Methods and Approaches of Futures Studies

For all of human history people have tried to develop methods for predicting the future, from reading palms to gazing at the stars. But in recent years, primarily since World War II, scientists, sociologists, operations researchers, and others, many of whom began to call themselves futurists, have developed quantitative and qualitative methods for rationally anticipating the future. What separates futurists from the soothsayers who came before is rationality, an awareness that the future cannot be known with absolute certainty, and the recognition that many different futures are possible, depending on decisions people make in the present.
Generally, methods for studying the future do not pretend to be able to predict the future, although assessing the probabilities of alternative futures is an important aspect of futures studies methods. Rather, futures studies methods are generally designed to help people better understand future possibilities in order to make better decisions today. Futurists often say they use their methods to reduce uncertainty, although it may be more accurate to say they are trying to manage uncertainty. Many decisions must be made today in the face of great uncertainty about what may happen in the future or even what the effects of today's decision might be in the future. Futures methods help people to deal with this uncertainty by clarifying what is known, what can be known, what the likely range of possibilities is, what the most desirable possibilities are, and how today's decisions may play out in each of a variety of possible futures.
Futures research methods are both descriptive and prescriptive. Descriptive methods, sometimes also called extrapolative, attempt to describe objectively what the future will be or could be. Prescriptive methods, also called normative, focus on what the future should be. Prescriptive methods try to help people clarify their values and preferences so they can develop visions of desirable futures. Once they understand what they would like the future to be, they're better able to take the appropriate steps to create that preferred future.
Although much has been learned about futures studies methods since most were developed in the 50s and 60s, they remain somewhat amorphous. One can probably identify as many futures studies methods as there are futurists, as each futurist develops his or her own style for looking ahead. But gradually, some consensus on methodologies is developing.
One principle upon which most futurists would agree is the need to use multiple methods to address most futures problems. One will gain much greater insight by developing a futures research program that combines environmental scanning, trend assessment, delphi, and scenarios, for example, than one could achieve using any single method alone. Thus, although several of the more popular methods are described individually in the pages that follow, they are ideally used in various combinations.
Another principle upon which some consensus is developing is that futures research should be participatory: it should involve stakeholders and decision-makers directly in the process of developing forecasts or creating scenarios, because that is the only way to enable people to fully appreciate and perceive the range of possible futures.
Although many futurist strive for objectivity, ultimately, most futures methods rely heavily on subjective human judgment. But there are various tools one can use to augment individual human judgment. A method's value often lies in amalgamating the judgment of many people, enhancing creativity, generating questions and ideas to produce different judgments, and demonstrating consistencies and inconsistencies among and within competing views of the future.
As discussed in Chapter 2, Principles of Futures Studies, futurists often divide the purposes of futures studies as imagining the possible, assessing the probable, and deciding on the preferable. Most futures studies methods focus on one or two of these goals, but not all three; thus one almost always will need multiple methods if on is to work through the full range of futures studies. For instance, analyzing a present trend will give some information about the possible and the probable, as we analyze what will happen if the trend continues or what may cause the trend to change, but it tells us relatively little about what we like to have happen. Visioning techniques may tell us something about the possible, as we brainstorm a range of alternatives, and the preferable, as we use visioning to imagine preferred futures, but it may tell us relatively little about the probabilities of our preferred futures without the help of other techniques.
Futures studies can also be thought of as encompassing five stages, although many individual projects will focus on one or two stages and leave the rest to other projects. The first stage is to identify and monitor change. The second stage is to critique and analyze change. The third stage is to imagine alternatives. The fourth stage is to envision the preferred alternative. And the fifth and final stage is to plan and implement steps to achieve the preferred vision.1
Most futures studies methods fall into one or two of these stages. For example, environmental scanning, trend analysis, and emerging issues analysis are very important for identifying and monitoring change, and they may also begin to critique change. Cross-impact analysis, technology assessment, and social change assessment are important for critiquing and analyzing change. Scenarios and related methods are prominent in imagining alternatives. Visioning methods are key to envisioning preferred futures. Strategic planning, political activism, and organizational development techniques are key to planning and implementing for the future.
