Background

If we expect PSPD to support creativity, we need to have an understanding of what we are trying to support. After all, how can we support what we don't understand!

The collaborative processes of generating design requirements and envisioned system designs remain something of a ‘magic art’, within both PD and other systems development approaches. In the PD literature, this ‘magic art’ is frequently referred to using terms such as creativity and innovation. However, while participatory design may be viewed as a collaborative or social creative process and PD researchers and practitioners use the term ‘creativity’ when referring to the design process, they provide little definition of what this term means and what is actually involved in this process of social creativity.

What is Creativity?

Definitions of creativity have been developed and evolved over several decades. There have been three main concepts by which creativity has been defined: the creative process, the creative person and the creative product.

The Creative Process

Early definitions of creativity defined it in terms of the creative process, that is a process essentially internal to an individual by which ideas are generated. Boden described the creative process as a person’s exploration and transformation of conceptual spaces. The exploration of conceptual spaces relates to recalling knowledge, while transforming conceptual spaces relates to forming new associations between knowledge structures.

Koestler proposed that creativity involves a ‘bisociative process’ whereby an individual deliberately connects previously unrelated ‘matrices of thought’ to produce a creative idea. When Koestler uses the term ‘matrix of thought’, he is referring to an idea or concept in an individual’s mind. However, when we view creativity as a collaborative or social process, the matrices of thought that are to be combined in the generation of creative ideas are not necessarily in the mind of a single individual but may come from more than one person in the group.

In addition, these combinations of matrices of thought are not necessarily just in the minds of the participants. Some theoretical perspectives, such as Distributed Cognition, move away from the individual mind and focus on the inclusion of significant features in the environment that support cognition. In our previous work on PD, we have explored the use of external shared representations to support the development of shared design ideas and understandings. From this perspective, we may consider the creative process as combining matrices of thought in our mind and our environment.

Furthermore, although Koestler refers to creativity as a ‘bisociative process of unrelated matrices of thought’, it should be considered as more of a multi-sociative process of related or unrelated matrices of thought as there may be occasions when the creative process involves the combination of more than two matrices of thought which are either related or unrelated.

Whilst these definitions provide a possible explanation as to how creativity comes about in the mind of the individual, they do not give us a complete understanding of creativity and provide no way of measuring when creativity has occurred or to what degree.

The Creative Person

A dominant approach in the 1950s was that of defining creativity in terms of the creative person. Guildford describes the creative person in terms of traits, ‘Creative personality is then a matter of those patterns of traits that are characteristic of creative persons'. It may be the case that these traits assist in the creative process, allowing the individual to explore and transform conceptual spaces in their mind more easily than a less creative person. However, while this apparently circular definition tells us that traits are what make a person creative, it fails to say what these traits of a creative person actually are.

Various creativity tests have been devised to assess the traits of creative individuals: personality inventories, biographical inventories and behaviour tests. These tests were developed based on studies using very large numbers of subjects in which the creativity of the subjects was assessed by ‘experts’ – an assessment that was itself subjective. Gough introduced a Creativity Personality Scale using an Adjective Check List where 18 adjectives positively relate to creativity and 12 adjectives negatively relate to creativity. In this test, individuals use the adjectives to describe themselves. The test adds a unit point for the use of a positive adjective and subtracts a unit point for the use of a negative adjective. Creative individuals tend to use the adjectives positively related to creativity rather than the negatively related adjectives and therefore score higher than less creative individuals.

However, Ward argues that test scores should not be considered a measure of creativity. While these ‘abilities’ are important to creativity, Amabile argues that it is inappropriate to label the results of these test scores as some direct indication of some global quality known as creativity. Although such tests may not be suitable for assessing creativity, they could be used to identify some of the attributes of creativity, as in studies by Elam and Mead and Marakas and Elam, to compare the potential for creativity of different people.

The Creative Product

Definitions of the creative product refer to the product’s reflecting some distinguishing signs of creativity – such as aesthetic responses brought about in the observers. Theorists who define creativity in terms of the creative product tend to include characteristics of ‘novelty’ and ‘appropriateness’.

How do we know then when an idea is novel? We consider a novel idea to mean the combination of two or more matrices of thought that are considered new or unusual. However, what may be considered novel to one person may not to another. Boden views novelty as belonging to one of two categories: Psychological Novelty (P-Novel) and Historical Novelty (H-Novel). P-Novel is an idea which is new to the mind in which it arose, though it may have been thought of by others before. H-Novel is an idea which is P-Novel and has never been thought of by anyone else before. To assess if an idea is P-Novel, one could simply ask the individual who thought of the idea using a retrospective protocol, but assessing an idea to be H-Novel is improbable as one would have to cross cultures and time to see if the idea had occurred before. Due to the more objective measure of P-Novelty, in our research we consider as novel an idea that is new or unusual to the mind in which it arose. This view has also been adopted by other creativity researchers.

