Normative Study

1. Point of View
2. Compressing the opinions
3. Logic of Normative Analysis

Normative research differs from informative studies because the target is not only to gather facts but also to point out in which respects the object of study can be improved. Usually the project even includes planning an approach that could be used to carry out the necessary changes in the object(s). Depending on whether the project continues as practical development or not, there are two styles of normative research:

1. General normative research (fig. on the right) produces theory of design (or of other development) like recommendations, rules, standards, algorithms, advices or other tools for improving the object of study, but includes no practical operations of development. This type of study is discussed under the title Theory of Design.

2. Normative case study (fig. on the left) purports to find out methods to ameliorate physically the object, for example by reducing known human problems in the daily lives and work processes of people, by developing an activity or a new product. When the project includes also carrying out the improvements, it is called "research and development", R&D.

Point of View

Normative research aims at improvements, which means that it includes evaluation of the present state of things and also of the direction of future development. By definition, evaluation is only possible from somebody's point of view. It thus at once becomes necessary to define whose point of view is used in the evaluation.

When assessing any normative proposal which is expected to affect the lives of many people, it helps if these people (which often are quite numerous) can be categorized into distinct groups. For a commercial or industrial company such interest groups exist both inside and outside of the company:

• The company itself, its general management and divisions with their specific points of view, like:
  • the manufacturing division
  • the logistics division
  • the marketing division
• Employees.
• Customers.
• Suppliers.
• Shareholders.

In more general terms, we can categorize the stakeholders of a development project:

• the agents, or the persons involved in producing and implementing the evaluand. These include:
• the developers of the evaluand (the proposal to be evalued)
• the funders
• the decision makers who are determined to utilize or develop the evaluand locally
• the providers of facilities and materials
• the personnel engaged in implementing the evaluand.
• the beneficiaries, those people who somehow profit from the use of the evaluand; these include
• the direct beneficiaries, the "target group," the persons for whom the evaluand was designed, the customers
• indirect beneficiaries, who are in a relation to the above persons and receive a positive influence from them.
• persons who gain from the existence of the evaluand, such as publishers of the materials, contractors who provide services,
• the victims, those persons who are negatively affected by the use of the evaluand; these include
• people that suffer negative side effects like noise, smell or wastes
• people originally intended as beneficiaries, who, however, suffer because of some failure in the evaluand.

The above list (which is abbreviated from Guba and Lincoln) contains people which live now and can participate in the evaluation. However, when assessing products or programs which shall exist a long time and will affect the lives of people not yet living, the list should be completed, e.g.:

• Future Generations
  • Future people who need raw materials and non-replenishable energy (i.e. some of these should be saved for future generations).
  • Future people who have to live with the wastes and pollution, which the various project alternatives often create in different amounts, and which therefore should be included in the evaluation. See also a discussion on the ethical considerations of Applying the Results of Research.

The category of agents contains mostly professionals in the field, such as artists, architects or industrial designers. These are today often organized into associations which may want to play a role in the evaluation, especially when the product is important and it has chances to become an exemplar which later indicates the direction of future design. Of course, the experts are well aware of, and usually wish to support, the needs of average beneficiaries. However, professionals often emphasize topics that are relevant only inside of the art in question, such as 'originality', 'novelty', 'witty deviation from tradition', 'boldness', 'logic' etc. Properties like these are sometimes taken - inside of the profession - as indicators of distinction, i.e. excellence which brings renown and authority to the maestro, especially when the products are published in professional journals, exhibitions etc. Cf. Bourdieu, 1984. All this means that it is usually advisable for a researcher to keep apart the opinions of the professionals and of the users of the products.

The alternatives and available procedures for demarcating a population of study are discussed elsewhere. If the target population turns out to be very large, it may be advisable to pick out a sample of it. All these operations must be done carefully, otherwise you risk getting bias or tendency in your results, which often would be very difficult to remove afterwards.

Compressing the opinions

No matter how meticulously you define the people whose evaluations shall dictate the direction of practical operations, it quite often turns out that these evaluations are more or less in conflict. Evidently each subjective evaluation is based on the person's individual values which in turn depend on the social environment and experiences. In informative study this variation is often welcome and interesting and the normal practice is to collect the opinion of each person. The study can then continue, for example, by finding out which groups of people agree with each opinion, how these groups relate with existing social classes or which are the social and psychological reasons for supporting each opinion.

