Many UX professionals speak of user research in terms of generative vs evaluative. Generative research occurs before a designed artifact has been produced and helps determine what kind of design will be most effective. Evaluative research occurs after a designed artifact has been produced and helps improve the design’s effectiveness.
I used a very similar dichotomy for some time, informing research vs validating research. I used to think the two dichotomies were more or less synonymous and interchangeable, but now I am realizing they are not. In fact, they are complementary — in an important way.
The distinction falls along the lines of holism and atomism.
Before I launch in, here is a short primer on holism and atomism: Holism understands realities essentially as wholes, within which features can be articulated by a variety of schema. Atomism understands realities essentially as parts, which combine and build up into wholes. “Right-brainers” tend to perceive things in holistic terms, grabbing things in big gestalt impressions. Holism tends to have a subjective character. “Left-brainers” are more often atomistic, dissecting things into small, manageable elements. Atomism tends to be have an objective character. There is actually no necessary conflict between holism and atomism — as long as they are seen as ways to perceive and schematize reality. But most people, without noticing, tend to project them into reality as metaphysical truths, and then succumb to reductionism, understanding the world to be essentially constituted of parts, or essentially constituted of a whole, and then further confounding their understanding of truth with truth itself. Once this occurs, the complementary view is experienced as defective. In my view, the antidote to reductionism is pluralism, the understanding that multiple valid ways exist to make sense of reality, each a particular pattern of perceptiveness, blindness, competence and incompetence, and all with the potential to complement the others. (This is why teams are so important, and why they must be seen as collaborative pluralities and not primarily as a sum of labor-power resources…)
Back to the forms of research. Oddly, each dichotomy has one atomistic and one holistic term. In generative-evaluative, generative is holistic and evaluative is atomistic. In informing-validating, informing is atomistic and validating is holistic.
Viewing research through this lens, here is how it looks to me:
Informing research is atomistic pre-design research focused on gathering information about the research participants, which is used to inform strategic decisions concerning requirements.
Generative research is holistic pre-design research, which pursues the perspectives of research participants. The researcher looks for insights that can help a design team intuitively see from the perspective of those who will experience their work. The reason such an approach is generative is that perspective shifts tend to free those who undergo them from inhibiting prejudices and causes new possibilities to become obvious. When this happens research is inspiring (meant in a non-romantic, almost technical sense).
Evaluative research is atomistic mid-post-design research which essentially QA tests a candidate design, to see if the team has designed something that works as intended, and if not, correcting errors and optimizing it. This seems to be the approach more analytical and systematic information architects tend to take.
Validating research is holistic mid-post-design research which treats a design artifact as a hypothesis. The hypothesis takes this form: “If we understand how this person sees, feels and responds to things, this design will be makes sense to them, they will know how to interact with it, and they will respond to in a particular way.”
Notice, the holistic approaches tend to tell us about the person experiencing the design where the atomistic approaches tend to tell us about the thing experienced by the person. And of course, an experience is determined by the interaction of that which is experienced and the one who experiences. This is the first reason I’ve come to see the distinction between informing and generative and validating and evaluative as important. But another reason is even more important.
Nearly everything in the business world — and in fact the entire world — has this dual dimension to it, and is often subjected to reductionistic thought. And the more important the thing is, the more likely this is to happen.
The two examples I have in mind are people and brands. I’ve already discussed people, but business is all about people, not only customers but employees, investors, gossipers, legislators, etc. There’s a strong tendency in institutions to reduce people to statistically behaving masses or an individually behaving entity, and to think of design as a means to modify behaviors. But people are really only understood when the experience that motivates these behaviors are understood as well, and this has practical and moral implications. This topic alone could expand into a multi-volume work.
But brand is significant for being the newest and most conspicuous convergence of holism and atomism at work in the world today. And you see the reductionistic split in approaches to brand strategy. Half the world wants to treat it as entirely irrational, subjective, emotional and subject only to the laws of intuition, at least until the brand is formalized into a brand identity system. And the other half of the world wants to reduce brand to a series of statements and attributes, documented and forgotten, except for the really tangible, enforceable parts, formalized in the brand identity system. About all everyone agrees on is that brand is at least somehow bound up with brand identity systems, and the simple fact that brand is much more than that. But precisely what that “more” is is practically impossible to think about and discuss. Until we learn to think about and talk about what brand really is and how it works, the existence of living, inside-out brands will happen more or less accidentally, and merely formal skin-deep branding will be the norm.