TV news has a problem — older people, especially, mostly get their news from television; however, younger people are getting their news, more and more, online. This development is proceeding rapidly. The Pew Research Center found that between 2016 and 2017 the share of Americans who often get their news from TV is down from 57% to 50 % over the course of a bit more than one year: “At the same time, the portion of Americans often getting news online . . . grew from 38% in early 2016 to 43% in the middle of 2017” (Gottfried & Shearer, 2017). In Germany, this development is slower but similar (Hölig & Hasebrink, 2017). Nielsen and Sambrook (2016) sum this up; “Television news as we know it . . . serves the past, not the future, and television news producers have to experiment with new formats and forms of distribution if they wish to remain relevant.” In this fight against irrelevance, globally, many broadcasters try to fascinate their viewers with giant and glittering news studio sets (cf. Pinterest, 2018). They want to make obtaining TV news from a big screen more attractive than from the small display of a Smartphone, for instance.
Since April 2014, the ARD has also broadcast its news from such a new studio. Similar to Al Jazeera, BBC News and CNN, ARD’s newscasters are positioned in front of a large media wall that shows photos, graphics, maps and sometimes videos. In this way, ARD wants to satisfy the “audiences’ requests for more visual stimuli,” as stated by the chief editor of Tagesschau, Kai Gniffke (personal communication, September 12, 2014). The question is, however, what more visual stimuli mean for viewers’ information processing. Therefore, we made the new Tagesschau design the topic of our research seminars to investigate the impact of large media walls and big background illustrations on the newscast’s ability to convey information. This sort of investigation is challenging because TV news exists not only as visual but also as multimodal stimuli (Piazza & Haarman, 2015, p. 462).
Viewers have to make meaning from a mix of spoken words, sounds, inscriptions, videos, photos and graphics, and in addition, process the hosts’ gestures, voice, hairstyle and fashion. Bucher and Schumacher (2012) call this kind of reception a “process of interpretation in which manifold symbol systems have to be combined to a coherent sense” (p. 10). Therefore, it is difficult to analyze the viewers’ processing of television news. The communication scholar, Ruth Teer-Tomaselli, noted that, all “these vectors slide across the analytical plate like peas eaten with chopsticks” (as cited in Jones, 2016, p. 117).
We sought “better cutlery” and found tools suited for this task: the main tool was eye tracking, a technique that analyzes what people are looking at; the second tool was a special method of thinking aloud and the third tool was classical interviews, based on a questionnaire, to measure the recall performance of our test subjects.
What happens in the minds of people when they watch TV news? Mayer and Moreno (2003), psychologists and experts in multimedia learning, work with a dual-channel assumption to describe this process. Both were inspired by Baddeley’s (1998) theory of working memory, Paivio’s (1986) dual-coding theory and Chandler and Sweller’s (1991) cognitive load theory. According to Mayer and Moreno (2003), “the human information-processing system consists of two separate channels — an auditory/verbal channel for processing auditory input and verbal representations and a visual/pictorial channel for processing visual input and pictorial representations” (p. 44). But each channel has limited capacity; “Only a limited amount of cognitive processing can take place in the verbal channel at any one time, and only a limited amount of cognitive processing can take place in the visual channel at any one time” (Mayer& Moreno, 2003, p. 44). Baddeley (2003) broadens the model. He introduces a third component, the “central executive” (p. 830). This controls and coordinates the processing of both channels; it organizes connections to the long-term memory (p. 835), where all stored knowledge and experiences are located.
In other words, meaningful information processing occurs if humans receive information from both channels, organize the information into a coherent representation and establish connections between corresponding representations. However, and quite importantly, connections can only be established if the corresponding pictorial and verbal information is processed in both channels at the same time. If auditory and visual information diverge, one or both channels can become overloaded. According to Baddeley (2003), the auditory channel — or in his words the phonological memory store — “can hold memory traces for a few seconds before they fade ” (p. 830). For this reason, channels can be overloaded when auditory and visual material is not presented simultaneously; “there is a need to hold one representation in one channel’s working memory until the corresponding material is presented in the other channel” (Mayer & Moreno, 2003, p. 50). This additive memory work can cause cognitive overload. For our study, the model is helpful to theoretically underpin our findings.
Former Studies of TV News
Research into the perception of TV news has a long history. Of particular interest are questions about overloaded working memory, and thereby, disturbed viewers’ information processing. In Germany, research into TV news, and particularly Tagesschau, began in the mid-1970s (cf. Bonfadelli & Friemel, 2015, p. 117 ff.; Renckstorf, 1980; Straßner, 1982). International research is also represented, for example, by Drew and Grimes (1987). They recorded several TV news reports and then manipulated them by cutting some of the material out. This created versions where the verbal and pictorial information was not synchronous, to greater and lesser degrees. In other versions, auditory and visual content were assimilated. The result was that the versions with synchronized information led to significantly better recall; in cases of mismatch, only the pictures were memorized.
Findahl (2009) followed a similar research design by modifying five TV news reports and creating eight versions of each one with different degrees of complexity. The result was comparable to the findings of Drew and Grimes (1987): the better the language and pictures were synchronized, the better the recall of the content was. Corroborating evidence also comes from other studies: Reese (1983) and Brosius, Donsbach and Birk (1996) also tested several variations of news reports; Bergen, Grimes and Potter (2005) studied CNN Headline News. Josephson and Holmes (2006) explored the impact of crawlers and headlines on recall using news reports from foreign countries to avoid the viewer having had already seen them. A more recent survey comes from Portugal: Rodrigues, Veloso and Mealha (2016) investigated the influence of crawlers and headlines on the perception of news presentations and reports. Our own study attempted to find out whether Tagesschau‘s new studio-background design, compared to the old one, is better, worse or just as well suited to successfully transfer information.
Working Hypothesis and Research Questions
In accordance with Goertz and Schönbach (1998), who discovered that too many inscriptions, background pictures or graphics have a negative impact on the recall of TV news, we formulated our working hypothesis:
The new design with the large media wall carries the danger of overloading the viewers’ pictorial channel and distracting them from the content that is verbally transmitted.
