Archives of Design Research

From loose thumbnail sketches to crisp renderings, digital CAD models to high-fidelity working prototypes (Cross, 2007; G. a. S. GOLDSCHMIDT, M, 2006; Pei, Evans, & Campbell, 2011; Visser, 2009), design representations are used for a wide variety of purposes, as means to support reflective activity (D. A. Schon & Wiggins, 1992), and as a way to communication design intentions to others (Cross, 2008; Self, Lee, & Bang, 2013). Considering its various and critical role in support of designerly work, studying design representation provide researchers, practitioners and educators an opportunity to extend their understanding of the nature of design practice and designerly ways of knowing and thinking (Cross, 2007; G Goldschmidt & Porter, 2004). For example. Pei et al. (2008) developed an extensive taxonomic classification of design representations as a means to clarify their role and to enhance communication between designers and engineers during new product development. An extension to Pei et al. (ibid) taxonomy has been proposed by Kim et al. (2013), who indicate the classification’s limitations in terms of conceptual design representations. Considering the various types of representation used in the practice of design, perhaps the most prolific, and most often associated with designerly activity, is the designer’s sketch (Bar- Eli, 2013; Goel, 1995). Due to its historic position as a critical non-verbal media through which design progresses and is communicated, a substantive body of knowledge now exists to extend an understanding of this most prolific and, arguably, most critical conceptual design tool.

In a previous study, Cross (1999) investigated the nature of design problems through examining the act of sketching as the means by which problems are explored. In a seminal work, Goel (1995) investigates design representation through a comparison of digital and conventional sketching to suggest important cognitive insights that take place during the conceptual design activity. It is also important to recognize that the ambiguous nature of sketching, combined with the semantic density and ability to provide opportunities for transformations between and among ideas are presented as distinctive characteristics during conceptual design. Using a notation system underpinned by Goel’s (ibid) design transformations construct. Do et al. (2000) interpret the designer’s thinking by investigating drawings to provide a greater understanding of the relationship between representation through drawing and its association with design practice. In further examples, McGown et al (1998) examine how computer-supported sketching tools can potentially improve design representation. However, Bilda et al (2006) employ the use of protocol analysis to suggest that there is no significant difference between the act of sketching and not sketching based on the design outcomes and the associated cognitive activities. In contrast, Bouchard et al’s (2006) study indicates the importance of sketching as a tool for solution-focused, externalisation of thought through the use of design representation, reflection and the further iteration of design intent. Goldschmidt, in an influential series of design experiments (G GOLDSCHMIDT, 1994, 1997; Gabriela Goldschmidt & Weil, 1998) discusses sketching’s importance as a means to facilitate visual thinking. They define the act of sketching as a ‘clue’ finding activity and the designer’s germane of their own visual thinking ‘on the fly’ and ‘in-situe’. Taking a step further, Menezes et al. (2006) explore the differences between novice and expert designers’ perception of conceptual sketching activity to indicate the kinds of knowledge and thought engaged during the reinterpretation of sketches during the ideation process. More recent studies aim to extend knowledge related to conceptual design activity and its application, such as the development of computer-aided conceptual design tools (Aldoy & Evans, 2011; Contero, Naya, Company, Saorin, & Conesa, 2005; Dorta, Pérez, & Lesage, 2008; Plimmer & Apperley, 2002).

However, a review of the literature has found that there is a disagreement among scholars on the extent to which sketching as tool for design representation influences conceptualisation of design intent, both in the framing of ill-defined design problems and the solution-focused generation and progression of design ideas.

The current study’s examination of the role and use of sketching during conceptual design employs the discursive epistemology of ways-of-knowing in practice first proposed by Schön (1983, 1987), and Wiggen and Schön (1983, 1987; 1992) and extended as a means to investigate design practice by Dorst and Dijkhuis (1995) and Valkenburg and Dorst (1998).

