Archives of Design Research

Through the innovation of digital technology, a great amount of information is being produced, and the development of mobile media has enabled access to information regardless of time or location. Thus, various attempts are being made to present information, and infographics have come into the spotlight lately.

Infographic is a compound word comprised of information and graphic. It is a visual representation of data, which delivers messages concisely. As stated above, the definition of infographics is “the visualization of data or a concept for prompt and effortless understanding by the user” (Smiciklas, 2012). Jung (2012) separated infographics from images and photographs by stating that infographics deliver information in detail, in a practical manner. She emphasized that infographics excel in the generation of interest, direct acquisition of information, and their ability to spread quickly. Infographics are used to turn vast and complex data into simple information that is easy to comprehend for the general public. They are also used as a means of analysis to identify parameters and causal relationship through information (Adams, 2011).

Edward R. Tufte is the most commonly mentioned name when searching for detailed definitions of infographics. Tufte created the notion of “chart junk,” and insisted that graphic elements unrelated to the information have to be eliminated in order to precisely represent the data.

British graphic designer Nigel Holmes is mentioned with an opinion contrary to that of Tufte. He suggested using ample amounts of graphics and design elements in information design. Visual metaphors and quality designs that have deep connections with their content support the subject and make it more attractive, which helps in recollection of the information. Their contrasting opinions were mentioned in various studies afterward. Bateman et al. (2010) set comprehension and recollection as dependent variables, and conducted a verification experiment in order to study the effects of chart junk. Decorative graphs and minimal graphs did not show a significant difference in terms of comprehension. However, the decorative graph received a higher score in comparison to the minimal graph in recollection after two to three weeks. Moreover, the participants preferred the decorative graph.

Haroz et al. (2015) applied pictograms to isotypes and bar graphs, and found that design elements were effective, aided in their recollection, and let participants look into the information in more detail when used as parts of data mapping. However, unnecessary decorative images distracted and confused the participants, acting as chart junk. Thus, it is essential in infographic design to be able to identify chart junk from valuable elements, and to utilize them appropriately.

The expectancy-value theory suggests that the pursued fulfillment is influenced by beliefs and evaluations of media objects (Hyun, 2008). Therefore, importance is now being given to users’ emotional evaluations. In gratification of the expectations, satisfaction refers to the emotional response to the product, which differs from attitude. Satisfaction influences attitude, and is thereby deeply connected with the behavioral intention that follows. When applied in media usage, the relationship between expectations and acquired gratification is linked to satisfaction, which is related to various results regarding usage in media. In satisfaction, emotional values such as satisfaction or dissatisfaction can be involved. Thus, Palmgreen and Rayburn (1982) and Blood and Galloway (1983) insisted that emotional evaluations should be included in the model.

Moon and Kim (2001) identified variables that affect perceived pleasure; Sim (2009) identified variables that affect participation; and Park and Kim (2010) identified variables that affect attitudes such as entertainment, participation, and reliability. DeLone and McLean (1992) measured online service systems’ information quality by informativeness and entertainment. Thus, existing studies support the notion that a user’s emotional value judgment affects the belief in, and evaluation of, the information in accepting it. Therefore, in this study, informativeness, entertainment, and reliability were used as variables for judging the value of information on the basis of these previous studies.

The AIDA model, DAGMAR model, Lavidge Steiner model, and McGuire model are representative models of the cognitive information processing of the audience, but the process from the exposure to the information to the action is shown although there is a slight difference in the process. According to the information processing process at the receiver’s end, the cognitive stage, which is exposed to information, and the affective stage, which has appeal or desire, leads to the behavioral stage (see Table 1).

According to this, a user can form an emotion after seeing the infographic, form a secondary attitude based on the emotion, and share and store the information. Therefore, the degree to which users can easily understand information, attitudes toward information, and intention to store or recommend information are all set as communication effect evaluation items.

As a part of the exploratory analysis, to understand the expression types of interactive infographics, and to formulate the experiment plan, we planned a case analysis.

The procedure of the case analysis is divided into four steps (see Table 2).

