- Open Access
Design preferences and cognitive styles: experimentation by automated website synthesis
© Leung et al; licensee BioMed Central Ltd. 2012
- Received: 28 July 2010
- Accepted: 25 June 2012
- Published: 25 June 2012
This article aims to demonstrate computational synthesis of Web-based experiments in undertaking experimentation on relationships among the participants' design preference, rationale, and cognitive test performance. The exemplified experiments were computationally synthesised, including the websites as materials, experiment protocols as methods, and cognitive tests as protocol modules. This work also exemplifies the use of a website synthesiser as an essential instrument enabling the participants to explore different possible designs, which were generated on the fly, before selection of preferred designs.
The participants were given interactive tree and table generators so that they could explore some different ways of presenting causality information in tables and trees as the visualisation formats. The participants gave their preference ratings for the available designs, as well as their rationale (criteria) for their design decisions. The participants were also asked to take four cognitive tests, which focus on the aspects of visualisation and analogy-making. The relationships among preference ratings, rationale, and the results of cognitive tests were analysed by conservative non-parametric statistics including Wilcoxon test, Krustal-Wallis test, and Kendall correlation.
In the test, 41 of the total 64 participants preferred graphical (tree-form) to tabular presentation. Despite the popular preference for graphical presentation, the given tabular presentation was generally rated to be easier than graphical presentation to interpret, especially by those who were scored lower in the visualization and analogy-making tests.
This piece of evidence helps generate a hypothesis that design preferences are related to specific cognitive abilities. Without the use of computational synthesis, the experiment setup and scientific results would be impractical to obtain.
- Cognitive Test
- Causality Information
- Design Preference
- Specific Cognitive Ability
- Visualisation Test
The manner of external representation (or presentation) could affect our way of working with the internal representation (mentally) and our understanding of the information , e.g. in cockpit information displays for aviation , but few results on graphical external representation can be generalised .
One problem affecting all websites is that there is no reliable, general and abstract method for predicting the effect of presentation rhetorics and modality on understanding of the information. To improve knowledge communication, we should investigate how sensitive people might be to differences in the way we construct our websites. It would be useful to conduct experiments quickly to compare different models of interpretation of information of a specific domain in particular cases. There may be certain styles of presentation or navigation that are generally demanded by users and can either hinder or support users' ability to interpret the information.
Tabular and graphical representations are common in constructing visual arguments  and presenting relational data (especially quantitative data) . Visualisation of aviation accident events generally use causal trees to represent the causal relations but there are few empirical studies on both preference and perception of causality visualisation. Specifically, we investigate users' preferences for information visualisation styles and their perception of causality as required by aviation accident reporting. As the Web is one of the main channels for publishing information of aviation accidents, it is desirable to know about how people would prefer the causal relations in accident events to be presented in a website and how they perceive this causality. The user preference data are useful in the design and re-design of websites. To elicit preferences from people, we provide multiple designs for selection and study the rationale of their design decisions. Automated website synthesis saves time and effort in building websites for such designs. Few models and theories are available to address computational website design. However, if we view websites as a form of information visualisation, we can borrow some findings from automated diagram design  to serve as our experiment hypotheses. Some systems for automated diagram design have incorporated text to enhance user understanding of graphical visualisation . This kind of multimodal visualisation should be applicable to website design. Expressiveness and effectiveness of graphical languages as proposed by Mackinlay  have been influential to diagram visualisation models, including source information characteristics , user-defined task specification , and user-defined layout preferences  for automated diagram design.
In general, we would like to see if people would prefer different representations to display the same information. In particular, this study aims to elicit preferences of designers (and users) about visualisation patterns, particularly the preferences for tables and trees in visualising causality information. In this case, we select trees and tables as the options for selection by users. Tree representations are commonly used to graphically represent causality in printed documents. The causal relations are normally represented by arrows or lines connecting causes and effects.
