Fixedness in problem-solving

As designers we place a lot of value in experience and creativity when tackling new problems. Experience informs us how similar problems were solved in the past - what worked well, and what didn’t. And we use creativity to help design engaging solutions that resonate with the target audience. Most times these skills serve us well as we build greater domain knowledge of different problem areas. We apply the knowledge we have built up (over hundreds or thousands of work hours) when we encounter new problems to solve.

Although we may see ourselves as ‘expert’ designers, we can be blind-sided by functional or mental fixedness by applying previous solutions to a problem that requires a more unique approach. As Öllinger et al outlined Abrham Luchins’ findings from his famous water jug experiments carried out in 1942,

Luchins proposed that the repeated application of a successful method makes blind any alternative approach, because of the mechanization of the particular solution method – resulting in what he termed mental set. (Öllinger et al, 2008)

Functional fixedness is when we have a ‘block’ in imagining or employing an object in any other way that deviates from its primary function. For example, a spoon is only seen as a tool for eating and this fixedness blocks our ability to see it being used as a balance weight, tool for digging or even a musical instrument. Or think of the multi-purpose ways of using a knife - as an instrument for cutting food, a temporary peg to hang things upon, a tent peg or even a tool for shaving.

Mental fixedness (commonly known as mental set or the Einstellung effect) occurs when we apply familiar (or rigid) strategies without carefully considering all the requirements of a given problem. This is problematic when we encounter problems that are unique, deviate from the norm or require us to solve them is different ways.

The concept of fixedness was first proposed by the Gestalt psychologist Karl Duncker in the 1920’s. He described this cognitive bias as a mental block that prevented people using objects in new (or atypical) ways to solve a problem (Wikipedia, 2013). This type of bias can limit our ability to envision objects as multi-purpose items, where they can be utilised in solving new problems in a manner they were not originally designed for. This becomes especially true where we have become ‘primed’ in the function of an item. Priming occurs when an object’s primary function is demonstrated or learnt in advance. This priming can then block people in imagining the object being used in any other way that differs from the functional use it was intended for.

Problem solving can be inefficient when the solution requires subjects to generate an atypical function for an object and the object’s typical function has been primed. Subjects become ‘‘fixed’’ on the design function of the object, and problem solving suffers relative to control conditions in which the object’s function is not demonstrated. (German & Clark Barrett, 2003)

It is hardly surprising that fixedness occurs in humans. We learn from an early age to understand the purpose and function of tools and objects around us. (German & Barrett, 2003). Lave and Wengers used their model of Situated Learning (Lave & Wenger, 1991) to describe how people learn by socialisation, visualisation, and imitation within social and cultural environments.

The issue for designers is if we fail to recognise fixedness is occurring; and not realising (through insight) that a final solution requires us to sidestep conventional (or previously employed) methods to solve a new problem (Öllinger et al, 2008). Gaining insight will not however provide a final solution. It only allows us to recognise a current solution will not solve the problem at hand. In addition, if we rely on prior knowledge to construct a series of problem-solving rules (also known as procedures), then these procedures are more likely to be selected again, and any new insights are inhibited.

The procedures required for solving insight problems therefore begin with a very low probability of selection. It is through the repeated failure of more high probability solution attempts (and thus a reduction in their probability of selection) that some time later an appropriate solution procedure is selected. (Öllinger et al, 2008)

And,

When a problem solver is confronted with an insight problem, there is an initial (often unconscious) activation of prior knowledge that was useful for solving apparently similar (noninsight) problems in the past, but is a hindrance for solving the insight problem. As a consequence a “biased” problem representation is established making it very difficult to access the operators that are necessary to transform the problem state into a proper solution. Without an appropriate solution procedure the problem solver gets stuck in an impasse. (Öllinger et al, 2008)

Taking a different view

Below are a number of methods and techniques to help you change decision-making strategies and also validation current assumptions and solutions.

1. Re-visiting the problem proposition (clarify the problem is not ill-defined)

Revisiting the problem proposition with the aim of establishing if the problem has been clearly defined, and has well stated goals. Examine the assumptions and supporting data or facts that have been used to define the problem. An ill-defined problem can lead to misleading assumptions and differing goal expectations. Also, ill-defined problems are hard to formalise and sometimes mask underlying sub-problems.

2. Restructuring the problem

One way of approaching a problem in a fresh way is to restructure it, or changing how it is represented. Can the problem be broken down into smaller sub-problems, and if so, is the sequence or goals of each sub-problem of equal weight or importance. Restructuring problems can promote insight and allow designers to circumvent a block of mental set fixedness.