Case Collection	Collection of interactive infographics examples dated before November 11, 2014 USA Today (18 examples) The New York Times (20 examples) Yonhap News Agency (22 examples)
Curation of Analysis Cards	Printed 60 sets of analysis cards with images and their properties
Classification of Design Types	Extraction of main design properties through classification, and charting employing the KJ method
Result	Preparation of the supporting data for the experiment standards

1) Case Collection: First, we searched for examples of interactive infographics dating before November 11, 2014. We collected 10 to 20 examples from each of the following three interactive news media websites by screen-capture or recording: Yonhap News Agency: Graphic Visual News, The New York Times: The Upshot, and USA Today: Interactive Graphics.

2) Curation of Cards for Analysis: We printed 60 sets of A6-size cards that enabled clear analysis of their contents and expressions. Since it was difficult to show all expression methods on one card, we created a set of cards, which consisted many copies of captured images in order to analyze every characteristic of the interaction (see Figure 1).

Figure 1 Cards for analysis

3) Classification of Design Types: We categorized and reorganized prepared cards, and recorded our notes of viewing them in different angles and criteria. Various methods of storytelling, colorful videos, and interaction techniques were used (see Table 2). Based on 60 interactive infographic cards, the affinity diagram was used to structure the characteristics of the interaction. We created 60 sentences by describing the interaction characteristics of each card in the post-it. We classified and grouped the statements about each interaction (see Figure 2).

Figure 2 Classification of design types using the affinity diagram

Through our assessment, we were able to divide them into two types: (i) functional expressions that improve the functionality by accentuating and adding information, and (ii) emotional expressions that entertain users and invite their participation. Functional expressions of interactive infographics included functional elements such as adding or subtracting information, accentuating necessary information through interaction, presenting relationships between multiple graphs, and providing selected information in detail through movements or changes in chart types. Emotional expressions included elements such as showing colorful animations—for example, tables put together—and presenting interesting movements without changing the provided information as users interact with them (see Table 3).

The survey was conducted on college students in Seoul and Chungcheong area, from March 24 to April 7, 2014. Of the 260 responses received, we excluded 8 incomplete responses and utilized the remaining 252 in our statistical analysis.

As noted above in our case analysis, we have found two main types of expressions: functional and emotional. We focused on expressions consisting movements or interactions, and did not focus on storytelling, topic, or forms of graphs as they have been extensively studied in preceding case analyses.

We created three variables: functional expressions, emotional expressions, and no movement (see Table 4).

Interactions were triggered when the user hovers the mouse over the human-shaped bar graph (see Figure 3). For functional expressions, new information was presented through a tooltip and the area was accentuated with color. For emotional expressions, the shape of the graph changed playfully, matching the topic of obsessive-compulsive disorder. The control group with no interactions showed no change upon mouseover.

Figure 3 Samples of interactive infographics for the experiment

Top: Functional Interaction / Middle: Emotional Interaction / Bottom: No Interaction

Referring to prior research on uses and gratifications theory, we picked “entertainment,” “informativeness,” and “reliability” of infographics as measures that can affect the value assessment of information; and as measures to see the effect of the aforementioned elements to user response, we added “attitude,” “cognitive effects,” and “acceptance intention” to the infographic (see Figure 4).

Figure 4 Measures

Part 1: Measures that affect the value assessment of the information

Participants check the interactive infographic samples on the monitor and evaluate their assessment in terms of entertainment, informativeness, and reliability on a 7-point scale.

Each measure is defined as follows:

•      -Entertainment: The amount of enjoyment gained through the infographics
•      -Reliability: The reliability the user feels about the information portrayed in the infographics
•      -Informativeness: The amount of social and personal benefits gained through the infographics
•      Q1–Entertainment: These infographics are interesting.
•                                      It was enjoyable to see these infographics.
•                                      These infographics are entertaining.

Upon reliability analysis, Cronbach’s alpha was at .903, which led us to judge it was suitable for measuring the entertainment element of the infographics.

•      Q2–Reliability: I can trust these infographics.
•                                These infographics were not exaggerated.
•                                These infographics seem to be true.

Upon reliability analysis, Cronbach’s alpha was at .881, which led us to judge it was suitable for measuring the reliability of the infographics.

•      Q3–Informativeness: These infographics are in-depth.
•                                          I find these infographics useful.
•                                         These infographics provide information that I need.

Upon reliability analysis, Cronbach’s alpha was at .793, which led us to judge it was suitable for measuring the informativeness of the infographics.