Novelty is a necessary but not sufficient feature of creativity. It is appropriateness that differentiates novelty from creativity. So, how do we assess a novel idea to be appropriate – and appropriate to what? Many creative process models have a stage of problem definition and preparation. When we are creative we generally start with a problem, whether it is producing a painting or identifying a solution to a design problem. During the problem definition and preparation stage of the process, the problem is explored, allowing characteristics of potential solutions to be determined. A solution is considered appropriate if it conforms to these characteristics. The characteristics that determine appropriateness are likely to vary from domain to domain. Relating value to our concept of appropriateness, Brannigan comments that ‘value judgments are to some extent culture-related, since what is valued by one person or social group may or may not be valued by another’. Thus, the appropriateness of a product should be assessed in relation to the setting for which the product was intended.

Boden argues that if we can identify the creative ideas generated through a creative process, we could develop some way of ‘counting’ them in order to measure the creativity of an individual or group. If we consider these ideas to be the ‘creative product’, such a measurement could identify when creativity has occurred and give some objective measure of its extent. However, Amabile argues that ‘assessment of creativity cannot be achieved by objective analysis alone. Some type of subjective analysis is required’. Amabile proposes the subjective analysis of experts to determine the degree of creativity of a product. Such a measurement has been used as a reliable measure for the degree of creativity associated with a product, although Marakas and Elam notes that research is needed on this issue of measurement.

Creative Process Models

The concept of ‘creative process’ described above views creativity as essentially the individual exploration and transformation of conceptual spaces to generate ideas. There is in addition a long history of research that uses creative process models to describe various phases that occur in the process of being creative, including but not limited to idea generation.

One of the first models of the process of being creative was proposed by Wallas, who described creativity as involving four phases: Preparation; Incubation; Illumination; and Verification. Preparation is a stage in which one clarifies the problem and develops an understanding of it, so that one is prepared for what may be needed in potential solutions to the problem. This stage may involve gathering relevant data about a problem and reviewing it. Incubation is when one no longer consciously considers the problem. However, although conscious thought is suspended, the problem remains as an ambient thought awaiting some creative insight (i.e. a ‘eureka’ moment). The Illumination stage is when this creative insight occurs. Nemiro describes this as ‘when there is a sudden change in perception, a new idea combination, or a transformation that produces an acceptable solution to the problem at hand’. This reflects both Koestler’s combination of matrices of thought and Boden’s transformation of conceptual spaces. The final stage, Verification involves making sure that one’s creative insight or novel idea is in fact an appropriate solution to one’s problem. This reflects the characteristics of novelty and appropriateness used to define a creative product.

Later models moved away from proposing unconscious stages of incubation and illumination, towards a more conscious process of deliberately coming up with ideas. Osborn described the creative process as comprising two main stages: Idea Generation and Idea Evaluation. Idea Generation is made up of a further two sub-stages: Fact-finding – the process of problem definition and preparation – and Idea-finding – the process of producing novel ideas through the combination of old, existing ideas. During the Idea Evaluation stage, these novel ideas are assessed for their ‘appropriateness’, allowing creative solutions to be identified.

Amabile provided a componential framework, showing how domain-relevant skills, creative-relevant skills and task motivation might contribute to the creative process. Domain-relevant skills are attributes such as factual knowledge and skills – attributes that will affect an individual’s performance in a given domain. Creative relevant skills include a person’s cognitive style – this will influence the way the individual explores and transforms conceptual spaces. Task motivation determines how an individual approaches a task – an individual’s enthusiasm for a task.

In Amabile’s creative process model there are five stages: problem and task presentation; preparation; response generation; response validation; and outcome. In problem and task presentation the individual is presented with the problem and what is involved. During this stage of the process, task motivation has to be high, so the individual has sufficient interest to pursue solving the problem. While Amabile does not acknowledge this, domain knowledge – does the individual understand the problem and what is involved – may also be important in this stage, having an influence on task motivation. Preparation involves the individual building up knowledge about the problem and researching what a potential solution may necessitate. Domain-relevant skills are particular important at this stage, as domain knowledge will play an important part in generating an acceptable solution. Response generation is heavily dependent on creative relevant skills and task motivation. Creative-relevant skills could influence the quality of the ideas produced, as the better the individual is at exploring conceptual spaces, the more novel the solutions are likely to be. Task motivation could also have an influence on the quantity of ideas produced. Osborn argues that ‘quantity breeds quality’ therefore a high task motivation is required so that more ideas are produced, as the more interested an individual is in a problem the more time and effort she is likely to spend generating ideas to solve it. Validation of the responses and solutions generated is heavily reliant on domain-relevant skills, as the individual must have knowledge by which to assess the appropriateness of the generated ideas. Finally, one of three possible outcomes is achieved: a solution is obtained and the process has been a success; all ideas for a solution are rejected and therefore the process has failed; the ideas generated have made a contribution to the problem, but it is not yet solved, in which case we return back to the first stage of the creative process and reassess the problem.