A Schösian account of design practice provides an alternative to a more rational, scientific approach to understanding designerly activity (J. S. Gero & Mc Neill, 1998; Simon, 1996). Schön himself (op cit) discusses the rational problem solving paradigm’s limitations in an inability to capture professional practice as it is experienced. Instead proposing the alternative of an epistemology of reflection-in-action, positioned as a means to explore the kinds of experiential knowledge deployed in an artistry of professional practices. Reflection-in-action, then, may be best described as a constructed account of designerly activity (Dorst and Dijkhuis, 1995). Design is seen as a reflective conversation with the situation (Lawson, 2004; D. A. Schon & Wiggins, 1992) where unique problems in unique situations are framed and explored by the designer or design team (Nelson & Stolterman, 2012; Stolterman, 2008). Moves are made to test and reframe the design problem through a solution-focused, constructed reflection-upon-activity, often supported by sketching (Self et al., 2014). As a result of this reflective-practice, in response to often ill-defined design problems (Rittel & Webber, 1973), design work develops and is progressed.

Adopting a similar conceptual lens to existing design practice research (Bar- Eli, 2013; G. a. S. GOLDSCHMIDT, M, 2006; D. A. Schon & Wiggins, 1992; Valkenburg & Dorst, 1998) we employ reflection-in-action as the paradigm through which an analysis of design activity through sketching proceeds, and provide evidence to indicate sketching’s influence when design practice research is based upon this constructionalist paradigm. The fact that our results may contradict those of others (Bilda et al., 2006; Jonson, 2005; Tang et al., 2011) provides further grounds for debate on the ways in which different approaches to the analysis of design practice may influence results.

This paper reports upon the conceptual design activity of a participant from a population sample of 4th year undergraduate industrial design students at the researchers’ institutions. The criteria for participant selection were as follows:

 • Successfully completed design fundamental courses and sketching skills courses

 • Enrolled as a full-time student in a 4 year ID degree program

Similarly to other seminal works examining design practice (D. Schon, 1983; D. A. Schon & Wiggins, 1992), and taking a depth first, fine-grained approach to the analysis of protocol data, here we present and discuss the design activity of a single participant taken from the sample population.

In doing, we examine and assess the ways in which design representation through sketching may provide the necessary context for problem-framing and act as catalyst for appositional, solution-focused conceptual design and development.

The participant’s design knowledge and skills may be described as competent according to the Dreyfus and Dreyfus (1980) model of skills acquisition (Table 1). The Dreyfus and Dreyfus model (ibid) derives stages of acquisition from the amount of experience an individual has had in practice. When applied to the criteria used in the selection of the participant, the model indicates the level of design competency. While not yet expert, the participant designer possesses a working knowledge of design practice, is able to cope with the complexities of design problems (Cross, 2007; Dorst, 1996; Rittel & Webber, 1973), and approaches design tasks with longer-term goals in mind (Dreyfus and Dreyfus op cit).

The participant’s design activity and think-aloud responses to a given design problem were recorded through 3 video recorders as shown in Figure 1.

Figure 1 Study’s task environment

The first camera (CAM.01) recorded an overview of the environment, the second camera (CAM.02), overlooked the participant, recorded the activity performed within the subject’s working environment (sketching, drawing, writing, reading). A third camera (CAM.03) was positioned to record the participant’s movements and body posture. The subject was provided with plain drawing paper, pens, pencils, coloured markers, shape templates and other analog tools commonly used as media during sketching activities.

The participant was provided with the task of developing a concept for a sports watch (Appendix A). Due to limitations of time and the nature of the protocol analysis as research method, a moderate level of difficulty was sort in setting the design problem. As such, design specifications documents used in similar published protocol studies (Cross, Christiaans, & Dorst, 1994; Prats, Lim, Jowers, Garner, & Chase, 2009; Tang et al., 2011) were employed as benchmakes for the design brief’s structure, composition and content. During the experiment, we identified the limitations of this approach, both in its inability to account for the rich context within which design practice is actually performed (Jonson, 2005), and facility to provide the designer with opportunities to develop understanding of the design problem through research and analysis such as leaving his “creative” environment to undertake user experience studies. As such, we acknowledge that such an experimental approach to understand the phenomena of design practice may in fact decontextualise what is in reality a highly collaborative and dynamically context-driven process as highlighted by Stolterman (2008). In defence of our methods, we suggest that conducting such an experimental methodology has its own merits such as providing a uniformly controlled environment for study, enabling a more focused understanding of the key aspects of the phenomena and also reducing unwanted noise that is often associated with “in-the-wild” design research (Michel, 2007). Given our aims of a fine-grained examination of the role and use of sketching during conceptual design practice; its influence upon problem-framing and solution-focused design work, our study required the inclusion of equipment to gather controlled response data. In this regard, our approach mirrors the methodological approaches employed by Cross et al’s Delft protocol workshops (1996) and Goel’s (1995) analysis of sketching’s association with cognition.