Part 2: Value assessment of the information and user response

In order to identify elements that affect value assessment of the information and their relationship to user response, we have defined “attitude,” “perceived cognitive effects,” and “acceptance intention” as follows:

•      -Attitude: The extent of positivity or negativity that a user feels to perceived information
•      -Cognitive effects: The perceived extent of how well the contents of the infographics are delivered to the user
•      -Acceptance intention: Intention to accept the information and save, recommend, or share it

Cognitive effects have always been deemed important in research related to infographics, and numerous tests have been conducted to measure them. However, in this experiment, we aimed to measure users’ perceived cognitive effects. Also, we have based the survey on advantages of infographics, which include their ease of understanding and eye-catching characteristics.

•      Q4–Attitude: I like these infographics.
•                            I’m satisfied with the contents of these infographics.
•                            I’m satisfied with the designs of these infographics.

Upon reliability analysis, Cronbach’s alpha was at .863, which led us to judge it was suitable for measuring the attitude towards the infographics.

•      Q5–Cognitive effects: These infographics are easy to remember.
•                                          These infographics have clear messages.
•                                          These infographics are eye-catching.

Upon reliability analysis, Cronbach’s alpha was at .851, which led us to judge it was suitable for measuring the perceived cognitive effects of the infographics.

•      Q6–Acceptance intention: I’d like to show these infographics to others.
•                                                 I’d like to share these infographics on social media.
•                                                 I’d like to download or save these infographics.

Upon reliability analysis, Cronbach’s alpha was at .907, which led us to judge it was suitable for measuring the acceptance intention for the infographics.

In order to find out how different types of expression in interactive infographics affect the value assessment of information, we conducted an analysis of variance (ANOVA) from the survey results of three groups—functional interactions, emotional interactions, and no interaction.

For the no interaction group, there is no change when you hover over a human-shaped bar chart. Functional interactions provide additional information to the tooltip when the mouse is over, and the selection information is highlighted in blue. In emotional interactions, we see change in the appearance and behavior of a human-shaped bar graph when the mouse is over (see Figure 3). As shown in Table 7, the entertainment element and the type of infographic interaction showed statistical significance (F = 40.764, df = 2. 249, p < 0.05). The entertainment element in each group was in the order of L3 > L2 > L1, but through post analysis, we established that all levels show distinct meaningful differences. Thus, “H1-1: Expression type of interactive infographics will exert positive influence on their entertainment” has been supported (see Table 7).

Results of the analysis indicate that the relationship between interaction types and reliability had no positive effect (F = 3.536, df = 2. 249, p < 0.05). The P value was 0.031, but the f value did not reach 4, and there was no statistically significant relationship between the groups in the post analysis interactions. Thus, “H1-2: Expression type of interactive infographics will exert positive influence on their reliability” has been rejected (see Table 8).

The types of interaction and informativeness showed statistical significance (F = 4.942, df = 2, 249, p < 0.05). Informativeness for each group were in the order of L2 > L3 > L1, but post analysis showed a distinct significant difference between L1 and L2. Thus, “H1-3: Expression type of interactive infographics will exert positive influence on their informativeness” has been supported (see Table 9).

As shown in Table 10, overall, “Reliability,” “Entertainment,” and “Informativeness” each influence “Attitude” (F= 212.165, df = 3, 248, p < 0.05). After verifying significance levels of individual variables, we have found “Entertainment” to have the most influence on “Attitude,” which leads us to accurately predict positive attitude when the entertainment value is high (β = 0.596). “Reliability” exerts considerable influence on “Attitude” (β = 0.353). In other words, reliable infographics tend to draw a more positive attitude. “Informativeness” also influenced “Attitude,” but its influence was weaker than that of the other two variables (β = 0.096).

As shown in Table 11, overall, “Reliability” and “Entertainment” influences “Perceived Cognitive Effects” (F= 114.652, df = 3, 248, p < 0.05). After verifying significance level of individual variables, we have found “Perceived Cognitive Effects” to change according to “Reliability” and “Entertainment.” However, “Informativeness” does not influence “Perceived Cognitive Effects” (β = 0.003, p > 0.05).

Table 12 shows that “Entertainment” and “Informativeness” influence “Acceptance Intention” (F = 78.196, df = 3, 248, p < 0.05). After verifying the significance level of individual variables, “Entertainment” and “Informativeness” have considerable influence (Entertainment: β = 0.398; Informativeness: β = 0.345). “Reliability” showed a low correlation with “Acceptance Intention” (β = 0.095, p < 0.05).