Our investigation was guided by the research question:
Which elements of the new studio design related to background illustrations and camera angles will support or impair viewers’ information processing?
Object of the research: Tagesschau
For our study, we had exceptionally good conditions related to our object of research. The public broadcasting service, ARD, had built a brand new TV studio in Hamburg using a very different, modern design. To test the new technical equipment and the different stage designs, the ARD produced several newscasts in two versions over a number of weeks. Version one was broadcast live from the old studio and version two was produced and concurrently recorded in the new studio. Both versions were almost identical in terms of content but different with regard to design and camera angles. We were able to access the two versions recorded on 11 April 2014, 5 p.m. Tagesschau program, for scientific purposes. By comparing both versions, the impact of the background design style and camera angles could be tested in a real-world setting. The project was designed as free university research, not paying contract research.
In the old studio’s design, the largest part of the background was blue, with blurred traces of a world map. A smaller section, approximately 30% of the background, to the right-hand of the presenter, was filled with photos related to the topic, and a headline was generally placed above the illustrations (Fig. 1). In the new studio with the large media wall, the background is normally filled with impressive pictures, also related to the topic. A broad, blue, half-transparent strip in the lower section separates the background and the host. If there is a headline, it is pushed down to the lower half of the picture (Fig. 2).
Fig. 1: Old studio design with S. Holst Fig. 2: New studio design with S. Stichler.
In both versions the editor team used the same background photo to topic 4, a software security gap caused by a programmer. In the old version, headlines sat above the relatively small background pictures; in the new version the headlines are placed in the bottom half of the full screen background illustration.
An additional camera angle, using a different projection onto the media wall, was developed for the afternoon edition of Tagesschau. It is an “American-type” shot that shows the presenter from the knees up. Next to the host, large panorama pictures, maps or videos are seen (Fig. 3). For these shots, the designer relinquished the headlines. To test this variant, we purposefully analyzed an afternoon edition.
Fig. 3: “American-type” shot.
At variance with most other studies in this sector, our research focused on the presenters’ activities. We displayed shortened versions of the news reports relating to each topic, but we displayed them only to preserve the impression of a complete newscast. In the end, both test videos were no longer than 6 minutes and 40 seconds. The original length of 15 minutes would have been too long for our project and our limited amount of time for testing. To analyze the viewers’ information processing of these shortened newscasts, we used a bundle of research tools: eye tracking, thinking aloud and interviews.
Eye tracking is a tool for analyzing what people are looking at. The eye tracker, a device placed in front of the viewer, beams a light with a frequency similar to that of infrared into the subjects’ eyes, and their eyes reflect this light. A small camera connected to a computer records these reflections. Special software analyzes and evaluates the data and records what the viewer focused on (Tobii, 2017). Currently, researchers in communication and media sciences explore quite different issues with eye tracking. They investigate, for instance, how people read their newspaper or use online pages (Bucher & Schumacher, 2012) and they try to determine, “the degrees of spectators’ visual concentration . . . during a TV newscast” (Geise, 2011, p. 153).
Scholars, however, state that eye tracking does not convey completely reliable data. In many cases, eye tracking shows a narrow correlation between recorded fixations and cognitive processing; but there are exceptions. Documented eye movements, for instance, are sometimes not connected with shifting cognitive focus. Reasons for this are effects like extrafoveal or peripheral vision or covert attention. Therefore, we noted, determining with the help of eye tracking what a participant is looking at does not generally mean we know what the participant is seeing (Schumacher, 2012, p. 115). To get more, and especially more reliable information it was necessary to combine eye tracking with other research tools like thinking aloud.
The think aloud method
Think aloud is an established and approved method (cf. Guan, Lee, & Cuddihy, 2006). In many eye-tracking studies, researchers apply simultaneous think aloud; subjects are asked to share all their thoughts during a test. However, if you try to analyze TV news with visual and auditory stimuli, simultaneous think aloud cannot be applied. The participants would not be able to hear the spoken text while they are talking. Therefore, we chose a variant called retrospective think aloud (Schumacher, 2012, p. 130-131).
The participants watched the test video again, after the eye tracking. This time the video ran without the soundtrack and showed the subject’s gaze plots. These colorful dots indicate where the gaze stops and how long the gaze remains at a certain point. The longer the gaze fixes on a point, the bigger the dot becomes. Therefore, the participants were able to see for themselves what they had been looking at. Additionally, we asked them to tell us their feelings and thoughts.
The recall measure
Ahead of the think-aloud session and some minutes after the eye tracking, we interviewed the participants. With the aid of a standardized questionnaire, we checked the quality of their information processing by measuring their recall performance. However, recall measures are disputed. Many scholars argue that this method does not measure the right issues; scholars differentiate between comprehension and memory as two distinctive processes. Comprehension methods test if participants have integrated the information into a meaningful system of already existing knowledge. Such measures should answer the question, how well did somebody grasp the gist of a story or can apply the newly gained information in other contexts. Recall, critics say, in contrast, only shows that a human is able to memorize more or less peripheral details like figures, names of people or cities (Grabe, Bas, & van Driel, 2015, p. 302). All these viewpoints are justified. We used the instrument of recall measuring, nonetheless, because we only intended to compare two different visual TV news versions with identical content. We did not want to find out the absolute quality of the comprehension. Therefore, we considered that recall measuring suited for our purpose.
Each version of the Tagesschau video was seen by a group of 16 test subjects. The total of 32 volunteers, although small, was large enough for our research, an explorative study (cf. Bortz & Döring, 2006, p. 71). The most senior individual in the group was 72 years old and the youngest was 21. The sample comprised three professors, of mechanical engineering, physics and law, respectively; scientific and technical employees; staff members of the university library and the canteen; secretaries and students. The participants in our research were recruited by asking them directly.
