Tired of Committees? Try the Delphi Method

by John L. Wheeler

Here’s a step-by-step account of an alternative to face-to-face meetings.

Have you ever gone home from a tense or frustrating committee meeting and thought that there must be a better way to make decisions? What would the outcome have been if so-and-so had not been so emphatic? That young, new guy seemed to have a good solution, but the chair wouldn’t let him explain. Do we really need to travel for hours and hours just to hold a face-to-face meeting?

A well-established and versatile group survey approach called Delphi can overcome these and other disadvantages of face-to-face committees and, sometimes, replace meetings altogether. It also has the virtue of involving people in decisions that will affect them. The Delphi method has exciting potential for Christian organizations.

There are usually advantages in using groups of people rather than individuals to arrive at priorities, formulate policy, develop budgets, and so on. Normally a group has more information and experience and takes more factors into consideration than an individual would.

Committees, however, have serious drawbacks when it comes to using all available information and arriving at genuinely unbiased decisions. There is often social pressure to agree with the majority, even when some members consider the decision to be wrong. Forceful individuals can sway meetings against the better judgment of the majority. The views of shy, modest, or less articulate people (however well informed) are frequently not duly considered. Martino says some group members may have a vested interest in certain points of view, especially if they have been presented strongly at the outset.1 Such people are usually more interested in winning than in reaching a valid conclusion. For most missionary organizations, a very significant disadvantage is the time and cost of meetings.

Basically, the Delphi method involves a sequence of surveys of a group for members’ opinions about an issue. Each new cycle of questions and answers provides more information until participants obtain a consensus or a clear description of views. Three or four cycles are usually enough. The method, according to Martino, “can be used for any purpose for which a committee can be used.”

In contrast to a face-to-face committee meeting, a Delphi survey is anonymous, proceeds through cycles with controlled feedback, and allows participants to see some statistical measure of the spread of opinion of the group, both during and after the process. Feedback given to the group with each cycle provides new or restated information that clarifies the group’s understanding or develops alternative approaches.

In effect, a Delphi is an almost utopian committee meeting in which all members consider the issues thoughtfully, express their opinions freely, albeit succinctly, but keep silent if they have no significant contribution to add. It is a meeting devoid of digressions, marked by a complete absence of inhibition or provocation of individuals by others.

As Lewis Thomas has put it,

Debating is what committees really do, not thinking. Take away the need for winning points, leading the discussion, protecting one’s face, gaining applause, shouting down opposition, scaring opponents, all that kind of noisy activity and a group of bright people can get down to quiet thought. (Delphi) is a nice idea and I’m glad it works.²

Although I hope this overstates how the committees of most church or mission organizations work, the underlying picture is not unfamiliar. Thus, I see a considerable need to judiciously adapt the Delphi technique in many areas of decision-making and policy development in Christian organizations. Could it promote team spirit, lead to better decisions, or reduce costs in your organization?

Possible uses in Christian organizations
Here are six areas in which you can use Delphis.³

1. Policy development and planning. This includes defining goals, probing constituents for their strategy preferences, and developing budgets. Should a church, as it moves into a new district, target particular groups? If so,which groups and how? What should be our targets over the next decade? How widespread are the objections to our proposed merger? What should be the selection criteria for the next director, and what weighting should be given to each factor? What do missionaries ending their first terms think needs to be taught in orientation courses, and what are the priorities?

2. Conflict resolution. The Delphi technique can help when opinions about policy or strategy emerging from face-to-face discussions have become strongly polarized. It also allows people to explore acknowledged differences. Hartman gives an excellent example of how a Delphi approach resolved a deadlock over curricula in a New Jersey school district.⁴ Twenty policy statements representing key issues dividing the two groups involved were listed. For each, council members were asked whether they (a) agreed, (b) disagreed, or (c) were willing to try the specific policy for a year on a trial basis. Panelists who disagreed were asked to amend the statements. In the next round, the vote totals were reported and respondents voted on the amended policy statements. This round resulted in a clear consensus.

3. Priority setting. Delphis can establish the criteria for setting priorities and then establish the priorities for organizational goals and activities. What are the criteria for choosing a group to receive a new Bible translation? What ranking should be given to the five objectives of the youth program next year?

4. Values clarification, that is, understanding more precisely the attitudes and expectations of constituents or target groups. For example, how do the financial supporters of a mission board understand stated policies on church planting or rural development?

5. Predictions of events, values, or demands. By what date will 75 percent literacy be achieved in a given people? What is the average inflation rate going to be in Namibia next year? What will be the five best selling books in our stores next quarter?