That futurists use multiple methods is illustrated by the results of a recent survey of professional members of the World Future Society. Asked to estimate the percentage of their futurist output that relied on each of 16 popular methods, the futurists showed that no method was dominant in their work. Trend analysis garnered the highest score, accounting for 20 percent of the futurists' output, with scenarios placing second at 14 percent. Visioning and environmental scanning followed, with 11 and 9 percent respectively. The other 12 methods all fell within a relatively narrow band of 2 to 7 percent.2
In addition, there was not a strong consensus among the futurists as to which methods were core to the field -- which methods anyone calling himself a futurist must be familiar with. Only trend analysis and visioning were considered core methods by more than half of the respondents, with 56 and 51 percent respectively. Scenarios was third, with 45 percent considering it to be a core method. However, nearly all the methods were considered very important. When respondents who considered a method to be essential were added to those who considered a method to be very important, only one of the 16 methods, precursor and bellwether analysis, fell below 50 percent, coming in at 36 percent. Eleven methods exceeded 60 percent of respondents ranking them as either essential or very important, with the remaining four exceeding 50 percent.
What does this mean for the aspiring futurist? The authors suggest that beginning futurists should seek a good understanding of trend analysis, scenarios, visioning, and environmental scanning early in their careers. These methods should be augmented based on individual interest by at least two other methods early on. This basic tool box can then be expanded as time and interest dictate, although some futurists will choose to specialize in just one or two methods. This often makes sense as that futurist can gain particular expertise in the method(s) of choice. But one must recognize that futures research will be improved by involving others with expertise in other methods in a particular futures project, and a basic understanding of other methods will aid the futurist in combining his or her work with the work of others.
This introductory text will only give the reader a basic sense of these methods. The reader should consult the resources listed at the back of the chapter to attain a solid grounding in these methods.
Trend Analysis: A Method Everyone Uses
Trend analysis involves the use of any of a variety of techniques based on historical data. Trend analysis involves several processes. One process is spotting an emerging trend, that is, identifying a change in the world around us. For example, you may notice that more and more people seem to be waiting until they are in their thirties to have children. You may have spotted a trend-i.e., that people are delaying child birth. Now you need to do some analysis to see what the nature of the trend is and what its implications might be. You could first look at historical data. What was the average age of women having their first child in 1950? In 1955? and so on. Do you see a pattern? Is the average age of women at the birth of their first child increasing?
You might see the age at birth of the first child is increasing by six months over each five-year interval. That is, perhaps the average age was 21 in 1950, 21.5 in 1955, 22 in 1960, and so on until 1995 when the average age is 26. Then you might extrapolate the trend into the future, to predict that the average age would be 26.5 in 2000 and 27 in the year 2005 and so on. But trend analysis requires that you do more than simply extrapolate the trend forward. You have to ask, what is causing this trend, and will those causes continue indefinitely? Are there upper limits to the trend? What other forces may affect the trend? At this point trend analysis relies more on subjective judgment rather than objective extrapolation of historical data.
Trend extrapolation is the most straight-forward and objective component of trend analysis. Extrapolation essentially consists of taking historical data, fitting a curve to the data, and extending the curve into the future. Trend extrapolation assumes that things will keep changing in the future the way they have been changing in the past. One simply extends the line or the curve forward to predict where things will be at a certain future time.
If the population of a city is known to be increasing at the rate of 2% a year, we assume that it will continue to do so in the future, and we can use simple arithmetic to calculate what the population will be in five years. In other words, we can generate a forecast by observing a change through time in the character of something and projecting (extrapolating) that change into the future. In making a forecast, we naturally disregard short-term changes or fluctuations, such as the swelling of a city's population each morning as people come to work. What is important is the longer-term change, that is, the trend.
Trend extrapolation is one of the most commonly used ways to generate a forecast. City planners, economists, demographers, and many other specialists constantly extrapolate trends -- consciously or unconsciously -- when they think about the future. So, too, do ordinary people. Assuming that the future will be like the past or that past changes will continue in the same direction and rate is a perfectly sensible way to begin trying to understand the future. It can not, however, be the end of our endeavors, or we would end up with absurd results. For example, we might estimate that a child aged four has grown at the rate of five inches a year, and then calculate that this rate of growth means he will be more than 13 feet tall at the age of 34! We would not accept this forecast, because we know that human beings never grow that tall. Long before he reaches the age of 34, we forecast, his rate of growth will slow and eventually halt at a height that will probably be somewhere between five feet and six and a half feet.
This slowing down of growth is frequently encountered among living things: an organism or a colony of bacteria will grow rapidly for a time and then its growth will slow and eventually stop. If growth did not stop, the organism or bacteria would eventually become bigger than the world itself -- and extrapolating still further, bigger than the solar system and universe. But growth does not continue indefinitely; eventually it slows because of limits either in the environment or in the organism itself. This type of growth-curve is generally called an S curve, because the shape of the curve resembles the letter S. See Figure ___.