The participant was given an introduction related to the aims and objectives of this research, their rights and obligations as a participant, the task procedure, methods of recording and time given to complete the task, after which the design problem was provided. An assistant researcher than withdrew from the environment, leaving the subject with 25 minutes to respond to the design problem. 20 minutes into the task, the participant was told 5 minutes remained. Immediately after the task, the subject’s recorded sketching activity was played back on a screen as part of a retrospective think-aloud session. During this session, the participant was asked to tell the researcher what they were thinking as they engaged the design task (Someren et al., 1994).

The think-aloud method and protocol analysis are established technique for design practice research data collection (Chai & Xiao, 2012). However, the types of think-aloud experiments and variations in their format differ within the literature, from concurrent think-aloud to retrospective studies (Perry & Krippendorff, 2013) where the participants’ design activity is played back to them after task completion. The length of time provided for a think aloud session can vary from a few hours to as little as fifteen minutes (Chai & Xiao, 2012; Jiang, 2009). As Perry and Krippendorff’s (op cit) study suggests, concurrent think-aloud may have implications for the subject’s ability to perform the design task as it has been found to be disruptive and may potentially affect the thinking process. As such, it was decided that the retrospective approach would be more suitable for this study.

The transcribed protocol of the design task was encoded through a qualitative content analysis (QCA, (Miles, Huberman, & Johnny, 2013; Schreier, 2012)). The transcribed think-aloud session was segmented using a thematic criterion where two researchers, working as coders separately looked for phrases, sentences and discourse that appeared to hang together as a single idea, action or thought. The importance of having two coders working separately provided the opportunity to check the segmentation of the protocol. Following segmentation, differences in the length and division of segments were discussed and decision rules applied (Schreier, 2012). This thereby limited the inherent subjectivity required in segmentation during QCA.

Following segmentation, coders assigned units of coding (segmented verbatim) to the dimensions of a coding frame adapted from Valkenburg and Dorst’s (1998) encoding classification system (Figure 2).

Figure 2 Five conceptual categories of coding frame (Valkenburg & Dorst, 1998)

The Valkenburg and Dorst’s model (ibid, Figure 3) provides four theoretical constructs through which the segmented protocol was encoded. The four concepts (naming, framing, moving and reflecting) originate from Schon’s (D. Schon, 1983; D. A. Schon & Wiggins, 1992) reflection-in-action epistemology of professional practice, with various studies having since extended their meaning (Bar-Eli, 2013; Dorst & Dijkhuis, 1995; J. Gero & Kannengiesser, 2008; Tang et al., 2011). A summative description of each concept is provided in Table 2.

Subsequent to the encoding of the protocol, coding was compared for consistency and inter-coder reliability. To achieve this, the same two coders worked seperatly in the encoding of the segmented protocol. Encoding was then compared for inconsistancies in the application of the coding frame when encoding of protocol segments (Schreier, 2012). Where inconsistancies were found, decision rules were then applied and segments recoded.

Although there will always be an element of subjectivite interpretation required in any QCA (Miles et al., 2013), through a process of encoding, comparison and revision, consideration for the reliability of the coding frame (its ability to classify the segmented protocol transcriptions) and validity (its ability to describe the participant’s design activity while sketching) was achieved.

Figure 3 illustrates subject A’s design activity from the beginning to the protocol’s end. The encoding of protocol segments as naming, moving and reflecting (see Table 2) are represented as red, yellow and blue respectively.