Shneiderman uses a four stage model to describe the creative process: Collect; Relate; Create; and Donate. Collect is the initial stage of collecting information about the problem from information resources such as digital libraries and the Web. The Relate stage in the model is when one consults with peers and mentors. This stage should be performed throughout the model as an iterative cycle, interleaved with the other stages. Create is the stage in which one explores, composes and evaluates possible solutions. Donate is the dissemination of the results to information resources. This stage may cause new needs to be identified or cause new ideas to be generated by the community who view the solutions, resulting in returning to previous stages in the model.

A Comparison of Creative Process Models

Creative process models have been evolving since at least 1926 when Wallas introduced one of the first. Since then, researchers in the area of creativity have been refining these models more accurately to reflect the process involved in generating creative solutions and to assist in making individuals and groups more creative. It should be noted that such models are not intended to be step-wise linear models, but rather models which show various phases of the intertwined and iterative nature of creativity – descriptive rather than prescriptive. We now consider the similarities and differences between these different models.

A common theme with all the models is the analytical stage of preparation. Before the generation of ideas to solve our problem, the individual views relevant information associated with the problem to develop an understanding of what is required in order to generate an acceptable solution. Amabile refers to this as the building up of one’s domain knowledge.

Once the individual understands the problem and has built up the relevant domain knowledge, the more specifically creative phase of the creative process model occurs – Idea Generation. Some models view this phase as a subconscious activity with stages of incubation and illumination, while others believe it to be a conscious activity where the individual deliberately generates new ideas through combining old, existing ideas. Models.

An argument against the subconscious view is offered by Perkins [41] who argues that subconscious mental processes are behind all forms of thinking, and are therefore not specifically related to creativity. Plsek argues that just because we cannot fully describe our subconscious thought process does not mean we are not in control of it.

All the creative process models have an Idea Evaluation stage. Such a stage is an important part of the creative process, as it is through the evaluation of the novel ideas produced during the Idea Generation stage that we judge their appropriateness, and so whether or not they are considered creative.

Shneiderman’s creative process model combines the stages of Idea Generation and Idea Evaluation into a Create stage. The combination of these two stages is unusual as all the other models separate Idea Generation and Idea Evaluation. Osborn says that the most important principle of Idea Generation is deferment of judgment. Amabile has shown in her studies how evaluation can decrease creativity and therefore the number of ideas which are generated. A decrease in the number of ideas produced will also violate another of Osborn’s principles: quantity is wanted. Hence, a refinement to Shneiderman’s creative process model would be to replace the create stage of his model with two stages which consider Idea Generation and Idea Evaluation separately.

Many of the models consider the creative process from the perspective of the individual. Shneiderman describes these as inspirationalist and structuralist views of creativity. Inspirationalists focus on the individual coming up with ideas, in a fashion such as the ‘eureka’ moment – a sudden change in perception giving rise to an idea from the subconscious. Structuralists apply more systematic approaches to exploring and transforming conceptual spaces, as described by the creative process models of Amabile and Osborn. Influenced by researchers such a Csikszentmihalyi, who emphasises the importance of the social nature of creativity, Shneiderman brings a social influence to the creative process model. The latter situationalist view of creativity moves away from the individual perspective on creativity and views creativity as more of a social process, placing more importance on interaction and collaboration with other individuals and the world around us.

Towards a Unified Understanding of Creativity

‘In our study of creativity in design then we need to examine not only products but also processes and persons’ Brian Lawson. While the focus of definitions of creativity has evolved over time, from process to person to product, these are all essentially important components of creativity. Each individual, or member of a group, has certain creative abilities; he/she may explore and transform conceptual spaces, combine matrices of thought to generate new ideas; and these ideas may consist in or lead to the development of a creative product. Previous research has tended to focus on the embodiment of such ideas in other subsequent products, viewing the latter as the ‘creative product’. In this research, we focus on creativity as the production of the ideas themselves. In line with Boden, we argue that considering the generated ideas as the creative product can help provide a basis for measuring or assessing the nature and extent of the creativity that has occurred.

Drawing together the strands of previous research, we propose a unified definition of creativity:

‘Creativity is the generation of ideas, which are a combination of two or more matrices of thought, which are considered unusual or new to the mind in which the ideas arose and are appropriate to the characteristics of a desired solution defined during the problem definition and preparation stage of the creative process’

From the perspective of design our products from the creative process are the design ideas generated to solve the design problem at hand. A design idea is deemed creative if it is new or unusual to the mind in which it arose (novelty) and conforms to the requirements of the design problem (appropriateness).

