated message involves a continuous interaction between the human information processing system and features of the mediated message. Processing media content involves allocating limited cognitive resources to the subprocesses of encoding, storage, and retrieval. The degree to which cognitive resources are allocated to each of these subprocesses varies based on both individual goals and message features. A. Lang (2006) theorized that individuals increase cognitive resources allocated to processing media content portraying motivationally important information (i.e., danger, food, sex, etc.). This notion seems particularly relevant to cognitive processing of news as much of the content of news stories deals with potential danger and other negative events. For this study, it means taking a closer look at the nature of encoding, storage, and retrieval of information contained in an online news story reporting unpleasant events.

Encoding, storage, and retrieval are not performed in serial order. Rather, as an individual encodes new information from a mediated message, information previously stored in long-term memory is retrieved as part of the process of storing the new information in memory (A. Lang, 2006). Retrieving information from long-term memory and holding it in short-term working memory along with any encoded new information from a message is a critical step in effectively storing information from a news story in long-term memory. At a minimum, an individual reading a news story must retrieve knowledge of language, stored in long-term memory, in order to make any sense out of information from the story that is being encoded into working memory. Thus, processing a news story involves the simultaneous allocation of cognitive resources to encoding, retrieval, and storage. It is critical to note that cognitive resources are not allocated equally among all three tasks. An individual's goals as well as features of a news story can elicit increased resource allocation to one of the three tasks, leaving fewer cognitive resources to be allocated to the others. For example, in a study on cognitive processing of radio advertisements it was found that as more cognitive resources were shifted to retrieving information out of long-term memory to aid in storing a message, fewer resources were allocated to encoding details of the message (Boils, in press).

The clear implication of the limited capacity model for how individuals process online news is that both individual goals and message features will affect the allocation of cognitive resources across the mental tasks involved in choosing, processing, and remembering news stories. Both individual goals and message features vary across the previously mentioned stages of selecting and reading online news. Therefore, cognitive resources are likely allocated among encoding, storage, and retrieval to varying degrees depending on the precise processing demands at each stage. This makes it important to consider the demands likely present at each stage of an online news-viewing episode: choosing a story from an array of pictures and headlines and then reading the text of the chosen story.

When a user visits a news Web site, they typically encounter a series of short headlines meant to grab attention and lead into the story. Sometimes, a related photograph accompanies these headlines. If someone has gone online simply to browse the news, without a particular informational need, the goal at this stage may simply be to choose a story that looks interesting (Tewksbury, 2003). The combination of Web site features and user goals at the story selection stage seems likely to elicit the allocation of processing resources to encoding features of the photographs and headlines. Previous research has demonstrated that negative, compelling visual images, such as those often found in news content, automatically increase resources allocated to encoding both photographs as well as video (P. J. Lang, Greenwald, Bradley, & Harem, 1993; Lang, Newhagen, & Reeves, 1996). It is important to note that such negative, compelling images are often perceived as such because the image represents a potential threat and therefore would seem to fit the description of motivationally relevant content. But what about storage? It seems likely there is little reason to store information in long-term memory at the point of selecting a story to read from an array of pictures and headlines. An individual simply needs to encode the available headlines and accompanying pictures in order to click through the link and read the full story. Storing detailed information contained in the pictures and headlines in long-term memory is not necessary to decide which story to select to read and could actually interfere with processing the oncoming story.

Once a person has chosen a story by clicking through a hyperlink, the mental task changes dramatically. After all, the stimulus package has changed from an array of unrelated topics and accompanying pictures to a string of related sentences. The task at this stage is to read the selected story, which should increase cognitive resources allocated to retrieving information from long-term memory in order to store information from the story in memory. Most of the mental work involved in reading text is focused on forming an abstract representation of the text in working memory (Rayner, Liversedge, White, & Vergilino-Perez, 2003). Linguistic and conceptual processing based on an individual's stored knowledge is the key to forming an abstract representation of text in working memory (Carpenter & Daneman, 1981). Thus, most of the work involved in reading text would seem to involve accessing linguistic and conceptual knowledge stored in long-term memory. Few resources would seem to be necessary to encode text appearing on screen into working memory. Indeed, it has been found that within 50 milliseconds, fluent readers can encode concrete visual features of the text that the eyes are fixated on (Rayner et al., 2003). Under A. Lang's limited capacity model described earlier, the mental task of reading the text of an online news story would appear to require cognitive resources to be shifted from encoding concrete features of text to retrieving from long-term memory previously stored knowledge needed for conceptual and linguistic processing of the text. This form of cognitive processing would be most likely to facilitate memory for story details.