Framing events are illustrated as rectangular boxes cutting across these three. In the left most column time stamps indicate the timing of each activity. To the right of this, notations provide an indication of what the subject was actually doing at each encoded segment. Absolute frequency counts (f) of encoded at naming (f=35) moving (f=44) and reflecting (f=25) indicate the amount of time spent on each activity.

From the flow-illustration we see that subject A commences the design task with a series of naming events (C/1:24-3:00, Figure 3). That is, she names the things she feels she must pay attention to within her task environment, ‘I have to consider all these things’ ([C/0:09min] N: concern over specifications, Figure 3)

Figure 3 Diagrammatic flow-illustration of subject A’s design activity

After an extended period of naming, subject A starts to sketch, whereupon her sketching moves appear to act as stimuli to further developing understanding of the strap design, ‘It will be a strap, kind of a strap. What could be the best for this design, proper for this design?’ (C/3:08, M: considering strap design, Figure 3). At this point the idea of a clasp design emerges as a possible partial-solution frame to address the sub-problem of strap design (C/3:49, F: clipping watch clasp). Subject A then engages in sustained sketching activity to test the frame’s suitability in light of her own emerging understanding of the design problem. Figure 4 illustrates subject A’s design work by the end of the protocol’s 4th minute.

Figure 4 Working environment, 4th minute of protocol

Following a further period of exploration and reflection (C.4:12, M: how to fit strap to wrist , C/4:15, R: importance of fix method for design, Figure 3), Subject A abruptly writes in capital letters, ‘Materials’ before enclosing the word within a dark boarder to signify its importance as something requiring her attention. She then proceeds to write, ‘rubber’ underneath, and reflects upon its suitability in light of her developing understanding of the design problem, ‘I think maybe rubber can be the best material for this design’ (C/4:48 R: on which material: rubber). Following this, subject A’s attention shifts as she draws a large rectangle at the top of her page and writes within it the digits, ‘9:00’ , ‘Err, this was kind of the display, digital or analog method’ (C/5:00 M: drawing different display). Following this, subject A continues to name design requirements, before proceeding to explore options for the design of the display ([C/5:27 N: importance of battery life; C/5:34 R: how to reduce weight; C/5.43 R: attractiveness of display).

At this point it appears that subject A’s attention towards important considerations within a future design proposition provide the necessary conditions for the frame of a new user interface to emerge, ‘So I started to think about a new UI maybe’ (C/5:58 F: new UI design, Figure 3). After the identification of this new frame of reference, subject A appears to immediately explore its suitability as a possible solution candidate through a period of sketching moves, ‘Err, that is my design, a new UI so that the display above will display the numbers for checking the hours. And for the bottom part for checking the minutes’ (C/6:18 M:UI display, hours/minutes). By the 5th minute of the design task, subject A has generated a sheet of sketch work, notes and schematic drawings (Figure 5).

Figure 5 Working environment, 5th minute of conventional sketching

At this point subject A continues her moves to explore solution ideas through a 3D sketch, notably including an indication of where the watch display interfaces with strap design. From exploring the new UI frame’s suitability to address the design sub-problem of display (C5/5:58 F: new UI design, Figure 3), subject A shifts her attention to the watch’s strap, as indicated by a new, emergent frame of reference, ‘Velcro I see. A sticking method. Yeah. Yeah a fixing method for Velcro.’ (C/6:58, F: Velcro on strap). It appears moves made in the testing of the UI display frame have included an incidental representation of the watch’s strap (Figure 6), prompting the reemergence of a focus upon strap design. This incidental sketch representation appears to have provided the stimuli through which subject A reengages the strap sub-solution, through which the Velcro insight emerges as a new frame worthy of exploration.

Figure 6 Incidental sketch, watch strap, 7th minute

At this point a further frame emerges, ‘because…just I thought a strap-tie would be good’’ (C/8:02 F: Strap-tie design). It appears that by the time subject A’s sketching move is complete her preference for Velcro has been replaced. As before, the strap-tie design appears to have emerged from the testing of the previous frame through sketch representation, which, itself, has been retrieved through an exploration of the UI display frame. That is, upon exploring the suitability of her frame through design orientated representation as sketching, her focus of attention has shifted back to a previous frame of reference. This has then provided a means to further develop the previous frame through reflection upon the current one, supported by subject A’s sketchwork. The act of sketching has provided the context through which previous frames may reemerge, be reevaluated combined and revised. Thus subject A’s sketch activity acts as a stimuli through which both the framing of the design problem, and its exploration, through solution propositions embodied as design-orientated sketch representations, has proceeded.

