I recently came across some old-ish work in analogical reasoning that tickled my brain about priming during experiments. What concerns me is the quality of conclusions reached by the research and how we might improve our experimental methods in studying analogical reasoning, especially as it pertains to design. References are at the end of this post.
Bear with me; we need to set things up first.
In a 1945 paper, Duncker offered the following scenario.
Suppose you are a doctor faced with a patient who has a malignant tumor in his stomach. It is impossible to operate on the patient, but unless the tumor is destroyed, the patient will die. There is a kind of ray that can be used to destroy the tumor. If the rays reach the tumor all at once at a sufficiently high intensity, the tumor will be destroyed. Unfortunately, at this intensity, the healthy tissue that the rays pass through on the way to the tumor will also be destroyed. At lower intensities, the rays are harmless to healthy tissue, but they will not affect the tumor either. What type of procedure might be used to destroy the tumor with the rays and at the same time avoid destroying the healthy tissue?
What immediately came to my mind was to install a tube, like a shunt, that would move aside any delicate tissues and organs; it would end at the tumour. The “rays” could be projected down the tube to the tumour. The tube could be removed once treatment is completed.
According to Loewenstein, Thompson, and Gentner (1999), 10% of people who were asked to solve this problem suggested using multiple, simultaneous low-dose rays from various angles. Each ray is low-dose, so wouldn’t damage the tissues it passed through to reach the tumour. All the rays would converge at the tumour, leading to a cumulative high dose only at the tumour.
In 1980, Gick and Holyoak recognized this particular solution as a potential testbed for examining whether analogical reasoning could be prompted. They wondered if individuals that had been primed by an analogous situation suggestive of the multi-directional solution would be more likely to think of it. They divided the participants in their experiment into two groups. Both groups were asked to solve Duncker’s radiation problem, but one group was first primed by having them read the following scenario:
A general wishes to capture a fortress located in the center of a country. There are many roads radiating outward from the fortress. All have been mined so that while small groups of men can pass over the roads safely, any large force will detonate the mines. A full- scale direct attack is therefore impossible. The general’s solution is to divide his army into small groups, send each group to the head of a different road, and have the groups converge simultaneously on the fortress.
The control group is the group that only solved the problem, without first reading the military scenario. It should come as no surprise that those who read the military scenario were more likely do identify the multi-ray solution when faced with the radiation scenario. I think this is an example of priming: the military scenario plants in the participants’ minds the seeds of the idea of attacking the tumour from multiple simultaneous directions via the analogy to multiple smaller military forces each coming from a different road.
While this constitutes are reasonable verification of the notion that analogies can play an important role in design ideation, I think there are experimental shortcomings here that could be addressed methodologically, and that in any case there are some important implications for the execution of design ideation in practice.
It’s important to keep in mind a fundamental imperfection of the entire experiment. In practice, an individual faced with the radiation problem would obviously (a) be a medical professional (i.e., a person having particular expertise uncommon in the population at large about cancers and their treatment), and (b) have access to a significant body of knowledge – including the expertise of others – about the “ray” machine itself. Butchers, accountants, and engineers would unlikely be solving this kind of problem in “real life.”
HOW MUCH INFORMATION IS ENOUGH INFORMATION?
Firstly, it is not clear that the control group had enough information about the scenario to be able to form the analogy. For instance, my personal medical experience is limited to machines like X-ray and ultrasound machines. These machines share a particular feature: they are uni-directional. So if I were in the control group, I would already be primed to think that available machines simply could not deliver multiple simultaneous “rays.” Whether such priming is accurate or not is irrelevant; that’s how my mind would function. Indeed, when I first read the military scenario (after having read the radiation scenario), I immediately saw the analogy myself with some surprise because it was so different from my own personal experience. This factor of experience appears unaccounted for in the research I’ve noted here.
So, if I were doing this experiment over (and I just might try it), I would want to control for the background of the test group. For instance, I might have one group with no medical or technical training whatsoever, a second group of medical professionals, and a third group of technical professionals (e.g., engineers). The professionals would have to have had some practical exposure to the kinds of equipment that might generate the “rays” in the radiation scenario. I would further exclude anyone with significant vocational or avocational expertise in military strategy (to help avoid any of the control group members from already knowing the tactic in the military scenario). Then I would prime half of each of the three group with the military scenario, and see what happens. Specifically, I could see whether, and to what degree, the military scenario helped promote analogical reasoning, depending on what existent knowledge about the radiation scenario the participants already had.
The research question here would be to see if relevant background of the participants helps or hinders the sought-after analogical reasoning. Knowing about design fixation, it is possible that people without appropriate background might outperform the professionals if, for instance, medical professionals are used to dealing with such situations in specific ways.
IS ANALOGY BETTER THAN “REGULAR” PRIMING?
Secondly, since it is possible to prime people without using analogy, one may ask what difference in performance might there have been if one compared the group primed with the military scenario to a group primed without such an analogous scenario but with a modified radiation problem instead.
In this case, I would have three groups. All three groups would be filtered so that no one had prior military, medical, or engineering training in relevant areas of knowledge. The control group would be given the original radiation scenario task. The second group would be primed with the military scenario, as was done originally. The third group would be given a suitably modified radiation scenario that provides more information about the nature and constitution of the equipment that could be used for treatment.
The modification to the radiation scenario could possibly take the form of images of various kinds of machines that could generate the appropriate “rays.” Since the participants would lack the training to understand the machines, the images would have to be carefully annotated so as to explain the equipment without “giving away” any particular solution. One of the images would be of equipment clearly able to project multiple simultaneous beams. An alternative modification could be to re-word the scenario to include new text describing available equipment; the description would include, among other details, something like: “The machine has four separate ray emitters.” If only the number and quality of emitters were added, this might draw participants attention artificially to sought-after analogy. It would therefore be important to include a wide variety of information about the equipment.
In this case, the research question would investigate the relative effectiveness of analogical reasoning compared to providing appropriate background about the machine. Since no participant has experience in the area, one would expect fixation to be minimal.
I could also repeat this experiment for participants that have appropriate medical training, and again for participants having appropriate engineering training. This would help determine the influence of background. It might be that one or another type of background impacts whether analogical reasoning is more likely used, or whether a simple description of the equipment is enough.
WHY ONLY THE MULTI-RAY SOLUTION?
Thirdly, it’s not clear that other solutions besides the multi-ray solution may have resulted from analogical reasoning. Only 10% of participants in the original study identified the multiple-ray solution. What were the other solutions? Were any of them driven by analogies?
This question is harder to address, since sometimes (often?) one performs analogical reasoning unconsciously – that is, due to prior priming or fixation, certain ideas are automatically promoted to consciousness. Much of our “thinking” occurs unconsciously – that is, we are not necessarily aware that our minds are solving problems till a solution is found. (For instance, see the work of Keith Sawyer.) Therefore, we cannot expect participants to know whether their solutions were driven by analogies. However, we can elicit, either via survey, recording, or interview, information about what they were thinking during the task. This might lead useable data about the prevalence of analogical reasoning throughout the experiment.
DOES COGNITIVE STYLE AFFECT ANALOGICAL REASONING?
Fourthly, there may be relevant correlations between other cognitive factors and analogical reasoning. For instance, it may be that only people with certain cognitive styles are prone to analogical reasoning. If this is the case, then cognitive style must be controlled for in the experiment if it is to yield useful results. The problem here is that there are many models of cognitive style (e.g., see the overview at Wikipedia). One might try to measure the participants’ cognitive styles using multiple instruments, each representing a different model, and then seek to correlate occurrence of analogical reasoning with each of the styles as indicated by each of the instruments.
DOES SOLUTION FEASIBILITY MATTER?
Finally, one must wonder whether the solutions generated by either of the groups were physically realizable. That is, if it simply is not possible to deliver multiple low-dose “rays” from different directions – as suggested above – then it may be that persons with suitable background in, say, medical sciences actually thought of that solution but immediately discounted it because they already knew it was impossible to execute. It should be obvious that this could artificially alter the experimental results if the background of the participants is not accounted for in the experimental design.
ANALOGY IN REAL LIFE
There’s one last point that I want to make here: there’s a very important difference between running these experiments and applying the results to real life. Let’s just say that after more experimentation, we find that there really is a good connection between analogical reasoning and design ideation – that people who can reason well analogically are, statistically at least, more likely to be innovative. How do we get practicing designers (and, indeed, everyone) to do more analogical reasoning?
Since analogical reasoning is founded on forming connections between problems you’re trying to solve, and things you already know about, and since the moment of realizing an analogy seems to occur unconsciously, what can we do to change how people do this?
I don’t know what the answer is – largely because we don’t seem to yet know enough about how analogical reasoning happens. However, it does seem to hinge on psychological priming. So one question bearing greater study is this: What is the impact of self-priming (i.e., getting a person to prime themselves) on analogical reasoning?
This is really interesting stuff for me. Clearly, there is very suggestive evidence already there, but I don’t think it’s been properly explored yet. That’s one of the great “fun” parts about scientific research – answering one question usually leads to some answers, but even more other questions. There’s always room for more exploration.
- Duncker, K. (1945). On Problem Solving. Psychological Monographs, 58(5), Whole No. 270.
- Gick, M. L., & Holyoak, K. J. (1980). Analogical problem solving. Cognitive Psychology, 12, 306-355.
- Loewenstein, J., Thompson, L., & Gentner, D. (1999). Analogical encoding facilitates knowledge transfer in negotiation. Psychonomic Bulletin & Review, 6(4), 586-597.