DELPHI TECHNIQUE FOR TECHNOLOGY FORECASTING
An objective forecast should have a sound logical framework to ensure its repeatability. Formal forecasting techniques are used to replace subjective opinion with objective data using a replicable method. Even though objective technique’s are desirable, there are three situations in which expert opinion (subjective or intuitive method) may be resorted to
i) when there is no historical data, especially situations in which new technologies are involved; expert opinion is the only possible source for a forecast,
ii) when impact of external factors is more important than the factors that governed the previous development of the technology. Under this condition, forecast using past data is irrelevant,
iii) when ethical considerations rather than technical and economic considerations govern the development of a technology.
Under such conditions a Committee approach is normally resorted to. The Committee approach has the following advantages and disadvantages.
i) The sum total of the information available to a group is at least as great as that available to any one individual.
ii) Committee can bring in interactive aspects which a single member cannot contribute.
i) Minority views are normally neglected if, either the group is highly bureaucratic or, consists of individuals from diverse disciplines.
ii) In a group the tendency is to agree with the majority even when the individual feels that the majority is wrong.
iii) A group has its own way of functioning i.e. it has its own internal dynamics. Reaching an acceptable goal becomes an end in itself. Group members may have vested interests in certain outcomes or a common bias and this can affect the results.
To overcome the disadvantages of the Committee approach and to make use of the advantages inherent in the committee, a method known as Delphi has been developed.
Delphi is a ‘programmed, sequential questionnaire’ approach. (The word has its origin in `Oracle of Delphi’ in which a group of learned persons used to make forecasts.) Instead of an individual making a forecast, a group prepares a forecast but with certain characteristics; (a) anonymity, (b) iteration with controlled feedback and (c) statistical group response taking into consideration minority views. The detailed steps involved in a Delphi exercise are as follows:
Step 1: Identify the specific area or field in which a Delphi exercise has to be carried out.
Step 2 : Identify a set of users, technology generators/experts, equipment manufacturers, development bankers, and social scientists and others who can help in preparing a set of questions for forecasting technological developments in the given area.
Step 3 : A small core group is formed to prepare the questionnaire with the help of persons at (2) to cover among other aspects the desirability of the technological development, its feasibility, time frame and resources needed for commercial use, and impact. Background data may be included to reduce subjectivity.
Step 4 : The questionnaire is then administered to a number of participants to cover a wide cross-section of interests.
Step 5 : The first round responses for the questionnaire is obtained and processed, wherein minority and even extreme views etc. are also included.
Step 6 : The second round questionnaire, containing the processed responses of the first round questionnaire, is sent back to the participants to give their revised comments.
Step 7 : The second round results are then processed to get the consensus results.
Thus Delphi provides a consensus view based on the opinion of a large number of participants and can be considered as a way of combining/integrating forecasts.
The Delphi method is used to :
a) identify new factors likely to influence the future state of technological development,
b) obtain probabilistic estimates of technological performance over a specified time horizon,
c) obtain forecasts of a time scale for an event where other methods cannot be used,
d) ascertain the feasibility of a given event occurring under specified conditions,
e) obtain subjective quantitative measures of technological performance in the absence of objective data.
The preparation of the questionnaire has to be done very carefully and should, among other conditions, ensure that:
a) The issues are unambiguously specified
b) The questions are unconditional and not interlinked to conditionality specified elsewhere
c) The definition of probability for occurrence of an event etc., sought in a question should be clearly defined/explained.
Delphi is a useful technique for (a) obtaining forecasts when there is a limited amount of historical data, and (b) for fields which are highly interactive and interdisciplinary involving diverse parameters such as social, technical, economic, political and managerial.
a) It does not have any ‘logic’ underlying each prediction and if repeated, it may not give reproducible results; and
b) Although it may produce a high degree of convergence, yet this convergence does not imply a high degree of reliability.