Such a convenient sample is not representative of educational background and can induce distortions when measuring recall performance (cf. Henrich, Heine & Norenzayan, 2010, p. 2). With a global view, Brookshire (2013) labels the participants of such panels as “overwhelming Western, educated, and from industrialized, rich, and democratic countries” (WEIRD). Brookshire (2013) states that WEIRD subjects represent only about 12 percent of the world’s population and differ from other populations, for instance, in “things like visual perception.” Studies can, therefore, be remarkably distorted when working with such panels. However, to test a German TV newscast to gain initial hints of the impact of distinctive background designs and camera angles, we deemed our test group appropriate. In addition, we did not check the subjects for existing special-topic interests or previous thematic knowledge, which can, by trend, also cause distortions (Bonfadelli & Friemel, 2015; Knobloch, Zillmann, Gibson, & Karrh et al. 2006). We accepted all these shortcomings in order to keep the interviews simple and make them easier to compare. Besides, possible discrepancies would have been conspicuous when evaluating the data due to the small sample.
The two participant groups, for the two Tagesschau versions, were formed by randomly dividing the sample. Group A watched the live broadcast of Tagesschau from April 11, 2014, presented by the anchor, Susanne Holst. Group B watched the second version, which was produced at the same time in the new studio with Susanne Stichler, who is the official replacement person for Susanne Holst in cases of illness or vacation.
The tests took place in two seminar rooms. For our experiment, we used two eye-trackers obtained from the Swedish manufacturer, Tobii Pro. The devices, which transmit infrared light and record eye reflections, were positioned at the lower edge of the TV screen. This was quite comfortable for the test subjects because it made it unnecessary to uncomfortably hold their head in the same position, and the situation was closer to reality. However, it was not the same as reality because an experiment always generates a more or less artificial ambiance. There was also an additional source of stress. Some subjects felt under pressure because of the eye tracker. They knew that they were being tested, and they were making an effort not to embarrass themselves. They reflected on the potential significance and ultimate intention of the experiment, because, to avoid influencing the results, we had purposefully not informed the test subjects of the research design. Despite these shortcomings and interferences, we can still say that the findings were informative and useful.
The presenters’ passages, analyzed by eye tracking, lasted between 20 and 30 seconds. The videos comprised five topics:
Topic 1: A debate in the German parliament, Bundestag, about genetically modified maize
Topic 2: A visit by the German chancellor, Angela Merkel, to Athens, Greece
Topic 3: The comeback of nuclear power in Japan
Topic 4: A software security gap unintentionally generated by a programmer
Topic 5: A report about the Berlin Airport during its construction and a further cost explosion
After the eye-tracking experiment was conducted, the test subjects were interviewed using a questionnaire to ascertain their recall. For this purpose, we used unaided as well as aided questions. The unaided questions, administered first, were supposed to help identify the topics the study participants remembered spontaneously. After this, each video topic was dealt with one by one, and the interviewer asked for certain details.
The test subjects were interviewed for five minutes after the eye-tracking experiment. For every correct answer, scores were given—three points for correct answers to unaided questions and a maximum of two points for correct answers to aided questions. A five-minute delay was necessary to avoid any distortion of recall via the short-term memory. Furthermore, it is recommended to interview subjects “after a task for distraction” (Gehl, 2012, p. 146). Our distraction task consisted of chatting with the students in a waiting room.
Following the interviews, we asked the participants to think aloud. The entire think-aloud session was recorded and later transcribed. We discovered that the think-aloud method could be optimized when the interviewer asked purposeful questions. In our preliminary tests, a variant of the method, we call stimulated retrospective think aloud, generated much better results than free association. Without stimulation, some subjects kept silent for a long time; they thought that their ideas and associations might not be interesting or relevant enough to be formulated. This behavior is mentioned in the scientific literature: “Among the disadvantages of the method is . . . that not all study participants are willing or able to verbalize their thoughts; time and again they say the logging would be unpleasant and embarrassing for them and they would feel ridiculous” (Geise, 2011, p. 237).
Only one person interviewed all the subjects during the think-aloud sessions; interviews often yield interesting points and these can be amplified with subsequent study participants. This method is not completely adequate because it makes the comparison of the data more difficult. However, the gain in findings compensates the method’s shortcomings. To concentrate on the thoughts and statements of the test subjects we sometimes stopped the video. The results of the think-aloud method yielded much information, in addition to the numerical data.
The recall performances for Topic 1 (parliamentary debate on genetically manipulated maize) and Topic 4 (software security gap) were almost the same for both studio versions. With respect to Topic 5 (Berlin Airport and cost explosion), the overall recall was slightly better for the new studio version, but the difference was not significant. We found significant differences between both versions of Topic 2 (Chancellor Merkel in Greece) and some smaller differences in single items related to Topic 3 (comeback of nuclear power in Japan). The old studio version generated better recall performance in both cases.
The recall related to Topic 2 was measured by four items. Subjects watching the old studio version scored an average of 1.8 points; however, the subjects watching the new studio version scored an average of only 0.5. This result is highly significant (t = 3.14, p < .01). The difference was elicited mainly by the question, “Which top politician did Merkel meet during her visit to Athens?” While the other single indicators related to the Merkel news item showed no significant differences, the overall index including all four indicators was significant (t = 2.14, p < .05).
During the presenter’s passage on Topic 2, the old studio design showed a photo of Chancellor Merkel jointly with Samaras, the former Greek prime minister. Below the photo, the inscriptions “Merkel” and “Samaras” were shown, and above the photo was shown the topic headline “Besuch in Athen” (Visit to Athens). The eye tracker revealed that the subjects gazed at the face of the Greek politician and read his name (Fig. 4).
Fig. 4: Spectators’ gaze plots for the old studio version.
In the new studio, the media wall showed only a large picture of chancellor Merkel pointing with her right hand. There were no inscriptions, and therefore, no specific visual indications of the topic (Fig. 5). The gaze plots disclosed that the test subjects looked at the presenter and repeatedly at Merkel’s larger-than-life face (Fig. 6). The extreme difference between both studio versions can be explained by Mayer and Moreno’s dual-channel theory; information processing will be particularly successful when humans receive information simultaneously via the auditory and the visual channel (cf. Mayer & Moreno 2003, p. 44). Related to Topic 2, the old studio version delivered the central news via both channels. Merkel and Samaras were mentioned in the spoken text (auditory channel), and both were shown at the same time in the background photo combined with a caption of their names (visual channel). The information that it was a political visit to Athens was also mentioned in the spoken text and in the headline at the same time. Therefore, information processing in the old studio version was much easier than in the new studio version, where only Merkel was shown and no further visual information was given.