As a contrast, in normative study which is expected to point out the optimal strategy for subsequent practical operations, a large variation in opinions can be a nuisance because it hinders selecting the goal for practical action.

Variation of opinions may be a problem, but even more problematic can be the task of reducing the opinions to one. Therefore, when it seems that people in the demarcated population have a variety of opinions on the direction of desirable development, the researcher should first find out whether it is really unavoidable to have only one single strategy for the subsequent practical development, or could it perhaps be possible to define two (or more) parallel outcomes (procedures or products), each one for the benefit of a different group of people? For example, in product development it is sometimes possible to create not just a single product but a few slightly different versions of it.

If parallel approaches seem possible, you do not need to compress all the contrasting opinions into one but instead into two (or more) alternative views, which can be easier to do. There are special methods for clustering a collection of opinions, cf. Exploratory Classification, after which you can proceed to defining the collective opinion of each cluster with the methods described below. The work should now be easier because there will be much less variation in each cluster than in the original population.

When circumstances allow only one strategy for realization, the researcher needs methods for condensing the distinct opinions into one collective resultant which then determines the choice of practical actions. In the following there are some approaches which you can use to attain such a consolidation:

• The most frequent view, or voting method.
• An average view is what remains after the random variation has been removed. If the material is given as fixed-choice alternatives, you can calculate the arithmetic mean or the median. If it, instead, is given in free form, you can try to compress all the different views into a summary, leaving out the extreme ones. This task can be quite difficult if the opinions are divergent.
• Acceptance by all can sometimes (but not always) be attained by combining the most stringent requirements in the hope that even the people with more lenient opinions can tolerate the amalgamation. For example, when defining the requirements for a new product, often many needs of special minorities can be included without protests from people who would not themselves need these refinements. In this way the signals that an appliance gives to the user can be made simultaneously visual and audible, which makes the product suitable to the blind and the deaf but preserves its usability for all the others, too.
• Expert opinion or political decision by locally elected representatives of people are sometimes taken as replacement of asking people directly. These methods are easily applied, but their validity in presenting the views of any larger population is usually questionable.
• Conferencing: the people that are being studied are called to a meeting and they are asked to discuss the problem and finally try to agree on a statement either unanimously or by majority. If enough time is available for discussion, people usually try to generate a compromise between several opinions, and often some people are willing to adjust their opinions for reaching a consensus. The consensus is not always quite genuine, because many people are reluctant to express an opinion at a public meeting. If you want to help such people, you can try to follow the procedures of "Democratic Dialogue" presented elsewhere.

Note finally, that the very concept of compressing people's opinions is somewhat crude. A personal opinion is something which a researcher can never profoundly understand, let alone competently modify. That is why it is advisable to use any of the preceding methods with caution. In any case, you should always keep the recordings of the original opinions ready to be consulted in the case of eventual revision.

Logic of Normative Analysis

Sometimes it is possible to carry out the normative process of research and development as a series of simple decisions, for example as follows:

1. defining the target, e.g. removing an existing inconvenience or creating a new product,
2. planning how it shall be done,
3. accepting the plan, etc.

When the normative project deals with complex practical problems the process can become more complicated, and it is often impossible to proceed straight to the synthesis and proposal. Some usual reasons for complications are:

• The outcome is in the future and therefore its detailed formulation takes time. The future context, too, can be difficult to predict.
• Because of the great number of people involved it is difficult to agree about the final outcome, about the methods of achieving it, or about the costs (see above).
• The proposals are made by researchers, planners and designers, i.e. by other people than the evaluators. These often reject essential parts of the proposals. The process then returns to preparing a new proposal.
• People need time for taking their viewpoint and expressing their wishes on the matter. They often want to continue the discussion later, and perhaps need clarification of some details.

Complications like these can compel redoing a part of the work, in other words returning to an earlier stage of the process. If there are many such backward returns the process begins to resemble more a circle than a linear succession of decisions. Indeed, a spiral like the one on the right has sometimes been defined as a typical model of a development project.
Typical phases in a "spiral of development" are as follows.

1. evaluative description of the initial state (perhaps including its earlier development) and defining the need for improvements
2. analysis of relationships and possibilities to change things
3. synthesis: proposal for improvement
4. evaluation of the proposal.