A useful method for restructuring a problem is the use of analogies. Analogical problem-solving allows us to look at how one problem (the target problem) and see a solution from another problem (the source problem) that could be applied to the target problem.

For example, in 1945 Druncker used the scenario of a multi-pronged attack on an enemy’s fortress as an analogy for a doctor trying to cure an inoperable tumour in a patient. One solution existed where an X-ray could be used to destroy the tumour. However, at the intensity required, the X-ray would not only destroy the tumour but also damage healthy tissue. And low intensity rays would neither damage vital organs or affect the tumour. By restructuring the problem using the analogy of a multi-pronged attack on a fortress a solution was arrived at where multiple low-intensity rays could be used at the same time that could be focused together to destroy the tumour without individually causing any damage to healthy tissue.

In 2008 Öllinger and associates put forward their Representational Change Theory (RCT) which proposed two possibilities for restructuring problems - chunk decomposition and constraint relaxation,

First, the relationship between the constituents of a given problem can be changed – for example, a problem entity may be perceived as a whole when in fact it can be broken down into further subcomponents. This is termed chunk decomposition. Second, the initial representation of the problem may place unnecessary constraints on the problem itself, and thus constraints need to be relaxed. This process is termed constraint relaxation. (Öllinger et al, 2008)

3. Using tools to validate decisions

Developed by psychologist Chris Argyris (as part of his Double Loop and Model I/Model II theories of learning), the Ladder of Inference helps explain the thinking process we go through to reach a decision. As a tool it can be used to test conclusions and also validate decisions, by looking at various steps in the decision-making process and questions assumptions, supporting facts or data, beliefs and conclusions reached.

4. Peer group reviews

Using team members and peers is a great way of articulating and getting feedback on design decisions. Although peer reviews need to be well facilitated to avoid subjective stances being taken, they can be an ideal method for explaining your understanding of the underlying problems, facts and data; the problem domain; the needs of end-users; and reasoning behind the decisions you have taken. These meetings can also be a way to let peers discuss similar problems and patterns they encountered and solved on other projects.

5. User testing

User testing rough screen flows and UI design early, and often in the project cycle to validate understanding and assumptions about business and end-user needs. Don’t be afraid to embrace mistakes if you discover some salient user experience issues with your rough concepts. End-users can offer valuable insights through post-test discussions and follow-up interviews.

6. Developing new problem solving methods through practice-based learning (PBL)

Learning does not happen in a void, but is situated and socially influenced (Gherardi, 2001). And while domain knowledge and design experience increases with age, we are all prone to falling into patterns that inhibit new ways of thinking about problem-solving. The need to stay open to developing new ways to learn and problem-solve can be helped through practice-based learning (PBL).

I first came into contact with PBL while doing an MA in 2004. PBL is not always focused on goal-based outcomes (or even a specific outcome) but instead promotes ‘learning how to learning’ through participation or 'doing'. This type of learning differs from tutor-led or classroom-based learning, in that PBL engages the learner in activities that may not have direct goal-based outcomes as experienced in traditional education structures. However, PBL allows learners to explore new ways to learn and discover different ways to frame and solve the same problem.

Unfortunately, classroom and real-world development experiences are typically provided independently as if there were no need to merge theory with practice. Work-based learning, on the other hand, deliberately merges theory with practice and acknowledges the intersection of explicit and tacit forms of knowing at both individual and collective levels. It recognizes that learning is acquired in the midst of practice and can occur while working on the tasks and relationships at hand. (Raelin, 1998)

The idea of learning through practice is also supported within David Kolb’s and Roger Fry’s cyclical model of experiential learning - (1) do or experience, (2) reflect on what was observed or learnt, (3) develop general internal theories about the learning, and (4) apply the learning within future experiences.

Kolb and Fry (1975) argue that the learning cycle can begin at any one of the four points – and that it should really be approached as a continuous spiral. However, it is suggested that the learning process often begins with a person carrying out a particular action and then seeing the effect of the action in this situation. (Smith, 2001, 2010)

PBL can be sequenced and self-paced, based on the learning style, skill level, preferences and time available on the part of the learner. PBL caters for people working individually or as part of a team. Another key component of PBL is reflecting through journals or logs to help track important learnings, patterns and outcomes from the experience (Raelin, 1998). PBL can be used for problem-solving on personal projects as well as employed within organisational work-based learning. However, facilitator or leader roles may prove beneficial when PBL is employed within organisations.

General Electric's (GE's) Executive Development Course is a month-long experience during which time promising executives assemble into teams to work on a specific assignment. The assignments vary by topic from year to year and, although sponsors get a completed project at the end of the month, the real issue for GE's Leadership Development Center at Crotonville is the value of the learning experience more than the assignment per se. (Raelin, 1998)