Growth curves may also be observed in social phenomena. For example, during the 1950s and 1960s scientific research grew very rapidly in the United States. In fact, if the percentage of the U.S. population trained as scientists were to increase in the years ahead as fast as it did during the 1950s and 1960s, there would soon be more scientists than people! Of course, this theoretical limit to the production of scientists will never be reached. Long before then, the production of scientists will be curbed by such factors as the reluctance of taxpayers to pay the ever-increasing costs of scientific research and the exhaustion of the pool of people who might become scientists.
Similarly, a new technology may exhibit a growth curve strikingly similar to those found in biology. When railroads were first developed in the early 19th century, they found many customers eager for rapid, inexpensive transportation. Railroads developed rapidly during the 19th century, but in the 20th their growth slowed, because they began to saturate their market and because new competition appeared in the form of automobiles, trucks, and airplanes.
When growth is plotted on a graph, the resulting curve often looks like an elongated S, which means that growth is slow at first, then becomes rapid, then slows again. Why? In the case of a biological system, growth normally begins with only a small base -- a single fertilized egg or a plant spore. Even if the cell grows and divides quickly, the first doubling in the number of cells only results in one additional cell. The second doubling adds only two more cells, and the third doubling only four. But as the base of cells grows larger, the number of cells added with each successive division grows larger, and growth is very rapid. Eventually, however, the growth encounters limitations, such as the exhaustion of the nutrients in the environment. Animal populations often increase rapidly until the animals have eaten up most of the food in their environment; at that point, the animals starve and the population shrinks. Somewhat similarly, a new technology or business organization generally starts with a small base -- little capital and few customers -- but grows more rapidly as banks and customers become confident and provide more capital and business. Eventually, however, the market is saturated, competition develops, or problems occur, and growth slows or ceases.
But sometimes growth does not follow the pattern we expect. Just when we expect growth to slow -- or even when it has actually started to slow -- it suddenly picks up again. Such situations have occurred in the history of technology, and when we consider the nature of technological development, we can understand why. In most technologies, many technical approaches are tried. Each approach encounters limitations, but just when the technology itself might seem to be encountering an insurmountable obstacle, a new technique arrives to keep the technology improving. For example, the top speed of aircraft has increased quite steadily during the 20th century despite the fact that the aircraft engineers kept encountering factors that seemed to make higher speed impossible. Almost as soon as a barrier was identified, a way around it was found, and airplanes got faster and faster. Fabric-and-wood airplanes gave way to all-metal craft; the open cockpit gave way to the enclosed cockpit, and the reciprocating engines gave way to jets. But forecasters of technology must be wary of assuming that the growth trend will always be saved by a new breakthrough; at some point, the curve will begin to bend over: Airplane speed will no longer increase so rapidly and may not increase at all. Interestingly -- and typical of futurics in general -- the forces that will change the trend in technology will probably not come from technology itself, but from other fields -- politics, say, or the natural environment -- because technology is shaped by social forces, which determine what technology is desirable or acceptable. These social forces influence the economic and managerial resources required to develop technology.
Basic trend extrapolation is an objective method. It relies on historical facts. The rate of change used to forecast future growth is determined by the rate of change in the past. It is not dependent on the subjective judgment of the forecaster. In contrast, qualitative trend analysis demands the futurist to use his or her judgment. "The identification and characterization of a given trend is itself a semi-empirical, semi-creative activity," says Joe Coates. "To identify trends [at least a decade ahead] one has to: develop a conceptual framework of the forces at play; identify what is known and unknown about them; look for theoretical constructs that shed light on those forces; seek out any information that may be relevant; and finally, one must come to the creation of the alternative futures implied by the examination of that system."3 Qualitative trend analysis focuses on areas that are more difficult to quantify but in which important changes seem to be occurring.
Trend extrapolation, while useful, is severely limited because it cannot address changes in the historical pattern of change. Straight-line trend extrapolation is one of most common causes of errors in forecasting. To avoid this common mistake, the futurist has to think about what might cause the patterns of change to speed up, slow down, or change direction.
Cyclical Pattern Analysis
Closely related to trend analysis is cyclical pattern analysis. Many phenomena appear to operate on cycles, and cyclical pattern analysis uses cyclic or recurring patterns (also referred to as waves, warps, bursts, surges, epochs, and episodes) as templates for anticipating future developments in various areas, such as public policy, the economy, etc. The "business cycle" is probably the best known example of this, in which a recession is followed by recovery, which leads to over-expansion of capacity, which in turn leads back to recession, and the cycle begins again. A similar, though much longer-range cycle, was proposed by Russian economist N.D. Kondratieff, who hypothesized that Western societies cycle through a pattern of long waves, characterized by recession-depression-revival-prosperity. The length of the overall cycle averages 56 years, with peaks in the occurring in 1800, 1856, 1916, and 1969.4 The Kondratieff Wave attracted great attention in the mid-1980s, when the cycle predicted depression, but has attracted less attention recently.