If people do prefer a representation, it would be interesting to see what rationale or criteria contribute to their preferences. We categorised common rationale/criteria mentioned in website design textbooks [11, 12]:
easy to learn: the users do not need much time and effort to understand how it works;
more visual: the users can understand through graphical illustrations;
more informative: the users can know more details;
more scalable: fewer changes are needed to handle more massive information;
more features represented: less characteristics (important information) are left out;
more suggestive: the users can understand without much guessing; and
more flexible: suitable for use in different situations.
As these seven rationales may not cover all possible rationales that are crucial to any particular preference, the experiment participants were asked (unprompted) for their rationale before seeing these seven rationales and then they were asked (prompted) to identify if any of these rationales were similar to their own rationales. They were also asked if any of their rationales was not covered by these seven rationales.
Software designs should reduce users' cognitive load . We hypothesise that participants prefer one design to other designs partially because the preferred design suits their cognitive abilities. If this is true, the cognitive abilities of the participants should be related to their preferences. The relationships among preferences of trees or tables, the cognitive test results of the participants, and rationale for their preferences were studied in this experiment. As it is impossible to test numerous cognitive factors in a single experiment, the participants were only tested on cognitive styles/abilities of visualisation and analogy-making, which we guessed were related to visual representations.
The main objectives of this experiment are as follows:
To see if there is any different preference for tables or trees in representing the given information;
To see if different preferences are based on different priority in criteria/rationale;
To see if the preferences are related to the cognitive test results; and
To see if the importance ratings of design criteria/rationale are related to cognitive test results.
Sixty four students from the University of Edinburgh participated in the experiment and received cash (GBP 10) as a reward. They were randomly assigned to one of the two groups according to a pre-generated random sequence. Each group had 32 participants. The treatments of these two groups differ in the order of using table and tree design generators (or simply called designers). All of the participants had the computer skills for browsing websites. The experiment took about 1.5 hours for each participant. No strict time limit was enforced for tasks except cognitive tests, for which data were automatically collected.
Computational website synthesis provided basic facilities for generating websites and their functional components such as menus and breadcrumbs. We mapped the information content items to appropriate components in specifications. Presentation of accident event (content) information in visualisation formats (tabular cells or graphical trees) required mappings of attributes between the content information and visualisation formats. One of the major factors affecting the mapping decision was designers' (and users') preferences.
All input from the participants were collected by standard HTML forms and CGI (common gateway interface) scripts, which were generated from simpler specifications for defining variables and variable types (e.g. multi-answers or long text) and special webpage elements. Subsequent minor modifications to the generated questionnaire forms were only for cosmetic purposes.
Tree and table generators
The trees and tables were generated on-the-fly and viewable in a separate window of the web browser together with other basic website navigation facilities, e.g. menus and hyperlinks. The tables were generated as standard HTML tables while the trees were generated as DOT diagram specifications for final image rendering by GraphViz on the server side before sending to the client side.
Each participant filled in a pre-experiment questionnaire which collected background information such as their familiarity with the Web and aviation operations. Then the participants used table and tree designers to express their preferences between tables and trees in representing a given structure of causality information. The order of using table and tree designers were randomly assigned. Participants assigned to Group A used the tree designer first and then the table designer while those in Group B used the table designer first and then the tree designer. In the designers, the participants used a customised drag-and-drop facility to design their preferred table and tree patterns. The participants submitted their preferred visualisation patterns as tables or trees. Their rationale for their preferred visualisation pattern were collected by a two-part post-experiment questionnaire. Part A of the questionnaire collected their rationale as the open-ended answers and their preference ratings of table and tree visualisation. Part B of the questionnaire collected their ratings of the importance against seven common preference criteria/rationale. The participants also took timed cognitive tests, including two (paper folding and surface development) visualisation tests and two (visual and verbal) analogy-making tests. The visualisation tests were licensed from the Educational Testing Services of the USA, as provided in a kit of factor-referenced cognitive tests . The visual analogy puzzles were the same as those selected by Thomas Evans  in his study of visual analogy problems. The verbal analogy test was the sample questions of the Miller Analogy Test (MAT).