6. Academic research, such as the “historic Delphi” developed by Strauss and Zeigler.⁵ For example, a Delphi could be conducted among a dozen or so specialists to determine what Adoniram Judson’s views might have been on several current issues in missiology. (Perhaps this is not to be encouraged, as there is still more than enough ancestor worship in most mission agencies!) More useful, perhaps, would be a Delphi among ethnic groups in a multicultural organization to determine their perceptions of the relative heinousness of certain sins (pride, lack of total honesty, financial misappropriation, nepotism) or of possible managerial attributes (drive, enthusiasm, tact, incisiveness, spirituality, listening skills, financial acumen).

THE DELPHI METHOD
1. Definitions. The “administrator,” “moderator,” or “monitor” conducts the survey and writes the report. (This can be more than one person.) The “panel” is the group of people whose opinions are sought, and the “questionnaire” is the form sent to them. (Actually, the term questionnaire is often inappropriate: Material sent to panel members may not be questions but rather a series of statements to be assessed.) The process of seeking a response from panel members, receiving and summarizing replies is a “round.”

2. Usual steps. Although they vary, the following steps are likely. (You may want to provide specific ideas for preliminary prayer and reflection with every communication to panel members.)

a) Appoint or nominate the administrator. There are advantages in objectivity, shared responsibility, and possibly also in time management if two or three people share the task. Apart from integrity, the most important attribute for a Delphi administrator is the ability to sift and summarize information fairly and succinctly.

b) Define the objectives. Decide as specifically as possible the information you want and the purposes for which it will be used. Define specifically the group of people whose responses you want.

c) Select and invite panel members. If you areseeking experts, choose them carefully and objectively. Remember that the “big names” may be too busy to reply in a reasonable time. Experimental evidence indicates that non-experts are as reliable as experts in forecasting, and they may be more innovative.6 When inviting people to participate, clearly explain the process, what you expect from participants, and the safeguards you’re adopting to preserve their anonymity. Provide a realistic estimate of the time it will take them to respond and emphasize that they need to return the questionnaires and complete all the rounds promptly. Between 20 and 50 panel members is usually adequate.

d) Prepare materials for the first round. Some Delphis begin with a blank sheet and an invitation for members to outline goals, policies, selection criteria, budget estimates, and so on. Others may provide an initial list for ranking by panelists, inviting additions that can be incorporated in subsequent questionnaires. Statements worded too concisely usually lead to widely varying interpretations; overly long statements, on the other hand, are difficult to understand and discourage response. Statements with both low and high numbers of words usually discourage consensus; in one study, 20 to 25 words per statement was the optimum.⁷

Ensure each participant’s anonymity by providing a standard blank inner envelope if you want him or her to return the questionnaire by mail. You can use fax machines to transmit questionnaires back and forth provided that you conceal the respondent’s identity.

e) Round 2. For each statement, calculate the mean, the mode (the most frequently given score), and some measure of variation (such as the standard error or the interquartile range, or perhaps simply the numbers voting for each). This is easy if someone enters the responses into a spreadsheet whenever they come in. Then return the statements, estimates, or whatever the Delphi consists of to panel members with these statistics. At this point, you usually remind members how they voted in the previous round. Ask those who differed significantly from the mode to reconsider their responses. If they wish to differ from the emerging consensus, ask them to state briefly their reasons.

f) Round 3 and subsequent rounds. The administrator summarizes the reasons of dissenters for persisting with their views and, after recalculating the statistics, sends out the questionnaire again with the additional information. The process is repeated until a consensus is evident (when there is little change between rounds) or the strength of the two or more differing positions is defined. In most cases, three rounds (or four, if the first is a blank sheet) are sufficient to reach consensus.

g) Final report. This presents the consensus and shows graphically or in other ways the spread of response. Reasons for and against the consensus may be presented. If the prearranged degree of anonymity permits, and you have information on the respondents, it may be possible to describe the characteristics of groups with certain views.

BENEFITS CLAIMED FOR DELPHI⁸
1. Respondents’ sense of participation in decisions. In many situations, Delphi allows all members or staff (or a substantial sample) to contribute freely to the organization’s decisions. People who know their opinions have been taken into account are likely to be more highly motivated to implement a policy than those who perceive it to have been formulated in a remote boardroom. Nonetheless, an unfamiliar, anonymous communication system such as Delphi can become a threat to established individuals and intraorganizational relationships.⁹

2. Consensus without bias. At least five sociopsychological pressures frequently bias the judgment of a face-to-face group. The opinions of the managing director, presiding minister, or other high-status person may receive more time and carry more weight than other equally well-informed views. Talkative, highly motivated, or persuasive speakers can, and frequently do, sway meetings. Groups tend to develop amindset in their approaches to problems and to resent or discourage other approaches. Individuals may be inhibited, or encouraged, by how they perceive their own expertise in relation to the group. There is usually group pressure, spoken or unspoken, real or perceived, on minorities (dissidents?) to go along with the opinions or decision of the group: Many “minority reports” are conceived, but few actually are born and come to public knowledge. Anonymous Delphis remove most of these biases against genuine response.