Other cycles futurists have explored include product life cycles, historical cycles, and generational cycles.
Environmental Scanning
Environmental scanning refers to the process of scanning the media (typically print media) to identify emerging issues to enable organizations or individuals to anticipate and respond to changes in the external environment.5 Scanning is meant to provide strategic intelligence to the strategic planning process by identifying changing trends and potential developments, monitoring them, forecasting their future pattern and assessing their impacts.6
"The objective of scanning is to look over the widest range of possible factors and to identify connections with the organization's function or business, and especially to identify the significant positive or negative effects those could have on the organization and its activities. In general, the objectives in monitoring and scanning are to:
* detect scientific, technical, economic, social, political and ecological events and other elements important to the company;
* define the potential threats or opportunities or major potential changes for the organization that are implied by those events;
* provide continuous awareness and evaluation of trends to guide planning and action choices;
* inform management and staff of the need for anticipatory action; minimize reaction; stimulate proaction;
* alert management and staff to trends which are converging, diverging, speeding up, slowing down, or interacting.7
Scanning may be active or passive. "Passive scanning is what most people do when they read journals or newspapers," writes James Morrison, an expert in and proponent of scanning.8 Active scanning is a more deliberate and conscious effort to review information from a broad array sources and subject areas.
Scanning projects run the gamut from extremely small, informal programs that may rely on one person's haphazard review of magazines at a newsstand to formalized programs that may involve a score or more people who systematically review a set of periodicals. For example, the State of Hawaii's scanning project involves a 14-member voluntary group (half of whom work in the Planning Office) who subscribe to several dozen periodicals. Volunteers are assigned periodicals in which they search for items indicating a trend, an innovation, or a newly forming issue. Items they find are written up for a monthly meeting, after which the coordinators prepare a monthly report. In turn, this leads to a quarterly newsletter, Future Wave, sent to legislators, state managers, and the general public.9 A number of futures consultancies run proprietary scanning projects for their clients. Future Survey, a monthly journal of the World Future Society, is essentially the result of a scanning project.
John Naisbitt's Megatrends and related publications present another example of scanning, combined with another social science research tool, content analysis. To produce Megatrends, Naisbitt's staff monitored 6,000 newspapers each month to pinpoint and evaluate important issues and trends. Applying content analysis, the researchers counted the column inches devoted to particular issues and tracked the trends in coverage of particular issues.

Scenarios: Making Up Stories About the Future
Scenario planning is the use of internally consistent narrative descriptions of possible states of affairs or development in the future. Usually, alternative scenarios are developed in order to allow people to conceptualize alternative futures and to clarify possible consequences of present developments and decisions.
A scenario is simply a series of events that we imagine happening in the future. Our everyday thinking is filled with little ventures into the mysterious world of tomorrow, or next week, or next year. And these ventures are scenarios, though rarely as well developed as the elaborate scenarios prepared by professional researchers working for government agencies, the military, and commercial enterprises.
A scenario begins when we ask, "What would happen if such-and-such occurred?" For example, "What would happen if we went to the theater on Saturday night?" Once this question is posed, we can begin to imagine the various consequences of the event. First, certain preparations would be necessary for this event to occur; for example, there would be the need for transportation to the theater. In addition, if the event does occur, there will be additional consequences, such as being absent from home at a time when we anticipate that a relative might come. In our minds, we may develop a large number of scenarios in an effort to decide whether or not to go to the theater on Saturday night. We develop these scenarios intuitively and rarely bother to write them down. We may, however, discuss them with each other and with friends.
What does a scenario do for us?
Fundamentally, scenarios are tools for ordering our perceptions about alternative futures in which today's decisions may play out. First, it makes us aware of potential problems that might occur if we were to take the proposed action. We can then (1) abandon the proposed action or (2) prepare to take precautions that will minimize the problems that might result.
Second, the scenario gives us an opportunity to escape from a potentially disastrous action -- or to realize a tremendous opportunity. Either eventuality may be tentatively identified by developing a number of scenarios. For example, as we develop a scenario we begin to think about how to get to the theater and how to get back. As we review in our minds the various alternative means of transportation, we recognize that the brakes on our automobile are defective. If we take the automobile to the repair shop today, it will be ready in time for our excursion to the theater. Otherwise, we might find ourselves using it on Saturday despite its unsafe condition, and possibly having a fatal accident.