Variables included (1) preferences of tables or trees, (2) importance ratings of preference criteria/rationale, and (3) results of cognitive tests. Data were presented in medians, means, and standard deviations (SD) for both parametric and non-parametric analyses, although only conservative non-parametric Wilcoxon test, Krustal-Wallis test, and Kendall rank correlation test were reported in this article. Both parametric (e.g. t-test and ANOVA) and non-parametric tests were run using R statistical software  and its μStat package. In multiple comparison, P values was adjusted by Bonferroni correction. P values less than 0.05 were considered statistically significant. In statistical tables, single asterisks (*) indicated P < 0.05 and double asterisks (**) indicated P < 0.01.
Background of participants
There was no significant difference in any background variable between two groups of participants as measured by the pre-experiment questionnaire and analysed by Student's t-test and Wilcoxon test.
Preferences were obtained from the preference ratings about participants' preferences between tables and trees as the representation of the given information. The participants rated the strength of their preferences as:
strong table preference,
moderate table preference,
marginal table preference,
marginal tree preference,
moderate tree preference, and
strong tree preference.
Preference of table and tree representations
No. of participants
Rationales for preferences
Easier to learn
Results of cognitive tests
Preferences and rationales
Relationship between preferences and rationales
Easier to learn
Ranking of rationales
Easier to learn
Preferences and cognitive tests
Relationship between cognitive test results and preferences
Comparison of cognitive test results between different groups of preferrers
Rationale and cognitive tests
Correlation between rationales and cognitive test results
Easier to learn
statistically significant rank correlations coefficient (τ) ranged between around 0.201 - 0.263, which are only low to moderate in strength.
This study used website synthesis to construct an experimental apparatus for Lab-on-the-Web. Using this approach we found a significant relationship among the participants' preferences, rationale, and cognitive ability test results. Most (41 out of 64) participants claimed themselves to be tree preferrers. The other participants (23 out of 64) said they preferred tables. Both table preferrers and tree preferrers found their preferred representations (tables or trees) more informative, without significant difference in rankings of this rationale. As the task given to the participants is to represent information, it is not surprising that the rationale "more informative" was one of the most important rationale. In further studies, it would be interesting to test whether the participants find the same rationale justifiable for their preferences when they are given different goals or under different conditions. This might give us more insights into how different preferrers perceive information.
To table preferrers, the rationale "easier to learn" was more important than "more informative". The rationale "easier to learn" was ranked as the most important by 13 out of 23 table preferrers, but not so important by tree preferrers. Tree preferrers ranked the rationale "more suggestive" significantly higher than the table preferrers did. To table preferrers, tables seemed to be easier to learn than trees. To tree preferrers, trees were more suggestive than tables. These discrepancies in rationale ranking indicate a perception difference between different preferrers in perceiving the given tables and trees.
Tree preferrers performed better than table preferrers in some tests of cognitive factors, particularly the paper folding visualisation test and visual analogy test. There was no significant difference between table preferrers and tree preferrers in their performance in other cognitive tests including the surface development visualisation test and the verbal analogy test. It is plausible that the interpretation of tree representations requires specific cognitive capabilities such as visualisation and visual analogy-making; thus, those who do not feel comfortable with these tasks would prefer tables and highly rank the rationale "easier to learn" for their table preference.
Among all participants, there is a statistically significant low-to-moderate rank correlation between the rankings of the rationale "easier to learn" and the results of visual analogy test and paper folding visualisation test. The low-to-moderate rank correlation indicates that the participants who performed better in such two cognitive tests ranked the rationale "easier to learn" to be less important. At similar correlation strength, the participants who performed better in visual analogy test ranked the rationale "more suggestive" to be more important. It seems that the visualisation and visual analogy-making abilities of the participants might play a role in their preferences for tables and trees. Possibly (although we cannot prove this), table preferrers lack sufficient cognitive capability to interpret graphical representations like trees; thus, they prefer tables as they are easier to learn. Tree preferrers would feel more comfortable in making visual analogy and find graphical representations like trees suggestive. As indicated by the low to moderate strengths of correlation in these results, it is probable that other factors (and other cognitive factors) may be also relevant to the participants' preferences. Further studies are required to delineate these relationships.
