This is an extract from Chapter 6 of Eric Jantsch's Technological Planning and Social Futures. For more on forecasting techniques and methods the reader is invited to refer to Eric Jantsch's Technological Forecasting in Perspective, 1967.
This extract on forecasting is here included because of the importance of being able to assess various future trends and to which one to give preferential treatment. In so doing we can prepare the way for a willed future instead of being forcefully channelled towards an unpalatable one.
It is a common misunderstanding that the use of techniques, or formalised approaches in general, should distinguish forecasting from mere speculation. Much of good forecasting is done without the explicit use of techniques. Techniques just serve to augment the capability of the forecaster, and, in general, follow the basic thinking procedures which the human brain is applying intuitively. Most of them have been designed for a subtle ‘man-technique’ dialogue and are very sensitive to man's knowledge and his capacity for imaginative thinking, technical and value judgment, and synthesis.
The most important contributions of special techniques related to forecasting may be summarized in three points :
-They elucidate the role of individual input factors, compel a comprehensive consideration of such factors, and assure some homogeneity in the results.
-They tend to reduce prejudice and bias.
-They permit the evaluation of vast amounts and complicated patterns of input information and facilitate the systematic evaluation of alternatives.
If forecasting exercises are to relate in a meaningful way to corporate planning, they have to employ a variety of approaches and combine them in ways which depend on the nature of the forecasting task. Both exploratory and normative thinking, and thus the use of techniques belonging to both ‘directions’, are necessary for a complete forecasting exercise. Simple techniques such as trend extrapolation, or scenario writing, may be used to generate information which is subsequently structured by other techniques and ‘processed’ for use in planning by still other procedures.
For the purpose of selecting the proper techniques, it might be of help to classify approaches to forecasting by their output - do they generate new information (which had not been explicitly on the table, although its elements might have been in people's minds), or do they simulate the use of this information? Again, the basic distinction between exploratory and normative approaches has to be made here. From this angle of view, forecasting techniques - explained here in terms of technological forecasting – may be generally grouped in the following way:
1. Techniques related to exploratory forecasting, and all kinds of formalised approaches in the same direction, discover two important tasks: the generation of new information about future technological systems and their performance, and the simulation of the various outcomes of the realisation of (technological) options in the context of a diversity of possible situations.
1.1. The generation of new information may be subdivided into extrapolative forecasts (where do trends lead to under linear or contingent assumptions?) and speculative forecasts (what is the range of alternatives?)
1.1.1. Extrapolative forecasting techniques are mainly based on trend extrapolation and its refinements, of which the envelope curve technique is of particular interest.
1.1.2. Speculative forecasting techniques have achieved some sophistication in techniques for improving group consensus or intuitive thinking – ranging from brainstorming to the Delphi technique – and in morphological analysis, systematically exploring all combinations of the qualitative variations the basic parameters of a concept (technological or other) can undergo, thereby bringing out the potential of new combinations.
1.2. The simulation of outcomes of the realisation of options in various systemic contexts finds a variety of techniques, including learning curves, gaming, input/output analysis, multivariate and structural models, scenario writing and cross-impact analysis.
2. Techniques related to normative forecasting, and formalised approaches in the same direction, also find two major tasks, again the generation of new information - but this time about needs, desires, values, functional requirements, and structural relationships - and the simulation of the implications of overall objectives (policies), goals (strategies) and specific operational targets in various systemic contexts.
2.1. The generation of new information may be subdivided into speculative techniques (what norms and what goals should we introduce into the planning process?) and structural techniques (what are the future relationships as affected by action we may take?)
2.1.1. Speculative forecasting techniques, in the normative direction, may again make use of improving group consensus through the Delphi technique.
2.1.2. Structural forecasting techniques have found their most elaborate example in relevance trees. Simpler applications of decision theory, such as decision matrices, are also in use, as well as network approaches to reasonably well-perceived goals. More recently, cross-impact analysis (which is also practised in a predominantly exploratory mood) has been developed as a means to structure and 'harmonise' future relationships in systemic contexts.
2.2. The simulation of implications of objectives and goals for action in the present again uses some of the structural approaches outlined above, such as relevance trees (in particular, in their numerical versions), all sorts of matrices or other simple procedures for priority ranking and rational resource allocation, usually basing on operations research and decision theory, dynamic modelling, occasionally game theory, and aspects of systems analysis. The aim of all these approaches is to guide the structuring of thinking by simulating the mutual consequences implied in the relationship between preconceived goals and recognised technologies or research elements. The criteria for investigations along these lines are generally preconceived, too.