Third, the scenario can mobilize others to get them involved in assessing a situation and planning action. People tend to become more involved in a situation when they are faced with a concrete choice. At that point, they must think about consequences and are led into the various aspects of the problem. Some writers have used the scenario as a useful technique to get people to focus on a certain problem. "The test of a good scenario," says Peter Schwartz, a leading scenario planner, "is not whether it portrays the future accurately but whether it helps an organization to learn and adapt."10
Scenarios should identify key driving forces and should be tailored to those who make decisions. For example, complex methodologies should be avoided, as they discourage involvement.11 Although it is possible for an individual to develop useful scenarios, current scenario planning usually involves a participatory workshop that directly involves the decision-makers. This is because scenario planning is a learning tool, in which the process used to create the scenario has at least as much and probably more to do with helping an organization learn and adapt than the final scenario product.
According to Schwartz, scenario planning follows systematic and recognizable stages. The first stage is to, isolate the decision to be made and to rigorously challenge the mental maps that shape one's perceptions. The second stage is to identify and prioritize driving forces, such as technology, economics, politics, etc. In this process one must identify predetermined elements -- those things that are almost certainly inevitable -- and one must identify the critical uncertainties -- those items that are important to the decision but extremely difficult to predict. Predetermined elements might be the location of a factory or a strict environmental regulation, assuming your analysis shows these elements are extremely unlikely to turn out differently from your projections. Uncertainties could be anything from the Dow Jones Industrial Average to the most popular television show in the year 2015, if these factors are important to the decision.
With these inputs, one can construct plausible scenario plots. These will typically be an extrapolative, business-as-usual scenario buttressed by other plausible scenarios. In the final stage, one identifies the implications of each scenario and tests possible decisions against the scenarios. If we chose X, what happens under each scenario? What happens if we choose Y?
Backcasting
A method closely related to scenarios is backcasting. Backcasting is concerned with how desirable futures can be created, rather than what futures are likely to occur. In backcasting, one envisions a desired future endpoint, and then works backward to determine what policy measures would be required to achieve such a future. Backcasting involves six steps: determine objective, specify goals and constraints, describe the present system, specify exogenous variables, undertake scenario analysis, and undertake impact analysis.12 The end result of a backcasting study is alternative images of the future, thoroughly analyzed as their feasibility and consequences.13
Visioning
Visioning has become one of the most popular and important futures studies methods, and a wide range of futurists have developed particularized techniques to help people develop their vision of a desirable future for themselves, their organization, or their community. (Visioning on larger scales, such as national or global scales, remains relatively undeveloped.) Generally, a visioning process will attempt to identify sources of pleasure and dismay in the past and present, will challenge people's current assumptions, will give people a sense of current drivers of change so they can imagine a range of alternative futures, and facilitates a process of achieving some consensus of a preferred vision for the future. "Visioning is a process of making images of the future sufficiently real and compelling to act as 'magnets,' or goals to achieve, or 'spurs' to present action. Visioning can be done by an individual, but it much more frequently takes place in futures workshops,"14 writes Australian futurist Richard Slaughter.
For example, Clem Bezold, who has been developing vision methods since the early 1980s, identifies five stages in building a vision: 1) identification of problems, 2) identification past successes 3) identification of future desires; 4) identification of measurable goals; and 5) identification of resources to achieve those goals.15
"If we can articulate what we want clearly enough, we will be better able to invent and create the future we most desire (our 'preferred' future)," says Bezold. "A preferred future encompasses our ideals (usually in the form of a vision statement or description) and our sense of the best outcome that might be achievable. A vision is a compelling, inspiring statement of the preferred future that the authors and those who subscribe to the vision want to create."16
The visioning concept owes a heavy debt to the future workshop developed by Robert Jungk. Jungk describes the future workshop as follows:
Typically, a future workshop can be divided into a preparatory phase and three workshop phases. The preparatory phase involves deciding on the topic and making the practical arrangements . . . The workshop itself begins with the critique phase, during which all the grievances and negative experiences related to the chosen topic are brought into the open. There then follows the fantasy phase, in which the participants come up with ideas in response to the problems, and with their desires, fantasies and alternative views. A selection is made of the most interesting notions and small working groups develop them into solutions and outline projects. The workshop concludes with the implementation phase, coming back down into the present with its power structures and constraints. It is at this stage that participants critically assess the chances of getting their projects implemented; identifying the obstacles and imaginatively seeking ways round them so as to draw up a plan of action.17
Jim Dator, another long-time expert of the visioning method, has modified Jungk's method in several ways, most significantly in emphasizing the role of the futurist in helping people think more broadly about alternative futures. "I think it is a serious mistake to ask people to engage in any kind of preferred futures envisioning exercise until they have first been challenged to examine their own ideas about the future," says Dator. "One part of the futurist's role is to present, in a dramatic, engaging way, some of the elements, forces or components in the past and present that might significantly influence the future."18
Delphi
Delphi is a method of soliciting and aggregating individual opinions or judgments, typically of a group of experts, to arrive at consensus views concerning such things as what may happen in the future. The Delphi technique keeps individual responses anonymous so that social influences (prestige of a certain participant, shyness of certain participants, etc.) are minimized, and poses the questions in a series of rounds. The results of each round are organized and presented to the participants in a carefully structured way.