The Problems with Collaborative Creativity

Gennari and Reddy describe the design process as ‘human activity, involving communication and creative thought among a group of participants’. A fundamental question about activities such as design that involve a social component is: what effect does the social nature of this activity have on creativity?

In 1958, Taylor et al conducted a study comparing real groups (i.e. face-to-face interacting groups) with nominal groups (i.e. individuals working on their own and then collating their outputs to form a cumulative output), to test Osborn’s claim that ‘the average person can think up twice as many ideas when working within a group than when working alone’. Taylor et al found that nominal groups produced nearly twice as many non-replicated ideas as real groups – refuting Osborn’s claim.  Since the Taylor et al study, over 50 years of empirical studies have shown nominal groups to outperform real groups.  The implication of this body of research is that collaborating groups, such as design teams, are not being as creative as they could be.
The three major explanations that have been explored thoroughly by the creativity community as to why nominal groups outperform real groups are the social influences of procedural mechanisms (i.e. production blocking), social psychological mechanisms (i.e. evaluation apprehension) and economic mechanisms (i.e. free riding).

Production blocking has been argued to be the most important cause of nominal groups’ outperforming real groups’.  Production blocking is common when ideas are expressed verbally within a group. Verbally expressing ideas is a form of asynchronous interaction. The problem with asynchronous forms of interaction is that group members are simultaneously prohibited from expressing their ideas. They may subsequently forget their ideas or suppress them because they may feel their ideas are less relevant as time passes. Another problem is that they may rehearse their ideas internally, preventing them from concentrating on what other members say.  Finally, if group members are prevented from expressing their ideas as they occur, they may be discouraged from producing further ideas.  To mitigate the effects of production blocking, researchers have moved towards using synchronous interaction techniques for expressing ideas.  Within the The CSCW community describe the use of single display groupware (SDG) and synchronous groupware systems (SGS) to prevent production blocking in real time collaboration.

Evaluation apprehension occurs when group members fear criticism from others within the group.  This can prevent group members from expressing ideas. The negative effect of evaluation apprehension reduces the quantity of ideas produced in groups.  To overcome the negative effects of evaluation apprehension, it has been suggested by some researchers that anonymous means of expressing ideas remove an individual’s identification with an idea and therefore helps encourage people to express their ideas without fear of criticism.  Some technologies described in the CSCW community allow for anonymizing individual input, thereby preventing evaluation apprehension.

Free riding, otherwise known as social loafing, is the result of group members’ becoming lazy, relying on other members in the group and not contributing as many ideas as they could.  When working in a group, group members may assume the group’s output to be assessed collectively, whereas when working alone one has to take responsibility for her performance.  Therefore, assessing members within a group on their individual performance rather than the group collectively reduces the effect of free-riding.  Paulus also refers to social stimulation as a way to reduce the impact of free-riding – encouraging a high motivation level in groups by increasing accountability for individual performance.  Various studies have shown that providing groups with a comparison standard increases their performance and providing explicit feedback about individual performance also increases performance of group members.

Much of the research aimed at increasing the effectiveness of group creativity has looked at production blocking, evaluation apprehension and free riding, presenting methods and techniques to reduce the dampening effects of these factors on social creativity.  Much of this research has only increased the effectiveness of real groups over nominal groups slightly in some conditions, while there are still many cases where nominal groups outperform real groups.

In our previous research we have developed a theory that shows how real groups have the potential to generate more creative ideas than nominal groups by taking advantage of shared domains of knowledge.  Fischer argues that ‘the unaided individual mind is highly overrated … much of our intelligence and creativity results from interaction and collaboration with other individuals’.  Each individual has a domain of knowledge and within this domain of knowledge has a collection of matrices of thought (i.e. existing ideas or concepts).  An individual has only the matrices of thought in her own domain of knowledge, whereas in a real group individuals can interact with each other, externalizing their matrices of thought and making them available to others.  Our theory explains how real groups have the potential to take advantage of these shared domains of knowledge thereby allowing them to generate more creative ideas than nominal groups.  By externalizing matrices of thought, more combinations of matrices of thought can be derived, therefore producing more creative ideas.

However, as nominal groups have frequently been shown to outperform real groups in terms of creativity, some researchers have suggested that the process of idea generation should be performed by nominal groups alone in order to increase creativity.  Building upon previous research in reducing the social influences on creativity, we extend consideration of procedural mechanisms, social psychological mechanisms and economic mechanisms to include a fourth type: operational mechanisms.  We define operational mechanisms to include, for example, group size and group composition. Looking at these different mechanisms we have to ask ourselves the question of, 'how can one go about supporting creativity?' - PSPD?