This experiment also demonstrated the technological significance of computational synthesis in enabling scientific experiments. This experiment re-used a website synthesiser to generate websites on the fly so that the designers/users can explore the design space. Without using a website synthesiser, this experiment would not be possible. Other than computational synthesis, the most relevant tool is Web content management system but it does not meet the requirement of this experiment. Apart from its high cost, there is no Web content management system so flexible as our our website synthesiser in accepting information mappings. This makes Web content management systems inapplicable to our experiment. Thus, computational synthesis is the only available solution for this experiment although computational synthesis following definite patterns would limit the variability of graphical presentation. To enrich the variability and functionality of graphical presentation, we would consider using web design patterns available in Web 2.0 and HTML5 to enhance further studies.
The experiment reported in this article found significant relationships between the participants' design preference, rationale, and cognitive test performance. This work also exemplifies the use of a website synthesiser as an essential instrument enabling the participants to explore different possible designs, which were generated on the fly, before they selected their preferred designs. In the tested sample, more people prefer graphical to tabular presentation. Despite the high preference for graphical presentation, the given tabular presentation was generally rated to be easier than graphical presentation to interpret, especially for those who score below average in the visualisation and analogy-making tests. This piece of evidence helps generate a hypothesis that design preferences are related to specific cognitive abilities. Without the use of computational synthesis, the experiment setup and scientific results would be impractical to obtain.
This study was partially supported by the UK EPSRC grant GR/M98302 for informatics research into communicating knowledge about accidents from synthesised websites.
- Zhang J: The nature of external representations in problem solving. Cognitive Science. 1997, 21 (2): 179-217. 10.1207/s15516709cog2102_3.View ArticleGoogle Scholar
- Zhang J: Distributed representations as a principle for the analysis of cockpit information displays. The International Journal of Aviation Psychology. 1997, 7 (2): 105-121. 10.1207/s15327108ijap0702_1.View ArticlePubMedGoogle Scholar
- Scaife M, Rogers Y: External cognition: how do graphical representations work?. International Journal of Human-Computer Studies. 1996, 45: 185-213. 10.1006/ijhc.1996.0048.View ArticleGoogle Scholar
- Oestermeier U, Hesse F: Verbal and visual causal arguments. Cognition. 2000, 75: 65-104. 10.1016/S0010-0277(00)00060-3.View ArticlePubMedGoogle Scholar
- Zhang J: A representational analysis of relational information displays. International Journal of Human-Computer Studies. 1996, 45: 59-74. 10.1006/ijhc.1996.0042.View ArticleGoogle Scholar
- Kamps T: Diagram Design. 1999, Springer-VerlagView ArticleGoogle Scholar
- Mittal V, Roth S, Moore J, Mattis J, Carenini G: Generating explanatory captions for information graphics. Proceeedings of the International Joint Conference on Artificial Intelligence. 1996, Montreal, CanadaGoogle Scholar
- Mackinlay J: Automatic Design of Graphical Presentation. Phd thesis. 1996, Computer Science Department, Stanford UniversityGoogle Scholar
- Roth S, Mattis J: Data characterization for graphic presentation. Proceedings of the onference on Human Factors in Computing Systems (CHI). 1990Google Scholar
- Casner S: A task-analytic approach to the automated design of graphic presentations. CM Transactions on Graphics. 1991, 10 (2): 111-151. 10.1145/108360.108361.View ArticleGoogle Scholar
- Krug S: Don't Make Me Think: A Common Sense Approach to Web Usability. 2000, New RidersGoogle Scholar
- Nielsen J: Designing Web Usability. 2000, New RidersGoogle Scholar
- Detienne F: Software Design - Cognitive Aspects. 2002, Springer-VerlagView ArticleGoogle Scholar
- Ekstrom R, French J, Harmon H: Manual for kit of factor-referenced cognitive tests. 1976, Education Testing ServiceGoogle Scholar
- Evans T: A heuristic program to solve geometric-analogy problems. Readings in computer vision. 1987, Morgan KaufmannGoogle Scholar
- R Development Core Team: R: A Language and Environment for Statistical Computing. Tech rep, R Foundation for Statistical Computing. 2012, [http://www.r-project.org]Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.