I removed one participant from the cluster analysis who sorted allthe cards, but in such a way as to skew the results. With a sharp wit, this participantcreated only two categories: “Byproducts of nuclear decay,” which containedonly one card,radiation; and “What an academic thinks of when he or shemisinterprets design as the synthesis of a set of design principles,” which con-tained all of the remaining cards. I did, however, include this participant’ssurvey responses, as they reflected an attitude toward design that favored artisticsensibility over using more formal methods such as design principles or designresearch. I address this attitude more fully in the conclusion to this article

However, I was confident that card sorting techniques would provide agood opportunity for exploring perceived relationships between common designprinciples. This methodology also allowed me to stay open to the possibilitythat design principles are so idiosyncratic that no meaningful structure or rela-tionships might emerg

I conducted an online card sorting exercise. Cardsorting is a flexible technique for discovering how people group items. Designersand usability experts often use this technique to help determine informationarchitecture, such as for website navigation orproduct list arrangeme

Having determined which design principles are mentioned most commonlyin literature on design, the next question is, how do these design principlesrelate to one another

This quantitative review produced a raw list of 198 design principles. Thisnumber might seem discouragingly high except for two outlying works: Leborg[54], which lists 41 principles, 33 of which are unique, and Lidwell et al. [58],which lists 100 principles, 87 of which are unique. Lidwell et al.’s broader focuson “universal principles of design” rather than visual design principles meansthat many of their principles fall outside of the scope of my study. (Some evenstretch the concept of design principle—for example, “uncertainty principle”and “normal distribution.”) If we were to exclude Leborg’s and Lidwell et al.’sunique principles, the list of design principles contracts from 198 to 77—stillhigh, but more reasonable. However, because I included unique principles fromother texts, I retained all of Leborg’s and Lidwell et al.’s principles in the raw list.As might be expected, some of these principles are considerably more commonthan most. Overall, of the 198 principles in the raw list, 160 were mentioned inonly one work. The 198 design principles were listed cumulatively 420 timesin the 46 texts; 61.9% of these listings referred to principles used in at leasttwo works. Despite the large number of unique principles, they were used onlyin 38.1% of texts.While the unique principles may have some heuristic value, they are clearlynot as widespread in the lore of design as other principles, so I excluded themfrom the card sorting exercise. Doing so produced a final list of 38 visual designprinciples referred to in at least two of the 46 works (see Figure 1). Figure 1visualizes the frequencies of the most common principles in a Pareto chart,excluding the long tail of individual mention

The primary advantage of this online approach was quick and efficient accessto participants. However, there were two disadvantages. First, in a face-to-facecard sorting exercise, the researcher often observes the participant as he or shesorts the cards and uses the talk-aloud protocol to gather further informationabout participants’ thinking during the exercise. Because I would not attend thecard sorting, this kind of observation was not possible. Second, I would havelittle control over who agreed to participate. The possibility for spam is alwayspresent with online media

Traditionally, in a card sorting exercise the researcher writes items on standard3” × 5” index cards and asks participants to sort the cards into categories. Theresearcher then records the groupings and category labels and compares themover multiple participants. To expand the reach of my study, I used an onlinecard sorting tool, Optimal Sort (http://www.optimalworkshop.com). Instead ofindex cards, this online tool displays digital shapes that participants can dragand drop on screen to create groups, which they can then label (see Figure 2).This tool also allowed participants to provide feedback through before and aftersurvey questions

Card sorting is typically conducted in one of two ways: closed sort or opensort. In a closed sort, participants sort the cards into predefined categories.In an open sort, participants sort the cards into categories they create themselves.Due to the exploratory nature of my study, I used an open sort so designerscould group the design principles as they saw fit. I also did not require partici-pants to sort all of the cards, but restricted my analysis to those who sorted atleast 75% of the cards

So at best, design principles are a kind of lore. Lore is a kind of contingentknowledge based in practice, and as North has argued for composition studies, ithas a value that is often overlooked. However, lore being what it is—anecdotal,implicit, and often idiosyncratic—I think there is good justification to bringresearch to bear on it. Accordingly, in this article I apply empirical techniques tounderstand the lore of design principles more fully

None of these questions have straightforward answers. Although many authorsrefer to design principles, hardly any define what they mean by the term. A few,such as Williams and Stimatz [8] or Blair-Early and Zender [9] generally refer todesign principles as “rules of thumb”; as such, principles are heuristic methodsthat help us make design decisions quickly. But these rules of thumb can be vari-able and ambiguous, because they are based on individual designers’experienceand on the advice they absorb from their training and reading. Depending on thesituation, designers may employ different design principles at different levelsof consistency. Designers, in fact, often talk about the need to “break the rules”on occasion to bring attention or interest to a design. Some designers even denyusing design principles at all, arguing that they rely on instinct, artistic sensi-bilities, or a “good eye.” This ambiguity leads Johnson to comment that th