In summary, the mental task of processing pictures and headlines in order to choose an online news story should elicit the allocation of processing resources to encoding information contained in the pictures and headlines with low demand for resources allocated to retrieval and storage. In contrast, reading the text of a story, once it has been chosen, should significantly increase the amount of cognitive resources allocated to retrieval.

A feature of news Web sites that clearly distinguishes the amount of information to be encoded at the stage of choosing a story to read is the number of hyperlinked stories presented on a particular Web site's main page. How might the number of hyperlinked stories an individual can choose from influence subsequent processing of the text of a selected story? A Web page with more hyperlinked stories ought to require more resources allocated to encoding the presented information in order to choose a story to read than a Web page with fewer hyperlinked stories. However, it seems unlikely that the demand on cognitive resources presented by the number of hyperlinked stories will significantly tax the human information processing system to the point of constraining resources available to process the text of a selected story. Encoding visuals in a mediated message has been found to be a relatively effortless, fairly automatic mental task (A. Lang, Potter, & Bolls, 1999). Further, the short headlines connected to photographs are also unlikely to be all that demanding to process. Rather than overloading cognitive resources, perhaps increasing the number of stories covering unpleasant events presented on a news Web site increases the motivational importance of information received leading individuals to increase resources allocated to reading a selected story.

Why might offering more hyperlinked stories on an online news Web site increase the motivational importance of a story an individual chooses to read? The human cognitive processing system attends preferentially to cues in the information environment that signal potential threats. Unpleasant stimuli in an individual's environment can automatically increase resources allocated to encoding information by activating what has been termed the aversive motivational system (A. Lang, 2006). This idea has been applied specifically to cognitive processing of news. Shoemaker (1996) argued that humans have a biological imperative to survey the environment for threats and therefore, may actually be somewhat hardwired to process unpleasant news. Features involved in the presentation of online news could serve as cues that signal the degree of threat in an individual's information environment. Such features should increase the motivational importance of received information leading individuals to allocate more effort to cognitively processing selected stories.

An example of a specific feature involved in the presentation of online news that has been found to affect cognitive processing is threatening news photographs. The presence of a threatening photo accompanying an online news headline has been found to increase effort put into processing the related story (Knobloch, Hastall, Zillmann, & Callison, 2003). A possible explanation of this finding is that unpleasant photographs on a news Web site increase the motivational relevance of presented information. Such an explanation is consistent with Lang's limited capacity model (A. Lang, 2006).

Increasing the number of unpleasant photos contained on a news Web site by increasing the number of hyperlinked stories available for an individual to click through to read could magnify the effect found by Knobloch and colleagues. This would theoretically occur by having hyperlinked unpleasant photo/headline combinations serve as a cue signaling the motivational importance of information presented on a news Web site. More such cues should result in an even higher degree of motivational importance leading to more effortful processing of selected stories. Manipulating the number of unpleasant, picture/headline hyperlinks presented to participants is one way to initially test this possibility.

One way of measuring mental effort invested in processing mediated information is to obtain individuals' heart rate during media exposure (A. Lang, 1994). Several studies have demonstrated the utility of heart rate as a measure of attention to media content (see Ravaja, 2004, for a review). As previously mentioned, attending to media content involves allocating cognitive resources to encoding, storage, and retrieval. Increasing cognitive resources allocated to encoding media content has been found to result in cardiac deceleration (A. Lang, 1990). However, increasing resources allocated to retrieving information out of long-term memory in order to store encoded media content has been found to result in cardiac acceleration (Bolls, 2002). Reading the text of a story chosen from an online news site primarily involves just such mental activity, so an increase in the cognitive resources allocated to reading the text should lead to faster heart rate. This leads to the first hypothesis:

[H.sub.1]: Reading stories chosen from an extensive array of hyperlinks will lead to greater cardiac acceleration than reading stories chosen from a limited array of hyperlinks.