One of the earliest formal futures methods, Delphi was invented in 1953 by Olaf Helmer and Norman Dalkey during a RAND study of the effects of a massive atomic attack on the United States.19 Being one of the oldest methods, Delphi is also one of the most used and most studied, though its use in recent years appears to be declining.
According to futurist and sociologist Wendell Bell, Delphi generally includes at least eight steps:
1. The specification of some topic or subject whose possible, probable, and preferable futures are to be investigated.
2. The construction of a questionnaire as an instrument of data collection.
3. The selection of some individuals (respondents) whose opinions are to be studied, usually experts on the topic being investigated.
4. The initial measurement of the opinions of the respondents by means of a questionnaire.
5. The preliminary organization and summery of the data resulting from the initial measurement.
6. The communication of the results of the initial measurement of opinions as feedback to all the respondents
7. A re-measurement of the opinions of the respondents as they have been informed and may have been changed by their knowledge of earlier results including of other respondents' supporting comments for their opinions.
8. An analysis, interpretation, and presentation of the data and the writing of a final report.20
Typically, as a Delphi study proceeds through several rounds, the feedback of result from earlier rounds tends to produce more similarity of opinion.21
While Delphi is one of the best known futures methods and has often been used, perhaps most successfully in Japan, it has come under increasing criticism in recent years. One review of Delphi from the early 1990s found that "the main claim of Delphi -- to remove the negative effects of unstructured, direct interaction -- cannot be substantiated," the authors wrote. "No evidence was found to support the view that Delphi is more accurate than other (simpler, faster, and cheaper) judgment methods, or that consensus in a Delphi is achieved by disseminating information to all participants. Rather, consensus is achieved mainly by group pressure to conform."22 Ted Gordon finds that "Delphi is expensive, cumbersome, misleading, and misused." He now prefers in-depth interviewing, "which is expensive but has an element of productivity that one does not find in a Delphi."23
Others have also critiqued Delphi: "A Delphi can sometimes confer a sense of methodological rigor that really is not there. A poorly designed Delphi survey is often dressed up and made to look credible. Its strength in reaching a consensus is tempered by its weakness in neglecting or downplaying the outlier or fringe ideas, which often are the most interesting. It is best used as an input to further thinking and analysis, rather than an output or final product," writes futurist Andy Hines.24

Cross-Impact Analysis
Cross-impact analysis is designed to identify the various effects that trends and future events may have on each other. For example, the development of an improved transportation system may reduce the need for better communications -- or vice versa. The structure for accomplishing such an analysis is called a cross-impact matrix, in which trends or specific developments are listed along both the horizontal and vertical axes and there are squares or boxes in which the analyst can note the impacts that two variables have on each other. Analysts then examine the potential interaction between each trend or event and decide whether the occurrence of one will make the occurrence of the other more or less likely, or have no effect.
For example, suppose a community is trying to decide whether to build a new elementary school. First, they identify a number of trends and potential events that are relevant to their decision. They may identify the following trends: the elementary school population is growing at five percent per year; local private school enrollment is increasing faster than public school enrollment; average family income is increasing by 10 percent per year; average test scores in the public schools are increasing faster than the state average. Next they may identify the following potential events: the county may impose a sewer moratorium to slow rapid residential development; the county may implement a voucher system to allow families to use some portion of public funds to pay private school tuition; a major local employer may close or move substantial portions of its operations.
Next the analysts array these trends and event along both axes of a matrix, and try to determine what affect the occurrence of an event or the continuation of a trend will have on the other trends and events. For example, if a major employer relocates, family income would be less likely to continue rising by 10 percent per year. If the county passes a voucher program, private school enrollment would be more likely to continue rising faster than public school enrollment. See Figure ___.