Fig. 5: New studio design. Fig. 6: Viewers’ divided attention.
The analysis of the areas of interest (AOI) provided more insights. The AOI can be programmed into the eye-tracking software to count gazes and fixations on the object being investigated. Here we analyzed how long and how intensively the subject looked at the presenter’s face. When purposefully asked, the subjects told us that during the think-aloud session, that their longer-standing gaze at the face meant they were paying attention to what the host was saying. However, comparing both studio versions, we found a considerable discrepancy (Table 1). In the new studio version, only 64% of the test subjects looked at the presenter’s face during the seconds when the crucial information that would enable them to answer the question correctly was given, while in the old studio version, 88% did this (see Percentage Fixated column, Table 1). In the new version, test subjects on average only looked once at the presenter’s face during this time, whereas in the old version, they did this 2.47 times (visit count). Focusing on the presenter’s face was for shorter periods for the new studio version compared to the old studio version. The relative fixation time was 28.40% in the old version and 20.33% in the new version.
Table 1: Analysis of areas of interest (AOI) related to the presenter’s face results.
|Relative Fixation Time||Fixation Count||Relative Visit Duration||Visit Count||Percentage Fixated|
We also discovered a significant difference in the recall performance of Topic 3, the comeback of nuclear power in Japan. The background picture in the old studio design showed the nuclear hulk of Fukushima. Under the photo was the inscription “Fukushima, März 2011,” (Fukushima, March 2011) and above the photo was the headline “Japan setzt wieder auf Atomkraft” (Japan relies on nuclear power again) (Fig. 7).
Fig. 7: Old studio design for Japan’s nuclear power comeback.
In the new studio design, a large picture of three men standing in front of contaminated water tanks was shown. The men are wearing breathing masks and protective clothing—a scene photographed in the hulk of Fukushima. A slightly blurred inscription could be identified on the protective clothing of the man on the right-hand side. The eye tracker revealed that most of the subjects fixated on the group of men, especially on the inscription on the protective suit (Fig. 8).
Fig. 8: Nuclear power in Japan—the new studio design.
To check the recall of this topic, we asked the subjects, “How did Japan first react to the catastrophe of Fukushima?” The correct answer was that Japan had reacted with a decision to abandon nuclear power. Here, we found the second biggest difference. In the old studio version, the subjects scored an average of 1.3 of two possible points, whereas in the new studio version they scored only 0.6 (t = 1.80, p < .10).
The explanation for the lower recall performance for both topics in the new studio design appears to be the same. We found a general distraction caused by the large background photos combined with the American shot. Compared with the “frontal medium” camera shot in the old and new studio design, the relative fixation time related to background pictures was much longer. The eye-tracking data showed the difference: the average value was about 44% in the case of the American shot and only about 13% in the case of the frontal medium shot. Our analysis was restricted to the shots of the new studio. Both topics, Merkel in Greece and nuclear power in Japan, were presented in the new studio with an American shot.
In the new design, the ARD purposely relinquished the headlines when using the American shot. The intention was for better and more impressive visual language. However, the headlines seem to be important for understanding and recall. The eye tracking showed that 100% of the subjects focused on the headline whenever a headline was presented (Table 2).
Table 2: Perception of headlines.
Relative Fixation Time Fixation Count Relative Visit Duration Visit Count Percentage Fixated
Studio old 4.43% 73.00 5.11% 29.35 1
Studio new 2.47% 42.79 2.95% 12.07 1
The lower figures for the new studio in all other categories are easily explained. In the new studio, three of the five shots were American shots without headlines. In the old studio design, all the shots were combined with headlines. Interestingly, there were small differences in both versions between relative visit duration and relative fixation time, indicating that all viewers read the headlines when they became aware of them. This conforms to what Mayer and Moreno (2003, p. 44) observed. According to their study, humans receive written words via the visual channel and spoken words via the auditory channel. If spoken and written words are presented simultaneously, the perception of content via the two channels at the same time will be more effective.
In the context of Topic 3, the comeback of nuclear power in Japan, our eye-tracking tests revealed another source of distraction, which was reviewed and confirmed in a subsequent investigation. If a background picture contains any kind of inscription—not a headline, but an inscription in the photo motif itself—spectators try to decode it. If the inscription is difficult to read because it is small or maybe slightly out of focus, they spend a lot of time trying to decipher it. The inscribed motif might be a sticker on a suitcase, for example, or an advertisement, or, as in our case, an inscription on protective clothing. In the new studio version, 9 of the 10 test subjects (93%) became aware of such inscriptions as part of the background motif (Table 3). In the old studio version, with its considerably smaller background illustrations, the awareness rate was 82%. In the new version, the relative fixation time was twice as high as for the old version (1.59% and 0.71%, respectively), and the visit-count rate was three times higher for the new studio version. This indicates that the subjects were making a special effort to read the inscriptions.
Table 3: Perception rates of inscriptions in the background photo motifs.
|Relative Fixation Time||Fixation Count||Relative Visit Duration||Visit Count||Percentage Fixated|
The slightly blurred inscription on the protective suit in the Fukushima photo obviously distracted the subjects’ attention and prevented them from concentrating on the relevant verbal information. Another reason for the distraction could have been a tracking shot in this passage—beginning with an American shot and zooming to a medium close-up. The combination of these factors probably overburdened some of the subjects’ cognitive capacity. These are plausible deductions; we did not find comparable differences with the other topics.
An exception occurred for Topic 1, the debate about genetically modified maize. A picture of a large corncob formed the background for the moderation in both versions. In the new studio version, combined with an American shot, the media wall additionally displayed a yellow warning sign with the inscription “GENFOOD.” In the old studio, the corncob photo was complemented with a headline. Despite these differences, the recall performance was almost the same in both versions. We believe that the background photos were well suited in both versions because they were simple to perceive and did not distract from the spoken content. The distribution of the gaze plots indicated that the subjects listened intently to the presenter because they focused on the newscaster’s face for both studio versions (Figs. 9 and 10).