By the end of the protocol’s 10th minute, subject A continues to test and explore the detail of her solution frames, ‘Yeah, the fixing method. Maybe mainly I’m thinking and considering the…fixing part as a very important issue…’ (C/10:44 M: drawing strap fixing method, Figure 3). Having completing her sketch of the watch-strap fixing, subject A returns to her previous frame of a new UI display. After some hesitation, she frames the display element of her design, ‘so I just decided to use that display.’ (C/11:54 F: Use of line display design), and proceeds to name the importance of clarity in the display of time (C/12:00 N: line displays time), before moving to explore the frame through sketching. However, after reflection upon the sketch representation of her intentions towards the new UI display, subject A abruptly names strap colour as an important consideration, ‘When people are purchasing the watch, they can select their own watch. So they can select this colour, this colour, this plain colour and strap colour’ (C/13:06 N: strap colour as important). Her reflections upon time and the telling of time indicate uncertainty towards the suitability of her UI display concept solution, ‘Most people consider, say about time, they just say about up or down, the numbers of the time. So, just five o’clock or five fifteen. They rather love to speak about time like that’ (C/12:45 R: on line display drawing). It appears her work in the testing and development of her initial line-display frame has resulted in some uncertainty in the evaluation of the solution’s suitability. And that her sketching work appears to be critical in her solution-focused conceptual development and evaluation.

After reflecting upon her sketch activity, subject A’s attention shifts to representing a battery detail, ‘…oh…battery, yeah…The same things, displaying it…and’ (C/15:48 M: drawing, battery display), before exploring the size and form of the design through a second sketch, ‘Um, I just wanted to find the better line that I wanted to draw’ (C:16:17 M: redrawing volume design). At this point subject A’s think-aloud statements point towards concerns over the practicality of her emerging solution concept, concerns she attempts to resolve through further moves to explore ideas supported by her sketching work. Thus she returns to the naming of materials as an important consideration, exploring ideas towards material combinations (C/16:41 M: combining materials, cloth/rubber). It is at this point that subject A’s 5th framing event is recorded as a decision to use rubber for the watch’s band material, ‘So, I think I finally decided the material as rubber.’ (C/17:03] F: rubber as band material). subject A again enters a period of moving activity immediately preceding this frame, sketching a strap concept and fixing mechanism, ‘And, again, I wanted to find a more fancy fixing method…So, just continue to think…’ (C.17:57 M: drawing fixing method, Figure 3).

Figure 7 Exploring fixing mechanism, 17th minute of conventional sketching protocol

It is during these sketching moves that two final problem frames emerge, ‘How about using some, like, pin-shape with rubber, not metal.’ (C/18:23 F: pin fixing). As seen earlier in the protocol, this latest framing event emerges as a direct result of subject A’s exploration through sketch representation. That is, the representation of design intent through sketching acts as a stimuli for an emerging understanding of the design problem through framing, as well as providing the necessary conditions for her solution-focused conceptual development work.

While continuing to explore her current understanding of the fixing method through further sketch activity, subject A’s attention shifts to ergonomic considerations as she writes the word ‘FIT’ next to her sketch, ‘Yeah, I think the most important thing is the fit to the body – it fits very well.’ (C/20:02 N: importance of fit to body). It appears subject A’s thoughts have shifted again from an exploration of strap-fixing to the notion of fit, previously encountered earlier on in the protocol. Her exploration of a currently focused aspect of her conceptual design work appears to provide opportunities to emerge related to previously explored frames. These moves, aimed at exploring solution conjectures under current consideration, act as a kind of catalyst for reengagement with previous frames in new ways. This thus results in shifts of attention between aspects of her design work. As such, the frame of current focus is explored in terms of the notion of fit through a sideelevation of her watch concept (Figure 8).

Figure 8 Exploration of current frame through sketching, 20th minute of conventional sketching

At the end of the 20th minute of the design task, subject A names a difficulty to know what to do next, (C/20:40 N: difficulty to think) and indicates a focus towards her drawing work, ‘so I draw it below this thing. And I keep drawing it. It will be better for it.’ (C/21:07 M: drawing, fit with body).

After further sketching moves to explore the suitability of a flexible plastic material, she rejects the idea to reaffirm fixing method as the frame through which her design work must now proceed, ‘Yeah. It would be better I think, but it doesn’t have the fixing method.’ (C/22:00] N: no fixing methods of flexi plastic). As indicated in the think-aloud protocol, subject A’s current focus upon her sketchwork is less about improving the quality of her sketches, and more a means to consider her next design move, ‘wanting to refresh myself and wanting to think about a new method…’ (C.23:00 M: trying to consider new approach). This is unsuccessful, however, and appears to instead trigger frustration as she names a desire for more information (C/23.23 N: desire for more info.). subject A then continues to shade as an aid for her thinking, ‘So just keep shadowing and thinking and err more ideas I need yeah.’ (C/23:43 M: shading to aid thinking). It is interesting to note the ways in which this apparently, somewhat undirected drawing activity is used as a means to facilitate thinking towards design intent. However, it appears to fail as stimuli, evidenced by subject A’s distraction as she considers starting a new page, ‘No, nothing, just thinking, will it be okay to go over there?’ (C/24:01 M: turning page in sketchbook).

In the final stages of the protocol, subject A is told that 3 minutes of time remain. This appears to change her focus from activity related to the development and testing of design intent to the clearer communication of her existing ideas, ‘So, I’m thinking that I have to choose the colour combination.’ (C/24.33 N: require choice of colours). At this point subject A opens a pack of markers and proceeds to create colour swatches for each (Figure 9), subsequently announcing her choice of colour (C.25:00 N: purple/pink as design colour). These final actions mark the end of the design protocol.

Figure 9 Choosing colour variations, 24th minute of protocol

An analysis of results indicates the ability of the sketching activity itself to provide the necessary conditions for framing and re-framing of the existing design problem. The study also indicated the way in which incidental representation (such as the strap design in Figure 6) made during sketching activity has the potential to prompt the reemergence of previous frames of reference. These Incidental representations appeared to provide the stimuli through which re-engagement with a sub-solution was achieved. That is, according to our qualitative analysis, it appears the activity of sketching provided the context through which previous problem frames re-emerged, were re-evaluated, combined and revised. It appeared that the designer’s sketching acts as a critical stimuli through which the framing and re-framing of the design problem is achieved. The act of sketching thus provides stimulus for the re-emergence of new problem frames. This then appears to provide the necessary condition for the propagation and progression of solution ideas, based upon a clearer framing of the design problem.

An analysis of the participant’s design orientated sketching work indicates the ways in which sketching acts as stimuli for solution-focused strategies aimed at the proposition and development of conceptual design ideas.

Following framing events related to developing an understanding of the design problem, the subject’s design-focused sketching work appeared to provide the necessary catalyst for the development and testing of the frame’s suitability to address an emerging understanding of the design problem.

That is, the participant’s sketching activity provided a means through which new problem frames were tested and refined through propositional solution candidates as sketches.

Moreover, results indicate it is through the use of design representation as sketches that design moves were self-interpreted in various ways to both explore a current focus of attention and provide emergent insights often related to previous frames of reference and previously proposed solution ideas. These insights were observed to be further tested and explored through further framing of different aspects of the design problem. That is, the expression and testing of a currently focused aspect of conceptual design work through sketching provided opportunities to emerge which relate to previously explored frames. Sketching moves that were aimed at exploring solution conjectures under current consideration, thus acted as a kind of catalyst for reengagement with previous frames through new solution-focused insights emergent from the participant’s sketching activity.

These were then progressed through further moves and reflections. This thus resulted in repeated shifts of attention between aspects of design work. Results indicate that these shifts, also identified by others (Cross 1990, Cross 2007, Lawson and Dorst 2009) as an indicator of expertise in design practice, are critically facilitated and enhanced through sketching activity.