The increased mental effort put into reading stories selected from an extensive array of hyperlinks ought to also be reflected in memory for story details. This leads to a second hypothesis:

[H.sub.2]: Details in stories chosen from an extensive array of hyperlinks will be more accurately recognized than details in stories chosen from a limited array of hyperlinks.

To test these hypotheses, an experiment was conducted in which participants selected individual news stories from a Web site containing both limited and extensive arrays of options. Participants' heart rate was recorded while they read each story. Later, their recognition memory for details from each story was tested.

Method

Participants

Thirty-four undergraduates (14 Male, 20 Female) participated in this study for credit in an advertising course at a large Midwestern university.

Design

This study was a 2 (Available Choice) x 3 (Repetition) repeated-measures design. Available Choice referred to the number of news stories that participants had to choose from and had two levels, limited (5) and extensive (15). These operational definitions are consistent with previous research which used 5 options as a limited-choice condition and described extensive-choice conditions as consisting of a "reasonably large but not ecologically unusual" number of options (lyengar & Lepper, 2000, p. 996). It was found that the "headline news" section of prominent news Web sites typically contains 10-12 headline links, and the researchers wanted to use a multiple of 5 for stimulus preparation and data analysis purposes, so 15 options were chosen for the extensive array. Participants chose and read three stories from both the limited and the extensive array of options, thus comprising the Repetition factor.

Stimuli

A news Web site containing two pages of hyperlinks was used. One page contained 5 (limited) hyperlinks; the other page contained 15 (extensive) hyperlinks. Each hyperlink consisted of a picture and a headline. When participants clicked on a hyperlink, they were taken to a news story that corresponded to the hyperlink.

The pictures came from the International Affective Picture System (lAPS) (Center for the Study of Emotion and Attention, 2001). The lAPS is a CD-ROM containing hundreds of digitized photographs that have been rated extensively in previous research, producing reliable ratings along the dimensions of valence and arousal (e.g., P. J. Lang et al., 1993). Since arousal and valence have been shown to significantly affect the allocation of cognitive resources to processing messages (A. Lang, Dhillon, & Dong, 1995), use of the lAPS photos provided the ability to control these emotional factors. All of the pictures used in this study were mildly unpleasant and moderately arousing. The mean valence and arousal ratings of the picture sample were 3.32 (SD = 1.62) and 4.62 (SD = 2.08), respectively, ranging from 1 (very unpleasant/boring) to 9 (very pleasant/arousing). Pairwise comparisons among our picture sample yielded no significant differences across both valence and arousal.

Mildly unpleasant pictures were used because previous research has shown that they elicit more attention than pleasant pictures (Bradley, Greenwald, Petry, & Lang, 1992). Moderately arousing pictures were used in order to generate interest in the hyperlinks without getting participants so aroused that they would have difficulty attending to the corresponding story.

Once the pictures had been chosen, a student enrolled in a news reporting class wrote a headline corresponding to each picture. These headlines were pretested in a small class where students (N = 15) rated each headline with the same instrument (The Self-Assessment Manikin [SAM]; Bradley & Lang, 1994) traditionally used to rate the valence and arousal of lAPS pictures. Students also rated how interested they would be in reading a story based on each headline if they saw it on a news Web site, using a 7-point scale anchored by "not interested" and "very interested." From these ratings, 20 headlines were chosen that clustered around the midpoint of both arousal (M = 4.49, SD = .53) and interest (M = 5.44, SD = .62). Pairwise comparisons yielded no significant differences for all measures among the chosen headlines. These headlines were randomly assigned to the two pages of hyperlinks, with 15 hyperlinks on one page and 5 hyperlinks on the other. After the headlines had been chosen, the same student wrote a short (110-130 words) fictitious news story for each of the 20 headlines. The limited and extensive arrays of hyperlinks are shown in Figure 1.