Fig. 9: New studio design for “gene maize.” Fig. 10: The old studio design.
In the new studio version, the American shot without a headline was used for three of the five moderation passages. The reason for this was that the American shot was a new element; the editors decided it should be tried repeatedly. However, this kind of shot is rarely used.
Topic 4, the software gap, was illustrated in both versions with a picture of a column of figures. In the center, it showed the inscription “password.” Both versions were given a headline (see Figs. 1 and 2). The recall performance was very similar for both versions.
Topic 5, the Berlin airport, was illustrated in the old studio design with a photo of a nameplate with the inscription “BER Berlin Brandenburg Airport.” In the new studio design, the background picture showed a Berlin airport terminal under construction. On the terminal building the name, “Flughafen Berlin Brandenburg Willy Brandt” (Airport Berlin Willy Brandt) can be seen in big letters (Figs. 11 and 12). Both studio versions contained a headline. The eye tracking revealed that viewers of both versions took a short glance at the pictures, read the headlines and inscriptions and then listened to the presenter. In this case, the recall performance was significantly (p < .10) better in favor of the new version. An explanation, therefore, could be that the headlines differed. The headline in the new version was more precise and meaningful. The phrase was, “Capital’s airport. Report about further cost increase.” The phrase in the old version was only, “Costs of Capital’s airport.” The issue of wording of headlines needs more investigation.
Fig. 11: Old studio design Fig 12: New studio design
Both pictures are related to Topic 5 “Cost explosion on Berlin airport.” The right-hand picture includes areas of interest (AOI) coding.
It was not always possible to compare the impact of identical photo motifs in the background because different illustrations had been used in some of the studio versions; the editing team had wanted to test the additional visual scopes of the new design. However, this was not a disadvantage for our research because we not only compared the material, we also separately analyzed every test subject’s percipience of every background motif by eye tracking and thinking aloud.
Graphical and numerical overview of the results
The subjects tested on the old studio version achieved an average score of 20.2 points, and the subjects tested on the new version scored 19.4 points. This difference was not significant (t = 0.37). However, the variation among the subjects was large: one subject scored only seven points and another subject scored 31 points, the highest score (the possible maximum score was 44 points). We concluded from these results that there was no general difference in recall performance between the old and new studio versions. Both versions were likewise suited to conveying the facts and information.
Chart 1 shows the comparison between the recall performances of the old and the new studio graphically. The vertical axis shows the scores of the recall performance and the horizontal axis documents the numbered questions (NQ), 1 to 18, from the questionnaire. The two columns at the bottom of the figure represent the exact mean for each item for both studio versions. The first question was unaided, while questions 2 through 18 were aided. Here, we asked the test subjects to mention all the topics presented in the newscast video they could remember. The possible maximum score was 15 points.
Chart 1: Comparison of recall performance between old and new studio versions. NQ = numbered questions.
Table 4 shows the differences in recall performance between both versions, related to significance.
Table 4: Significance of the recall performances.
Topic Group Mean Significance
Genetic Maize (index out of 4 items) old .59 not significant
new .56 Merkel in Athens (index out of 4 items) old .816 p < .05
new .416 Nuclear Power in Japan (index out of 4 items) old .80 not significant
new .66 Software Security (index out of 3 items) old .63 not significant
new .57 Berlin Airport (index out of 2 items) old .72 p < .10
We can summarize, there was almost no divergence in recall performance between the versions when medium shots were used and headlines were offered. However, when American shots were used with large background illustrations on the media wall without headlines, the quality of the information transfer was seriously diminished. In Chart 1, the line for the old studio version indicates the difference; two peaks on the graph mark the much better recall performance in comparison to the new studio version.
The average recall rate was 46.0% for the old studio version and 43.6% for the new studio version. The test subjects were not able to remember half of the news content in the moderation passages. This seems very modest at first sight, however, in comparison, the recall figures in many other studies are lower. Earlier investigations show average recall rates of 25% (Ruhrmann, 1989, p. 25; Bonfadelli & Friemel, 2015, p. 119). A recent, Portuguese TV news eye-tracking project surveyed the recall performance of 80 students aged between 18 and 30 years old and reported an average recall rate of 40% (Rodrigues et al. 2016, p. 77).
In a parallel line of research, we wanted to determine the correlation between recall performances and test subjects’ educational level. Many studies show “a significant relationship between education and recall performance” (Bonfadelli & Friemel, 2015, p. 120). The results of our research also confirmed this; the higher the subject’s level of education, the higher the recall rate.
Review of the findings
In an additional line of research, we reviewed our findings to learn more about the impact of studio design and camera shots on viewers’ information processing. We produced our own newscast in the university’s TV studio lab. For our experimental newscast, we made up stories. The advantage of this was that the subjects would definitely not recognize the news, because this news had never been broadcast, and the remembering of former information was excluded. It was essential to create the stories in a natural realistic manner. The material for the newscasts was taken from social media, news agencies and Tagesschau. We edited these videos and supplied them with new text. The newscaster’s passages contained exclusive information because they had been fabricated. None of the subjects realized that they were confronted with manipulated material.
The design and camera shots were similar to those of the Tagesschau’s newscasts we had previously tested. We introduced some modifications in terms of a different intro jingle and different branding. The subjects were told that this would be the pilot of a new TV news format. We created four versions with identical content. The presenter was a female student who had already moderated regional TV programs in the past. In each version, we varied the background and the camera shots and we either used headlines or did not (Figs. 13, 14, 15, 16).
With this material, we also tested the medium close up shot, which is sometimes used in the afternoon emission of Tagesschau. This shot generally lacks a headline. The background is blue with a part of the world map (Fig. 15). Note that the medium close up was not part of the initially tested original Tagesschau.
Four design variants of one topic, here “Terror in Kiev. Shooting in front of a police station”. From clockwise –
Fig. 13: Medium shot with headline and background photo; Fig. 14: Medium shot with headline and map in the background; Fig. 15: Medium close up shot with world map in the background; Fig.16: American shot without headline.
For this supplementary investigation, in the framework of a research seminar, we recruited for convenience, a panel of 20 volunteers who were all fellow students of the seminar participants. The test was conducted in the same way as described in the Methods Section. We again used eye tracking, thinking aloud and interviews with a questionnaire.
A panel size of 20 test subjects meant that there were only 5 subjects available for each version. This is a very small group. Because of that we do not present any results in form of statistics or detailed figures. However, combined with the results of the think-aloud sessions and compared with our former findings, we were able to obtain further information; it confirmed the results of our first survey and yielded further indications for our research question: “Which elements of the new studio design related to background illustrations and camera angles will support or impair viewers’ information processing?”
The central insight coming from the comparison of the original versions of Tagesschau, and confirmed by our self-produced material, is that headlines in the presenter’s area are very helpful for viewer’s information processing. The eye tracking, as well as the think-aloud results, showed that viewers value such headlines. The measured recall performance was generally better in passages with headlines. During their thinking aloud, some participants said it would be helpful for comprehension if the headline was a displayed word-for-word of the presenter’s spoken text.
Also confirmed had been the insight that background pictures with any kind of inscription (not headlines) are extremely distracting. The subjects of the new project too regularly tried to read the inscription, and during this time, they did not follow the spoken information. This can seriously impair the process of comprehension.
Confirmation we got too for the insight that the American shot without a headline, but with large background photos, did not aid the perception of more complex topics. Our supplementary tests, however, showed that the American shot is well-suited for transferring amusing tales with a simple storyline. In one particular self-produced test version, a video, rather than a photo of a giant crawling salamander was projected on the media wall behind the host. Four of the five subjects were able to remember this and achieved a full score. Then we tested another background motif. In a similar shot, subjects saw a war scene from the Middle East on the media wall. The video contained hard cuts and some of the pictures were unsteady. The distraction this caused was remarkable—not one of the subjects could recall the topic or its content.
Related to the medium close-up camera shot, we found surprising cues. This shot (Fig. 17) is part of the new Tagesschau design but was not part of the originally tested versions. Therefore, we could analyze its impact only in our self-produced news. The shot seemed to pose a problem for the majority of collegiate subjects. During the think-aloud sessions, the following types of comments were common: “The host is gazing at me,” “I looked only at her face. I couldn’t follow what she said,” “I didn’t get a chance to look at anything apart from her face. I did not like this” and “That is much too close.”
Fig. 17: Medium close-up shot.
Such reactions were accompanied by a poorer recall performance. The moderation part of our fabricated topic called “Shooting in Kiev,” for instance, was presented in a medium close-up shot and a medium shot with a headline (compare Figs. 15 and 13). Viewers of the medium-shot version with a headline scored an average of 2.4 points; the viewers of the medium close-up shot, however, only scored 0.4 points. The differences were not always large, but the medium close-up shot tended to generate lower recall performance. This is a baffling result; it would appear to dissent from some American research. Graber (1990, p. 146) found that close-up shots of familiar people are especially well suited for transferring information. Graber, however, referred to close-ups in reports, not in moderation portions. An explanation for this might be that our test subjects were sitting closer to the monitor because of the demands of the eye-tracking technique set-up. The host possibly appeared much larger-than-life and invaded their personal space. In addition, we conceded that our host was not familiar to our subjects through previous TV programs; unknown faces possibly distract more from content and the face of a well-known presenter might produce better recall performances. More research is recommended.
We randomly obtained another insight; we found that on the first view, appropriate background pictures could have a negative effect. In our self-produced material, we used a medium shot with a headline and a photo of a big white column with a VW logo (Fig. 18). This well-structured motif has often been part of a background design on Tagesschau.
We thought that using this might lead to good recall performance, but we were wrong; during the think-aloud sessions, the study participants told us why. The VW logo was a fatal cue because VW’s exhaust scandal had reached its peak by the time of this study. A subject said, “I saw the logo and I knew what a totally boring topic dealing with VW would follow. I switched off immediately.”
Fig. 18: Stele with VW logo.
Repetitive use of the same background illustrations accompanying long-term themes can apparently cause tedium, and in that way, a seriously reduced willingness to perceive the information.
We compared two versions of the same TV newscast, with identical verbal content but different visual presentations, in order to learn something about the impact of various camera angles and background arrangements on viewers’ information processing. As a departure point, we tested original material from the German public broadcasting system, ARD. For review purposes, we additionally produced, in the university’s media lab, our own newscast. Our working hypothesis was that the new design with the large media wall carries the danger of overloading the viewers’ pictorial channel and distracting them from the verbally transmitted content.
This hypothesis was not confirmed at first sight. We found no general difference between the old studio design of Tagesschau and the new one. Both studio designs were applicable for supporting information gain, presupposing the use of a medium shot with headlines. This was not surprising because the medium shot in the new studio version was visually more up-to-date but similar to the medium shot in the old version. Moreover, our study showed that the better the background illustrations and headlines contents’ accorded with the presenters’ spoken text, the better the recall performances were.
This corroborates Mayer and Moreno’s (2003) dual-channel theory of information processing; humans receive information via two separate channels, the auditory/verbal one for auditory input and verbal representations and the visual/pictorial one for visual input and pictorial representations (p. 44). Meaningful information processing will occur if humans receive information via both channels simultaneously. If auditory and visual information diverge, however, one or both channels can be overloaded (p. 50). In such cases, meaningful information processing will be complicated or rendered impossible.
We found such a mechanism related to the American shot, which is only part of the new studio design. Our hypothesis was supported by this shot. In the new Tagesschau format, the American shot is generally used without a headline in order not to derogate from the impressive background motif. In comparison to the medium shot with a headline, the American shot seemed to be worse suited to conveying complex information. We found an exceedingly significant difference between both shots in two topics of the original ARD material. For Topic 1, the visit by the German chancellor, Merkel, to Athens, in the old studio version, the background was a medium shot presented with a headline and a photo of the two politicians (Merkel and Samaras), who were mentioned by the newscaster. In the new studio version, in the background, on the large media wall, there was presented an American shot of only one (Merkel) of the two mentioned politicians and no headline. The recall performance in the first version was more than three times higher than the second version. In version one, the verbal and pictorial information supplemented each other. In version two, there was a lack of pictorial information because of the absent second portrait. Moreover, the viewers were distracted by the large photo of Merkel’s face combined with her big hand-gesture and they received no support from a headline; the viewers’ information processing was disturbed.