This simplified version of cross-impact analysis forces participants to think through the ways in which various future events may interact and thus can help clarify people's conceptions of the future. In its more sophisticated form, cross-impact also tests the consistency of participants' judgment, by asking them to determine the probability of an event occurring under different circumstances.
The more sophisticated cross-impact exercises seem to have fallen out of favor with many practitioners, because their complexity is both time-consuming and may make it more difficult for decision-makers to understand the results. "Cross-impact has been superseded by 'softer' qualitative descriptions of interactions or by far simpler spreadsheet models,"25 says Roy Amara of the Institute for the Future.
Models, Simulations and Games
Models, including computer and physical models, simulations, which are typically computer simulations, and games, which are generally role-playing exercises, can be used to test the effects of various developments or events on the system being studied.26 War games might be the most familiar of these methods to many people.
Everyone is familiar with models, in the form of miniature replicas of airplanes, ships, and rockets. There are also models of human beings, or dolls, and model towns that children can build with blocks. Generally, a model is something that is smaller and easier to handle than the thing it represents. But since it has many characteristics of the original object, we can learn a lot about the original by studying the model.
We also use mental models everyday, although this is frequently unsystematic and unconscious. Through generalization and abstraction, we construct mental models to imagine how alternative courses of action might play out. For example, if we're planning a dinner a party, we might imagine how the conversation might flow if we seat Betty next to Harry, or if we seat Betty next to Jim.
Futurists often try to improve on everyday mental models by creating more formal models. "A 'formal model' is, like an everyday mental model, a set of generalizations or assumptions about the world, but, unlike an everyday mental model, it is explicitly stated in some form." writes futurist Wendell Bell. "Formal models, compared with every day mental models, have several possible advantages, including greater rigor (e.g., they are explicit and precise), comprehensiveness (e.g., more information can be included), logical coherence (e.g., conclusions can be error-free with the system because strict deducibility is possible), and accessibility (e.g., other people can see what is being done and can examine the procedures for reliability and validity)."27
In order to process more information than the human mind can process, futurists and others often construct computer models. A complex system such as the economy of a city can be simulated by means of a mathematical model, that is, a series of equations showing how different variables affect each other. The equations can then be fed into a computer, and then the computer can be given data representing the situation as it is. At that point, researchers can change one or more of the variables to see what happens.
What would happen, for example, if the real estate tax were raised 10% or the sales tax abolished? If the model is properly designed and operated, it can allow people to anticipate the effects of a policy before actually implementing it in the real world. The great advantage of such models is that they can model the interdependence of variables and show the effects of feedback loops. A change in a single variable will often ripple through the system creating a range of hard-to-anticipate changes in unexpected places. For example, we would expect the abolition of the sales tax to create a rise in the amount of shopping. But we might also learn that the increased traffic in the shopping district would accelerate the deterioration of roads, while the elimination of the sales tax would leave us without funds to repair the roads sooner than originally planned. The deterioration might be so severe that shoppers would abandon the shopping district for other, more pleasant surroundings, even if they have to pay the sales tax. Our policy change would have ended up having the exact opposite effect we intended. If the model is equipped with the good data and good equations (two very important ifs), it can show us a fuller range effects of the policy change than we could generate on our own.
Computer models are now a standard apparatus in the fashioning of economic policies. The number of equations is often staggering: a single model may have more thousands of separate equations. One may represent automobile sales; another the existing stock of cars; still another, consumer debt.
Computer models are also used outside of economics. For example, scientists are developing models that may be able to predict the course of a forest fire almost as soon as it is detected.28 Computer models are gaining increasing predictive success, although such success is largely limited to areas such as engineering or economics, where there are sufficiently reliable data and fairly well-grounded laws, theories, and assumptions about the behavior in the system being modelled. Thus Nicholas Rescher, a philosopher and one of the original futurists at the RAND Corporation in the 1950s, writes: "In their technical refinements, their precision, and their capacity to combine both scientific findings (natural laws) and the rules of thumb used in informed judgment ("expert systems"), computer models are the most flexible and powerful predictive tools we have."29
Global Models
Global models are a specific type of model that has had significant influence in futures studies. Global models use sets of mathematical equations, programmed to run on a computer, that describe problems of global scope.30 Probably the best-known global model was that developed by Jay Forrester and his colleagues at the Massachusetts Institute of Technology, which resulted in the publication of The Limits to Growth in 1972. Although Limits was criticized from the beginning for certain shortcomings in its methodology, (see Chapter ___), it nonetheless sparked great interest in global modeling. This controversial "Limits to Growth" study, sponsored by the Club of Rome, was followed by other world-model studies that were less pessimistic.