3. More informed decision making. Controlling the feedback of information helps ensure that all relevant information that individuals possess is available to, and can be objectively considered by, all members. Note, however, that the quality of this information depends on the integrity as well as the intellectual and literary skills of the administrator(s). Because the process is anonymous, panel members who learn additional facts feel freer to change their positions.

4. Economies in time and travel. Many organizations working across a country or internationally hold meetings to decide policy, discuss difficult management decisions, formulate budgets, and so on. Provided that mail or fax services are reasonably reliable, much of this work could be done by Delphi.

5. Exposure of divergent views or uncertainty. Although a Delphi normally produces a consensus, it also reveals and clarifies divergent views. Moreover, the process quantifies the support for the two or more views. Unconventional solutions or opinions, which initially might be viewed as “unspiritual” by some traditions, would seem to have a better chance of being shared with the group using Delphi rather than face-to-face meetings.

OBJECTIONS AND PROBLEMS
In the 25 years or so that people have actively used Delphis, the method has been thoroughly appraised. Linstone has distilled the criticisms into a checklist of pitfalls for Delphi designers.10 Philosophically, some object to the quality of the consensus obtained. They cite the lack of an adversary process and the risk of deliberate or inadvertent bias or manipulation. Also, a Delphi conducted by mail may take months to complete, which eliminates it for some purposes. (Delphis run by facsimile or electronic mail, however, can be completed in a few days.)

Organizations considering using Delphis may experience two problems. Autocrats or those who feel threatened by worker participation or team decision making are unlikely to find the technique appealing. On the other hand, rank-and-file members of organizations, perhaps accustomed to personally briefing their board or committee representatives on their views, or those who enjoy criticizing decisions they have not had to make, may not like actually making decisions. A solution in both cases is to carefully introduce the idea on a couple of minor issues so that people do not feel threatened.

CONCLUSION
An alternative to face-to-face committee meetings exists. Provided that you use it thoughtfully, and with the above safeguards, the Delphi method can be an extremely useful and economic way to get a consensus. The time and volume of work involved for administrators probably mean that, despite its advantages, Delphi will not be practical for everyday business or simple, uncontentious issues. However, people have used Delphi creatively and successfully in hundreds of applications—in education administration, the social sciences, and business. Try it in your field. Although the name may not appeal, I hope the concept behind it will be attractive to organizations that, by definition, are committed to working peacefully, valuing the opinions of individuals, and seeking God’s will in their activities.

Endnotes
1. Joseph P. Martino, Technological Forecasting for Decision Making (New York: North Holland, 1983 2nd ed.).
2. Lewis Thomas, The Medusa and the Snail (New York: Bantam Books, 1988). Quoted by Hartman, op. cit. sub.
3. Norman Hale, “Problem solving techniques for administrators,” ACSA School ManagementDigest, Series 1,No. 11(Burlingame, Calif.: Association of California School Administrators, 1978). Hale gives, in a very readable discussion of the Delphi technique, other examples of the use of Delphis in education administration. Modified Delphis have been used before and during conferences, alone, or in conjunction with face-to-face meetings.
4. Adele Hartman, “Reaching consensus using the Delphi technique.” Educational Leadership, 38 (1981), pp. 495-97.
5. H. L. Strauss and L. H. Zeigler, “The Delphi technique and its uses in social science research.” Journal of Creative Behavior, 9 (1975).
6. Peter W. O’Brien, “The Delphi technique, a review of research.” South Australian Journal of Education Research, 1 (1978), pp. 57-75.
7. Harold A. Linstone and Murray Turoff, eds., The Delphi Method: Techniques and Applications. (Reading, Mass.: Addison-Wesley, 1975), p. 232.
8. See also Linstone and Turoff, op. cit. and Hale, op. cit.
9. Linstone and Turoff, op. cit., p. 585.
10. Linstone and Turoff, op. cit., pp. 573-86.

—–

EMQ, Vol. 29, No. 1, pp. 42-48. Copyright © 1993 Evangelism and Missions Information Service (EMIS). All rights reserved. Not to be reproduced or copied in any form without written permission from EMIS.

Comments are closed.