Translation of user requirements into house designs: A multi-dimensional scaling analysis

An experiment was conducted to determine how user requirements for single-storey houses are translated into actual design layouts. Architectural students were asked to provide floorplans of houses for four quite different hypothetical users. The briefs contained information either about spatial requirements or about life-style characteristics. Another group of architectural students later evaluated the appropriateness of the designs for the intended user. An INDSCAL analysis of the data (inter-room distances) produced a 4-dimensional solution that accounted for 65% of the variance between house designs. The salience of the dimensions seemed to be more a function of the designer than of the requirements of the user.It was found that when users' requirements are expressed in spatial terms only, the designs produced are clearly a function of the user, not the designer. The more information supplied about the user, the greater the tendency for congruence in the designs. When the brief consisted of life-style characteristics only, the results remained equivocal. There appeared to be little consensus on what designs were appropriate for the hypothetical users.The experiment strongly suggested that (a) architecture students handle information about spatial requirements with ease and with a good deal of consensus on what they consider to be the appropriate design solution for each user; (b) information about life-style characteristics and activity profiles, on the other hand, is handled quite differently by different students. In the latter case, the house designs produced probably follow implicit design-behaviour theories that are quite idiosyncratic. A number of theoretical and methodological issues that require further investigation are raised in the discussion.