For Topic 3, the comeback of nuclear power in Japan, the old studio version presented in a medium shot, a relatively small photo of the destroyed nuclear plant in Fukushima combined with a headline and a cutline. The new studio version initially presented the issue with an American shot and then zoomed to a medium close-up. On the media wall, one could initially see three workers, standing in front of large tanks, wearing breathing masks and protective clothing. On one protective suit, an inscription could be identified. The eye tracker revealed that many of the subjects tried to decipher the inscription and did not listen during this time to what the newscaster said. The recall performance in the old studio version was more than two times higher than the new studio version. Therefore, we learnt that the American shot could seriously impair viewer’s information processing.
We concluded that the new studio design is not worse, but is not better suited for conveying news information when using medium shots with a headline. However, when using an American shot, the new studio design was considerably less-suited to supporting the viewers’ information processing. The American shot had a tendency to create a distracting effect, and seems, therefore, not to be appropriate for transferring complex and difficult issues.
Therefore, we could answer our leading research question: Which elements of design support or impair information processing as measured by recall performance? Information gain will be supported by medium shots with headlines and will be impaired by American shots with large, detailed photos without headlines. Especially distracting, are all kinds of inscriptions inside the background motif.
Future studies should check if a newscaster’s text that integrates the background motif of an American shot with deictic wording causes better information processing. The studies should investigate what impact on viewers’ information gain the media wall has combined with long shots, which are used in other TV news’ formats. Future investigations may generally lead to finding more answers to the question: Which kinds of background arrangement are particularly valuable for successful information processing?
Because of the small and potentially distorting panels in both steps of our investigation, we have to be careful to generalize the results. Many of our subjects have a very high educational level and are skilled in decoding verbal information. This could be one explanation for the subject’s often-wished headlines. In addition, this could be one explanation for why many test subjects tried to decipher inscriptions wherever they appeared in a background photo. Perhaps disparities exist for people with lower education levels and other ethnic groups (Brookshire, 2013). Future studies should, therefore, perform tests like ours with a much larger group of participants. The subjects should be chosen to represent a wider age range and a more diverse educational and ethnic background to check the extent of possible distortions. Nevertheless, our results agreed with information processing theory and they are coherent; therefore, we believe that in spite of the previously mentioned constraints, our findings are adjuvant cues for future research and news journalism practice.
The comprehension of TV news is not only a multimodal but also a multifactorial action. A person’s interest in certain topics, their previous knowledge about an issue, the intensity of their media use, their formal education and other factors may assume a greater role for understanding TV news (Bonfadelli & Friemel, 2015, p. 120). TV news producers cannot influence all these factors; however, they can optimize the news studio’s arrangement to foster information processing, and therefore, intelligibility.
The German TV newscast, Tagesschau, that we tested, is, despite its problem of being linear television in times of digital media, still one of the most successful news formats worldwide. Related to its design, it has one of the more conservative formats. It operates, for example, without permanent crawls or other data on the screen, like weather information or stock exchange news. The use of a media wall to show impressive background pictures during moderation passages, often combined with wide camera angles, is, however, the international standard. The results of our tests, therefore, are also helpful indicators for producers of other newscast programs.
Baddeley, A. (1998). Human memory. Boston: Allyn & Bacon.
Baddeley, A. (2003). Working memory: Looking back and looking forward. Nature Reviews Neuroscience, 4, 829-839.
Bergen, L., Grimes, T., & Potter, D. (2005). How attention partitions itself during simultaneous message presentations. Human Communication Research, 31(3), 311-336.
Bonfadelli, H., & Friemel, T. N. (2015). Medienwirkungsforschung [Research of media impact]. Konstanz/Germany: UVK, 5th ed.
Bortz, J., & Döring, N. (2006). Forschungsmethoden und Evaluation für Human- und Sozialwissenschaftler[Research Methods and Evaluation for Human- and Social Scientists]. Heidelberg/Germany: Springer, 4th revised ed.
Brookshire, B. (2013, May 8). Psychology Is WEIRD. Western college students are not the best representatives of human emotion, behavior, and sexuality. Retrieved from http://www.slate.com/articles/health_and_science/science/2013/05/weird_psychology_social_science_researchers_rely_too_much_on_western_college.html
Brosius, H.-B., Donsbach, W., & Birk, M. (1996). How do text-picture relations affect the informational effectiveness of television newscasts? Journal of Broadcasting & Electronic Media, 40(2), 180-195.
Bucher, H. J. (2012). Intermodale Effekte in der audio-visuellen Kommunikation: Blickaufzeichnungsstudie zur Rezeption von zwei Werbespots [Intermodale effects in the audio-visual communication: eye tracking study of the reception of two advertising spots]. In: H.-J. Bucher & P. Schumacher (Hrsg.): Interaktionale Rezeptionsforschung, 257-296.
Bucher, H.-J., & Schumacher, P. (Hrsg.) (2012). Interaktionale Rezeptionsforschung. Theorie und Methode der Blickaufzeichnung in der Medienforschung [Interactional reception research. Theory and method of eye tracking in media research]. Wiesbaden/Germany: Springer VS.
Chandler, P., & Sweller, J. (1991). Cognitive load theory and the format of instruction. Cognition and Instruction, 8, 293–332.
Drew, D. G., & Grimes, T. (1987). Audio-visual redundancy and TV news recall. Communication Research, 14, 452-461.
Findahl, O. (2009). The effect of visual illustrations upon perception and retention of news programmes. Communications, 7(2-3), 151-168.