By repeating the sequence from 2 to 4, and by gradually improving the proposal, an acceptable result is usually found. The process is, in other words, iterative.

Iteration means that you make small "incremental" modifications to the proposal, compare it to the old one and continue with whichever of them is best. It is a powerful method, but when using it you should keep in mind two inherent weaknesses of it:

1. The method of iteration cannot really well handle more than one feature of the object at a time. If the alternatives diverge from each other in more than one respect, it will be difficult to compare and rank order them.
2. Another weakness of iteration is that while it usually leads to a better solution, it may nevertheless fail to find the best alternative of all. An example of this is illustrated in the figure on the left: if the iteration process is started at option A, it will eventually lead to alternative C. This is, however, only a partial optimum: while it certainly is better than the neighbours, it is far from the absolute optimum S, which is radically different and could never have been found by iteration only. Because of this weakness it is usually advisable to complement the iterative approach with other, more innovative methods that are capable of producing proposals that are more distant to the already known ones.

Exploiting the advantages of the iterative approach. In a development project it is the final assessment that counts. The proposals of the project usually take shape gradually and in the introductory stage they are just preliminary. Provided that the final testing is done carefully, it is permissible to use a bolder, more innovative style of work during most of the time. It could mean, for example:

• Easier sampling: in the preliminary surveys you can use very small samples, and samples of convenience or of volunteers, that in other contexts are to be avoided because of their indefinite bias. Thus you can, for example, invite "lead customers" such as racing car drivers, which are no typical customers, to participate in the development sessions for a new product.
• Easier surveys: in a questionnaire a high percentage of no-reply is not fatal.
• Bolder proposals. If you do not have enough facts to base your first proposals on, you can simply make guesses. Bad guesses will then be eliminated in the evaluation phase. If you, for example, want to create a product that certainly is different from the competition, you can try to give people something that they do not (yet) know that they need.

Working on a lower level of exactness in the initial phases of the project means not only speed and economy. With less detailing and exactness it is easier to maintain innovation and create alternatives that are not too close to the ones existing and known today.

First in the final phase of planning the proposals should be evaluated as carefully as possible, if possible by appropriate random samples selected from the future users of the results. The final evaluation will have better validity already because the final proposals are more detailed and more realistic.

How to evaluate a thing in the future? The proposals of a R&D project are something which shall exist in the future, while the wishes that you can collect from people today, exist now. It is quite possible that people will have other wishes after those months or years that it takes to realize the new development. How to estimate the difference?

There is not much that you can do to compensate for the temporal difference, but one thing is to make the final evaluation as late as possible, in other words as near as possible to the date of realization of the proposals. In this way the evaluation will be better grounded.

Another method is to extrapolate and suppose that e.g. the rising trend in the standard of living will continue and people will constantly buy larger dwellings, heavier cars and more automatic machines. The method is uncertain, because the development of fashions and lifestyles can make jumps that are difficult to predict.

Examples of complete research and development processes are given on separate www-pages:

• Developing an Activity, including among others the Action Research approach,
• Artistic Research, an approach that combines some scientific methods and creation of a work of art,
• Developing an Industrial Product.

Beside the complete research and development projects which aim at immediate improvements, there are research projects that do not include the operations themselves but instead aim at developing Design Theory, which means that the project produces recommendations, rules, standards, algorithms, advices or other tools for practical operations but does not carry out the operations in reality. There are two usual types of design theory (cf. Paradigms Of Design Theory):

• product-specific theories, e.g. handbooks and other bases of knowledge for designers of a certain type of product. As an example of this type of study there is an account of current theory of furniture on a separate page. Besides, you might want to have a look on a historical review on the theories of architecture from Vitruve until postmodernism.
• goal-specific theories can help you to attain a certain type of target when you are designing a product of any sort. Of these theories, the following ones have specific www-pages:
  • usability, utility, function
  • beauty, aesthetics
  • meaning, message
  • ecology, impact to environment
  • economy, value, price
  • safety.

There are also separate pages which explain some typical sub-routines or limited procedures that are often used in normative studies:

• Normative Case Study
• Normative Comparison
• Normative Classification
• Normative Study of Variables
• Normative Study of Development
• Assessing Practical Output
• The Ethics of Application.

February 19, 2005. Original location: http://www2.uiah.fi/projects/metodi
Comments to the author: