For a good design, it is necessary to know the problem to be solved by the design, to know the solution to the problem, and to know whether the solution was appropriate or not so that it can be utilized in the next design. All three of these are related to the psychology and physiology of the user. The subjective evaluation is answered by words (including numerical values), but since it is not easy to grasp these in a reproducible form, various efforts have been made, including how to analyze them.

For example, if the size or shape of a window changes in a certain space, how much space is perceived by people will also change. To some extent, there is a common tendency for people to feel this way. There is also a common tendency in how the impression of a space changes depending on the color of the walls. On the other hand, each individual may have a different tendency as to which color they prefer more. To grasp such subjective evaluations and accumulate them as knowledge that can be applied to design, it is necessary to have a method that can grasp not only impressions but also psychological quantities.

The history of subjective evaluation methods goes back to the 19th century. The earliest method of grasping psychological quantity was proposed by G.T. Fechner in Germany. He is regarded as the founder of psychophysics, the study of the relationship between stimuli and sensations. He proposed the method of limits, the method of constant stimuli, and the method of adjustment for measuring psychophysical quantities such as stimulus thresholds, difference thresholds, and the point of subjective equality. The psychophysical scaling method is to construct a sensory scale based on these measurement methods. Once sensory scales related to the intensity of sensation and stimulus, such as light intensity and brightness, sound pressure and loudness, etc., were constructed and became known as the Weber–Fechner law, subjective evaluation began to aim for higher-order evaluations that include cognitive content that can be called more subjective, i.e., evaluations based on the evaluator’s personal values. For this reason, psychophysical measurement methods came to be treated as psychological experimental methods rather than subjective evaluation methods.

Magnitude estimation (ME) is a method devised by S. S. Stevens, an American scientist in the 20th century, and is a representative example of a subjective evaluation method that is still applied to a variety of evaluations. The ME method, which is a direct ratio scaling, led to Stevens’ law, which states that the magnitude of a sensation is proportional to the power of the magnitude of the stimulus.

Another representative subjective evaluation method is the semantic differential (SD) technique. SD was devised by psychologist C. E. Osgood as a way to objectively measure the subjective meaning (positioning) of concepts (adjectives). Using a large number of adjective pairs with opposite meanings, such as light and dark, hard and soft, etc., the subject was asked to rate the object on a scale of generally 5 to 7. By analyzing the factors, he showed that the potential semantic concepts that we use to describe things are based on the three axes of evaluation, potency, and activity.

Currently, “subjective evaluation” refers to the evaluation of impressions and preferences regarding an object based on the subjectivity of the individual evaluator, and is sometimes called impression evaluation. By contrast, “subjective evaluation method” refers to the method regarding how the evaluator answers the evaluation in the subjective evaluation process, and often does not include the way of presenting the objects of evaluation. The words and concepts used in the subjective evaluation are called “subjective evaluation scale,” and they are answered in the subjective evaluation method in a stepwise or non-stepwise manner, either by filling in the evaluation form or orally.

Responses using the subjective evaluation method are often quantified and analyzed. In the case of the ME method, the subject judges the ratio of the amount of sensation between the standard stimulus and the comparison stimulus, and answers directly as a numerical value in comparison to the one specified for the standard stimulus. For example, when the subject is asked to evaluate the sense of volume of a certain room, the subject is first asked to remember the feeling of the size of the standard room as 100, and if the subject judges that the room to be evaluated is twice the size, the subject is instructed to answer 200. Then, the participants are asked to move to a room with a different width, depth, and ceiling height, and to give a numerical answer of 150 if they feel that the room is 1.5 times larger, and 50 if it is half the size. This response format is often used for quantitative evaluations, such as the size of a space, because it is easy to conduct quantitative analysis, but it is difficult to apply this format to cases that do not fit quantitative evaluations, such as preferences.

Even if it is not a quantitative evaluation, if the subjective evaluation can be expressed as a psychological continuum, i.e. a continuous numerical value, such as the strength of a subjective impression, it is possible to grasp the evaluation position of the impression on a psychological scale based on the results of impression measurement. This method is called the psychological scaling. The impression measurement procedures used for this purpose include the method of paired comparison, the method of rank-order, and the method of rating scale.

The ME method is suitable for evaluating quantitative measures. For this reason, it is often used to evaluate the sense of volume in various forms of space, the brightness in lighting and lighting design, and the loudness of various types of sound. A somewhat higher-order judgment is the evaluation of spaciousness. Whether the sense of spaciousness in a space with windows is a common measure among people has been verified by model and real space experiments, and it has been confirmed that it can be evaluated by the ME method. Based on these evaluation experiments, prediction formulas and allowable ranges based on four physical quantities, namely the luminance of the sky seen through the window, the room illuminance, the room volume, and the size of the window ( configuration factor), are presented and used for the actual design.

In the SD method, a stepwise evaluation of the paired terms (bipolar scale) is performed. By considering the evaluation steps as equally spaced, a numerical value is given to each step and analysis is conducted. The advantage of the SD method is that the concept of the design can be organized by consolidating the scales used, and the evaluation target can be positioned at the same time using the scales.

For example, during the ME method evaluation experiment of the sense of spaciousness described above, the SD method can be used to evaluate a number of scales related to the atmosphere of the space, such as like/dislike, calmness/unsettledness, etc., to understand how the atmosphere of the space changes along with the sense of spaciousness. Especially since the 1980s, the SD method has been widely used in architectural spaces and urban landscapes, where there are many components and it is difficult to narrow down the parameters for evaluation.

The application of these subjective evaluation methods in the field of design is thought to have been triggered by the interaction between researchers in psychology and architecture. In the background of the application and spread of the SD method to a variety of subjects, the development and spread of computers since the 1970s has made it easier to conduct factor analysis that deals with a large number of scales and objects . Today, with the accumulation of many evaluation cases, the number of scales used is not as large as it was in the beginning, and factor analysis is not necessarily involved in many cases.

In addition, since the SD method was originally devised for grasping semantic concepts that are common to many people, it is not suitable for cases where individual differences in evaluation need to be handled. In addition, criticism has been raised since the late 1980s, especially in Japan, about the arbitrary selection of scales and targets by the experimenter in setting the evaluation frame. Since then, the study of the cognitive structure itself, which is the scale on which individuals make evaluative judgments about various objects, has emerged, and various evaluation methods have been devised according to the purpose of evaluation, mainly based on Kelly’s personal construct theory.

In psychological evaluation for design, it is important to determine how to combine the subject of evaluation, the target of evaluation, and the subjective evaluation method according to the purpose of evaluation. We are entering an era in which the subjective evaluation methods that have been established up to now are being employed with slight modifications depending on the purpose and target of evaluation.

(Naoyuki Oi)

Related Classes

Environmental Design Course Environmental Information I・II

References

• Osgood, C. E., George J. Suci, and Percy H. Tannenbaum(1957) The Measurement of Meaning. Urbana, University of Illinois Press.
• 岩下豊彦（1983）『SD法によるイメージの測定』川島書店
• 日本建築学会編（2011）『住まいとまちをつくるための調査のデザイン』オーム社
• 日本建築学会編（2013）『建築環境心理生理用語集』彰国社