Gehl, D. (2012). Concept mapping und eyetracking: Eine Methodenkombination zur Diagnose medial initiierter Wissenszuwächse [Concept mapping and eye tracking: A combination of methods to diagnose knowledge increases initiated by media]. In: H.-J. Bucher, & P. Schumacher (Hrsg.): Interaktionale Rezeptionsforschung, 135-155.
Geise, S. (2011). Eyetracking in der Kommunikations- und Medienwissenschaft: Theorie, Methode und kritische Reflexion [Eye tracking in communication- and media science: theory, method and critical reflection]. Studies in Communication/Media, 2011, (2), 149-263.
Goertz, L., & Schönbach, K. (1998). Zwischen Attraktivität und Verständlichkeit. Balanceakt der Informationsvermittlung [Between attractiveness and intelligibility. Balancing act of the information transmission]. In: K. Kamps, & M. Meckel (Hrsg.): Fernsehnachrichten. Prozesse, Strukturen, Funktionen. Opladen/Germany: Westdeutscher Verlag, 111-126.
Gottfried, J., & Shearer, E. (2017, September 7). Americans’ online news use is closing in on TV news use. Retrieved from http://www.pewresearch.org/fact-tank/2017/09/07/americans-online-news-use-vs-tv-news-use/
Grabe, M. E., Bas, O., & van Driel, I. I. (2015). Defecting from the Gutenberg legacy: Employing images to test knowledge gaps. Journal of Communication, 65, 300-319.
Graber, D. A. (1988). Processing the news: How people tame the information tide. White Plains, NY: Longman.
Graber, D. A. (1990). Seeing is remembering: How visuals contribute to learning from television news. Journal of Communication, 40, 134-156.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The Weirdest People in the World? Working paper Nr. 139, Berlin: German Data Forum. Retrieved from https://www.ratswd.de/download/RatSWD_WP_2010/RatSWD_WP_139.pdf
Hertel, N. (2012). Animierte Informationsgrafiken im Fernsehen: Zur Rezeption non-linearer Darstellungsformen in einem linearen Medium [Animated information graphics in TV: about the reception of non-linear display formats in linear media]. In: H.-J. Bucher, & P. Schumacher (Hrsg.): Interaktionale Rezeptionsforschung, 239-255.
Hölig, S., & Hasebrink, U. (2017). Reuters Institute Digital News Survey 2017. Ergebnisse für Deutschland [Reuters Institute Digital News Survey 2017. Results of Germany]. Arbeitspapiere des Hans-Bredow-Instituts Nr. 42 [Publication of Hans-Bredow-Institute, Hamburg/Germany]. Retrieved from https://www.hans-bredow-institut.de/uploads/media/Publikationen/cms/media/2d87ccdfc2823806045f142bebc42f5f039d0f11.pdf
Jones, B. (2016). Television news and the digital environment: A triadic multimodal approach for analysing moving image media. African Journalism Studies, 37(2), 116-137.
Josephson, S., & Holmes, M. E. (2006, March). Clutter or content? How on-screen enhancements affect how TV viewers scan and what they learn. In Proceedings of the 2006 on Eye tracking research & applications (ETRA) (pp. 155-162). New York: ACM Press.
Game, A. (2001). Creative ways of being. In J. R. Morss, N. Stephenson, & J.
F. H. V. Rappard (Eds.), Theoretical issues in psychology: Proceedings of the International Society for Theoretical Psychology 1999 Conference (pp. 3-12). Sydney: Springer.
Knobloch, S., Zillmann, D., Gibson, R., & Karrh, J. A. (2002). Effects of Salient News Items on Information Acquisition and Issue Perception. Zeitschrift für Medienpsychologie, 14, 14-22. https://doi.org/10.1026//1617-63126.96.36.199
Mayer, R., & Moreno, R. (2003). Nine ways to reduce cognitive load in multimedia learning. Educational Psychologist, 38(1), 43-52.
Nielsen, R. K., Sambrook, R. (2016). What is happening to Television News? Retrieved from http://www.digitalnewsreport.org/publications/2016/what-is-happening-to-television-news/
Paivio, A. (1986). Mental representations: A dual coding approach. Oxford, UK: Oxford University Press.
Piazza, R., & Haarman, L. (2015). A pragmatic cognitive model for the interpretation of verbal-visual communication in television news progammes. Visual Communications, 15(4), 461-486.
Pinterest (2018). Photo gallery of glittering news studio sets. Retrieved from https://www.pinterest.de/ricardaschlegel/news-studios/
Prior, M. (2014). Visual political knowledge: A different road to competence. Journal of Politics, 76(1), 41-57.
Reese, S. D. (1983). Improving audience learning from television news through between-channel redundancy. ERIC No. ED229777.
Renckstorf, K. (1980). Nachrichtensendungen im Fernsehen (1). Zur Wirkung von Darstellungsformen in Fernsehnachrichten [Newscasts in television (1). The impact of display formats in the TV news]. Berlin/Germany: Spiess.
Rodrigues, R., Veloso, A., & Mealha, O. (2016). Influence of the graphical layout of television news on the viewers: An eye tracking study. Observatorio 10(1), 67-82. Retrieved from https://www.researchgate.net/publication/296484365_Influence_of_the_graphical_layout_of_television_news_on_the_viewers_An_eye_tracking_study
Ruhrmann, G. (1989). Rezipient und Nachricht. Struktur und Prozess der Nachrichtenrekonstruktion [Recipient and news. Structure and process of news reconstruction]. Opladen/Germany: Westdeutscher Verlag.
Schumacher, P. (2012). Blickaufzeichnung in der Rezeptionsforschung: Befunde, Probleme und Perspektiven [Eye tracking in the reception research: findings, problems and perspectives]. In: H.-J. Bucher, & P. Schumacher (Hrsg.): Interaktionale Rezeptionsforschung, 111-134.
Straßner, E. (1982). Fernsehnachrichten: Eine Produktion-, Produkt- und Rezeptionsanalyse [Television news: an analysis of production, product and reception]. Tübingen/Germany: Niemeyer.
Tobii (2017). Wie funktioniert Eyetracking? [How does eye tracking work?] Retrieved from http://www.tobiidynavox.de/wie-funktioniert-eyetracking/