[FIGURE 1 OMITTED]

Physiological Recording

Physiological signals were measured, amplified, and recorded using Coulbourn modules linked to a PC. The VPM software program (Cook, Atkinson, & Lang, 1987) coordinated the sampling and storage of physiology data.

Heart Rate

Two 8mm Ag/AgCl (silver/silver-chloride) electrodes were placed on each participant's forearms, a few inches above the wrist. The signal was amplified and filtered by a Coulbourn bioamplifier equipped with high and low pass filtering. A Schmitt trigger interrupted the PC every time it detected the R-spike of the cardiac wave. Data were initially collected as interbeat intervals, or milliseconds between consecutive R-spikes in the QRS complex of the cardiac cycle. Data were then cleaned for movement artifact and transformed to an average beat per minute value for each second of data collection.

Skin Conductance

Skin conductance was measured in order to control for the potential confound of varying levels of arousal evoked by the news stories. Two 8mm Ag/AgCl (silver/silver-chloride) electrodes were placed on the palm of each participant's nondominant hand. The electrodes were attached to a Coulbourn $71-23 isolated skin conductance coupler. This signal was sampled 20 times per second and converted to conductance values in microSeimens (uS).

Recognition

Recognition of details contained in each story was measured through a four-alternative multiple choice test. Four questions were asked about each story participants read. Each question was based on a factual element within the story. The questions from each story were presented in the same block; the first question of each block identified which story the question referred to. MediaLab software randomized presentation of each block of questions. When a question appeared on the screen, participants used a mouse to point and click on what they believed to be the correct answer.

Procedure

Participants entered the laboratory, sat down at a computer terminal, and provided written informed consent. An experimenter prepared each participant's skin for data collection and attached sensors to the participant, The experimenter told each participant that they would be reading stories from a news Web site. No criteria were given to guide participants' choice of stories, they were simply told to click on any headline that they'd like to read more about. The experiment began once the participant expressed understanding of the procedure.

Stimulus presentation began with instructions, followed by a 10-second baseline period in which participants saw a black screen. After this baseline period, the first news Web site appeared containing either 5 or 15 options. The MediaLab software program (Jarvis, 2004) controlled stimulus presentation and randomized Web site presentation order. Participants used a mouse to point and click on the headline of their choosing, taking as much or as little time as they wanted to choose each headline. After pointing and clicking on a headline of their choice, each participant read the corresponding story. Again, participants could take as much or as little time as they needed to read each story. Upon reading the first story, participants clicked a "Continue" button that returned them to the headline array. Participants then repeated this choosing/reading task two more times, until they had chosen and read three stories from the first Web page. The entire process was repeated for the second Web page.

Next, participants took part in an unrelated experiment that served as a distraction task for this study. The distraction task lasted approximately 30 minutes and required each participant to watch 10 television commercials, completing a short questionnaire after each commercial. During this task, the experimenter created a recognition test corresponding to the stories each participant had chosen to read from each Web site. Upon completing the television advertising experiment, participants completed the recognition test over the stories they selected to read. At the end of the recognition test, the experimenter removed electrodes from participants, then debriefed, thanked, and dismissed them. The entire experiment lasted approximately 1 hour.

Data Reduction

Response curves were created for both heart rate and skin conductance by computing change scores across the period when participants read each story. Because of individual differences in reading speed, the time spent reading each article varied across participants. In order to use time as a repeated measure in these analyses, it was necessary to transform these reading periods into uniform lengths. This was done by dividing each reading period into thirds (i.e., beginning, middle, end), then computing the average heart rate for each third. For example, if a participant took 30 seconds to read a particular story, that period was divided into three equal segments of 10 seconds each. The average heart rate/skin conductance during each of those 10-second segments was then computed. Finally, each 10-second average was subtracted from the value taken in the second immediately prior to the appearance of the story.

If the time spent reading a particular story was not divisible by 3, remaining seconds were divided among the first two segments. For instance, a 32-second reading period would have been divided into segments of 11, 11, and 10 seconds. This technique yielded an equal number of segments that allowed repeated measures analyses of tonic heart rate and skin conductance level (SCL) by yielding 4 data points for each segment, counting the onset of each selected story. (1)

Analyses

Both heart rate and skin conductance data were analyzed with a 2 (Choice) x 2 (Repetition) x 4 (Time) repeated-measures analysis of variance (ANOVA). Univariate analyses that violated the assumption of sphericity were adjusted with the Huynh-Feldt degrees of freedom correction. There are only 2 (and not 3) levels to the repetition factor because the first story that subjects read from each Web site was removed from physiological data analyses. Because stimulus novelty is an essential component of the orienting response (A. Lang, 1994), which is accompanied by decreased heart rate and increased skin conductance, it was feared that novelty associated with the first time of performing a task could potentially skew the physiological data.

Each question on the recognition test was coded as either a "1" if it was correctly answered or "0" for an incorrect answer. These numbers were then summed across each condition and divided by the total number of questions (12). This yielded separate recognition accuracy scores for each condition that were compared with a dependent-samples t test.

To demonstrate that hypothesized effects could not be caused by differences in story content, the skin conductance data recorded while participants read each story was analyzed. If the stories that people chose and read from one array were more inherently arousing than the stories they chose and read from the other array, there would have been greater skin conductance throughout the reading period. This would confound the results since arousal is associated with accelerated heart rate and better recognition memory (Bradley et al., 1992). The Choice x Time interaction was not significant, F(3, 81) = .07, ns, indicating no differences in participants' arousal while reading stories from each condition.

The amount of time it took for participants to read the stories in each condition was analyzed to make sure that stories in one condition weren't inherently more difficult to read than stories in another condition. Such a finding would confound the recognition data since people should remember fewer details from more difficult stories. The main effect of Choice on time spent reading was not significant, F(1, 27) = .20, ns; Limited: M = 26.51, SD = 1.25; Extensive: M = 26.84, SD = 1.38, indicating that the stories in each condition were of equal difficulty.

Results

Hypothesis 1 predicted that participants' heart rate would be faster while reading news stories selected from an extensive array of options versus stories selected from a limited array. This pattern of results would reflect greater mental effort put into reading a story selected from an extensive array of hyperlinks and pictures. Analysis of the heart rate data revealed a significant Choice x Time interaction, F(3, 81) = 3.43, p <.03, partial-[[eta].sup.2] = .11. As Figure 2 shows, participants experienced greater heart rate acceleration while reading news stories selected from the extensive array of options in comparison to heart rate experienced while reading stories chosen from the limited array of options. This result supports Hypothesis 1, suggesting that individuals put more mental effort into reading the text of an online news story when it is selected from a higher number of available stories.

[FIGURE 2 OMITTED]

According to Hypothesis 2, recognition will be more accurate for specific facts in news stories chosen from the extensive array of options. There was a significant main effect for Choice on the recognition data, t(27) = -4.06, p <.01. Participants' recognition was more accurate for stories chosen from the extensive array of options (M = .77, SD = .03), compared to recognition of facts in stories chosen from the limited array of options (M = .61, SD = .03). This result supports Hypothesis 2, indicating that increased mental effort put into reading stories does indeed result in more accurate recognition of story details.

Discussion

This experiment explored how the number of available hyperlinked stories affects cognitive processing of online news. Participants chose an online news story from a Web page containing either 5 or 15 headline/photo combinations, then read the story they had chosen. It was predicted that people would invest more cognitive effort into reading the stories that they had chosen from the Web site with more hyperlinked photo/headline options. This prediction was derived from Lang's Limited Capacity Model of Motivated Media Message Processing (A. Lang, 2006). This model proposes that features involved in the presentation of media content can serve as cues, increasing the motivational importance of received information. It was theorized that increasing the number of unpleasant headline/photo links could increase the motivational relevance of presented stories leading participants to put more mental effort into reading selected stories. If people exerted more cognitive effort processing the stories chosen from the larger array, it was expected that they would also have better recognition of details from these stories.

The results of this experiment support the hypotheses. Participants showed greater heart rate acceleration, indicative of cognitive effort, while reading stories chosen from the larger array. They also had more accurate recognition for details of the stories chosen from the larger array.

These findings have both theoretical and practical implications for understanding how people process online news. It does appear that unpleasant photo/headline combinations contained on news Web sites act as cues capable of increasing the motivational relevance of presented stories. This increase in motivation could have come from an increase in perceived threat due to the presence of more unpleasant stories. Knobloch and colleagues (2003) found that the presence of a threat-related photograph increased the effort people invested in processing online stories. These results suggest that increasing the number of unpleasant, potentially threatening photographs linked to different stories presented to an individual could magnify this effect. It is important to note that no manipulation of perceived threat occurred in this study, so this idea is speculative. However, given the nature of the stories used in this experiment, as well as an apparent emphasis on threat reporting in today's news environment, it seems plausible that someone's general perception of threat would vary with the number of bad news stories to which they are exposed.

Results of this experiment suggest that features of online news presentation that could impact motivational relevance of information presented are critical variables in determining how stories will be cognitively processed. A. Lang's (2006) limited capacity model specifically includes motivation, in the form of appetitive and aversive activation, as a variable influencing how people process mediated messages. Future research should more directly apply Lang's model in studying cognitive processing of online news by directly measuring how features of news Web sites affect aversive and appetitive motivational activation. The current study indicates that motivational activation should be considered across mental tasks engaged during media exposure. Here, these tasks included choosing and reading a story. This appears to be especially important in the context of interactive media where features of every mediated interface (i.e., Web page) could evoke varying patterns of aversive and appetitive motivational activation that influences cognitive processing of subsequent media content. The nature of photos, headlines, and stories likely increased aversive motivational activation. Future research should also explore features of online news that are likely to increase appetitive activation. After all, not all news is bad news.

This study serves as a reminder of the importance of considering the precise nature of mental tasks involved in cognitively processing media content. Results reported here illustrate how the exact mental processes involved in encoding a mediated message can vary. It appears that the mental processes engaged when an individual invests more effort into encoding the text of a selected online news story are similar to mental processes engaged by encoding high-imagery radio ads. Both tasks appear to result in cardiac acceleration, which in theory, is due to more resources allocated to retrieving information from long-term memory to aid encoding of the message. This is a different pattern of results than reported in studies of cognitive processing of television where greater effort invested in encoding has been found to result in heart rate deceleration (i.e., A. Lang, Bolls, Potter, & Kawahara, 1999). Cardiac deceleration is thought to be due to an increase in cognitive resources allocated to encoding concrete features of a message. Researchers need to theoretically consider how features of messages they are studying might engage these two different forms of encoding. Varying mental processes underlying encoding could substantially change observed message effects.

Combining results of this study with other recent research allows some interesting albeit preliminary conclusions to be drawn about how available choice moderates the processing of online media. If supported by further research, these conclusions have practical implications for designing news Web sites to be most informative. In a study of responses to selecting photographs for further viewing, Wise and Pepple (in press) found that people had weaker orienting responses and poorer recognition for pictures chosen from an extensive, compared to a limited, array of options. The results reported here show exactly the opposite effect when the stimulus is the text of a story. Taken together, these two studies indicate that the degree to which features of a Web page engage similar mental processes at the stages of choosing and then processing selected content may moderate the effect available choice has on memory for online content. In the Wise and Pepple study, both stages primarily engaged visual encoding of the stimulus. Thus, the mental processes at both stages were highly similar, leading to poorer memory when choosing content placed a greater demand on cognitive resources (extensive choice). In the current study, features of the Web page engaged significantly different mental processes at the choice stage compared to processing selected content. Thus, increasing the demand placed on cognitive resources in choosing content did not have a detrimental effect on memory of selected content.

The practical implication of this study for online news producers is that offering more hyperlinked photo/headline elements connected to stories on a Web page seems to lead audience members to invest more effort into processing the text of the story they select. This contradicts the assumption that offering a lot of stories on a Web page might somehow lead to information overload for a Web site visitor. The highest number of hyperlinked stories tested in this study was 15. Some online news portals offer more links than this. There could be a point where the number of stories offered on a Web page does overwhelm users leading them to disengage from information processing of presented news. It is also important to note that this study only looked at cognitive processing of online news. There could be negative effects due to offering a lot of stories on a news Web site on attitudinal variables such as attraction, involvement, or credibility, which could have just as important implications for online news consumption. Future research should examine the interaction of cognitive processing and attitudinal responses evoked by features involved in the presentation of online news.

The broader implication of this study is that producers need to learn more about how features of online news presentation impact motivational activation during news consumption. It is well established that variation in underlying appetitive and aversive motivation affects how individuals process information in their environment (Cacioppo, Gardner, & Berntson, 1999; A. Lang, 2006). Content characteristics as well as features involved in presenting stories on a news Web site could engage varying patterns of motivational activation. Understanding how different ways of producing an online news stories engage motivational activation could provide insight into how to design news Web sites that are maximally effective at encouraging individuals to cognitively process news in a way that makes stories more memorable. It appears that up to a point, increasing the number of stories a user can choose from is one way to achieve this.

The limitations of this study point to several interesting avenues for future research. Online content often combines text with multimedia information, such as text stories that give the user the option of viewing video content of a particular on-scene or studio report. This study investigated the presentation of simple photographs, headlines, and written text. This is likely the simplest form of online news content. Clearly, online news portals are increasingly becoming more complex and dynamic in combining text, pictures, audio, and video in the presentation of news. It would be interesting to explore how different formats for presenting available content a user can choose from interacts with processing of audio/visual online news content.

A second limitation is that the recognition memory test used in this experiment lacked some sensitivity. Difficulty of individual multiple-choice questions was not systematically controlled for each story. Admittedly, the difficulty of each question on the recognition test probably varied somewhat. However, following procedure established in previous research, recognition for each story was computed as the average of four questions per story (A. Lang et al., 1999). This should allow the variation due to varying level of difficulty between questions to go into the error term of the analysis. Future research could provide even more insight into recognition of details in a news story by switching from a forced choice multiple-choice test to a speeded recognition test that would allow signal detection analysis of the data. Such an analysis would provide insight into memory sensitivity as well as simple accuracy.

Ultimately, results of this experiment further demonstrate that the mental demands of negotiating online media to acquire selected content affect cognitive processing of that content when it is delivered. This could be considered as the distinction between "getting there" and "being there." In this case, getting there required people to choose from either a limited or extensive array of hyperlinks, while being there involved reading the story acquired by clicking the chosen hyperlink. When the demands of getting there changed, so did people's cognitive responses to the content they acquired. Understanding what people must go through to acquire content from interactive media is crucial to understanding how they process content once they've gotten there.

References

Bolls, P. D. (2002). I can hear you but can I see you? The use of visual cognition during exposure to high imagery radio advertisements. Communication Research, 29, 537-563.

Bolls, P. D. (in press). It's just your imagination: The effect of imagery on product versus non-product information in radio advertisements. Journal of Radio Studies.

Bradley, M. M., Greenwald, M. K., Petry, M. C., & Lang, P. J. (1992). Remembering pictures: Pleasure and arousal in memory. Journal of Experimental Psychology: Learning, Memory, and Cognition 18, 379-390.

Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The Self-Assessment Manikin and the Semantic Differential. Journal of Behavior Therapy and Experimental Psychiatry, 25, 49-59.

Cacioppo, J. T., Gardner, W. L., & Berntson, G. G. (1999). The affect system has parallel and integrative processing components: Form follows function. Journal of Personality and Social Psychology, 76, 839-855.

Carpenter, P. A., & Daneman, M. (1981). Lexical retrieval and error recovery in reading: A model based on eye fixations. Journal of Verbal Learning and Verbal Behavior, 20, 137-160.

Center for the Study of Emotion and Attention [CSEA-NIMH] (2001). The international affective picture system: Digitized photographs. Gainesville, FL: The Center for Research in Psychophysiology, University of Florida.

Cook, E. W. Ill., Atkinson, L., & Lang, K. G. (1987). Stimulus control and data acquisition for IBM PCs and compatibles. Psychophysiology, 24, 726-727.

Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Can one desire too much of a good thing? Journal of Personality and Social Psychology, 79, 995-1006.

Jarvis, W. B. G. (2004). MediaLab [Computer software]. New York: Empirisoft.

Knobloch, S., Hastall, M., Zillmann, D., & Callison, C. (2003). Imagery effects on the selective reading of Internet newsmagazines. Communication Research, 30, 3-29.

Lang, A. (1990). Involuntary attention and physiological arousal evoked by structural features and mild emotion in TV commercials. Communication Research, 17, 275-299.

Lang, A. (1994). What can the heart tell us about thinking? In A. Lang (Ed.), Measuring psychological responses to media (pp. 99-113). Hillsdale, N J: Lawrence Erlbaum Associates, Inc.

Lang, A. (2000). The limited capacity model of mediated message processing. Journal of Communication, 50, 46-70.

Lang, A. (2006). Using the limited capacity model of motivated mediated message processing to design effective cancer communication. Journal of Communication, 56, S57-S80.

Lang, A., Bolls, P. D., Potter, R. F., & Kawahara, K. (1999). The effects of production pacing and arousing content on the information processing of television messages. Journal of Broadcasting & Electronic Media, 43, 451-475.

Lang, A., Dhillon, K., & Dong, Q. (1995). The effects of emotional arousal and valence on television viewers' cognitive capacity and memory. Journal of Broadcasting & Electronic Media, 39, 313-327.

Lang, A., Newhagen, J., & Reeves, B. (1996). Negative video as structure: Emotion, attention, capacity, and memory. Journal of Broadcasting & Electronic Media, 40, 460-477.

Lang, A., Potter, R. F., & Bolls, P. D. (1999). Something for nothing: Is visual encoding automatic? Media Psychology, 1, 145-163.

Lang, P. J., Greenwald, M., Bradley, M. M., & Hamm, A. O. (1993). Looking at pictures: Evaluative, facial, visceral, and behavioral responses. Psychophysiology, 30, 261-273.

Ravaja, N. (2004). Contributions of psychophysiology to media research: Review and recommendations. Media Psychology, 6, 193-235.

Rayner, K., Liversedge, S. P., White, S. J., & Vergilino-Perez, D. (2003). Reading disappearing text: Cognitive control of eye movements. Psychological Science, 14, 385-388.

Shoemaker, P. J. (1996). Hardwired for news: Using biological and cultural evolution to explain the surveillance function. Journal of Communication, 46, 32-47.

Tewksbury, D. (2003). What do Americans really want to know? Tracking the behavior of news readers on the Internet. Journal of Communication, 53, 694-710.

Wise, K., & Pepple, K. (in press). The effect of complexity of choice on the cognitive processing of pictures. Computers in Human Behavior.

Note

(1) This technique would not be advisable for analyzing phasic responses because it dilutes the second-by-second fluctuations that distinguish phasic responses (i.e., orienting, startle, defense). However, averages across time periods are acceptable in analysis of tonic responses.

Kevin Wise (Ph.D., Stanford) is an Assistant Professor of Strategic Communication and Co-Director of the PRIME Lab (Psychological Research on Information and Media Effects) at the Missouri School of Journalism. His research interests include the cognitive and emotional processing of interactive media.

Paul D. Bolls (Ph.D., Indiana) is an Associate Professor of Strategic Communication and Co-Director of the PRIME Lab at the Missouri School of Journalism. His research agenda is focused on cognitive and emotional processing of media content.

Samantha R. Schaefer is a junior Strategic Communication major at the Missouri School of Journalism.