"At its most ambitious, global modelling appears to be an attempt to build a computer-based model of the world economy, environment, and population in order to analyze and forecast events over several decades at a level of detail, reliability and flexibility to be relevant to practical, present day policy,"31 says global modeler John Richardson.
Games
Games are not always played for entertainment. War games are a standard way for military commanders and their troops to prepare for combat. In a typical military exercise a Red army will contend with a Blue army, with actual soldiers playing the role of contenders. Such games provide military commanders with an opportunity to test alternative strategies; at the same time, both the commanders and the troops gain experience under combat-like conditions.
War games need not involve actual soldiers. The games can also be played on a more abstract level by military experts using only pencil and paper. These games are, in effect, sophisticated versions of "Battleship," the child's game in which players secretly place models of warships on a grid, and then take turns calling out the coordinates of squares in the hope of "hitting" the opponent's ships. Today's war games for the U.S. military may be exceedingly elaborate: A Game may involve a number of teams located at various think tanks, military installations, and government agencies, and use a nationwide network of computers.
Games played by Rand Corporation analysts are credited with having had a major influence on the development of U.S. military policy. For example, a series of games begun during the 1950s examined the possible role of the Air Force in a war in the Middle East, while another series of games initiated during that period dealt with the role of the military in limited wars. Together these games cast doubt on the then prevalent doctrine of "massive retaliation," and suggested that the nation must also be prepared for limited warfare.
Games can also be used to simulate international political affairs, because the military planners must be alert to circumstances in which political leaders, such as the President of he United States, may require military support. In some games, players assume the roles of individual leaders (e.g., the President of France or the Secretary General of the United Nations) or of entities such as East Germany or the Soviet Union.
Games enable decisionmakers to see more clearly what might happen if they took certain actions. Such bloodless battles can prevent the shedding of blood in actual encounters: If Napoleon and Hitler had known from war games that they would lose their campaigns against Russia, the battles of Baritone and Stalingrad would presumably never have occurred. War games have probably played a major role in preventing World War III by showing both sides what they would lose if their forces engaged in actual combat.
Games can also be used to understand the functioning of a city and explore new ways for a city to solve its problems. One player may be designated as the mayor; another plays the role of a slum lord; a third represents the labor unions, etc. The game master can then set the players a problem to solve, such as how to dispose of the city's garbage when the local landfill is closed. Each player tries to respond on the basis of the interests of the individual or group he represents, and the interactions in the group can reveal the acceptability and wisdom of alternative solutions to city problems.
Technological Forecasting
A technology forecaster generally makes forecasts concerning how soon various types of technologies will be possible and what characteristics they may have, rather than what they will have, because the actual technology that will be used in the future depends on economic, social, and political considerations, which are normally beyond the province of the technology forecaster. For example, a technology forecaster might forecast that it will be possible by the year 2050 to produce electricity from nuclear fusion, but whether thermonuclear fusion will actually be used for that purpose may depend on a variety of non-technological considerations.
Technology forecasting is differentiated from the other methods described in this chapter by the subject area of the forecasts rather than the methodology used. Technology forecasting could, theoretically, employ almost any of the other methods described here. However, technology forecasting has developed as a distinct endeavor within futures studies, with its own concepts, literature, and practitioners, so it is useful to address it as an independent method.
An important concept used by technological forecasters is "stages of innovation." Every technological advance passes through certain stages, with each stage representing a greater degree of practicality or use. According to technological forecaster Joseph P. Martino, these stages are: "scientific findings," when some basic scientific understanding has been developed; "laboratory feasibility," when a specific solution to a specific problem has been identifies and a laboratory model has been created; "operating prototype," when a device intended for a particular operational environment has been built; "commercial introduction or operational use," at which point the innovation at which point the innovation technologically successful but is also economically feasible; "widespread adoption," at which point the innovation has shown itself to be in some way superior to whatever method was used previously to perform its function and the innovation replaces a some portion of those previous methods; "diffusion to other areas," at which point the innovation becomes adopted for purposes other than those originally intended; "social and economic impact," at which point the innovation has changed the behavior of society or has somehow involved a substantial portion of the economy.32
Technology Assessment
Technology assessment is a term for a certain form of policy research aimed at the provision of a balanced appraisal for decision-makers of potential dangers, as well as the benefits, of new technologies. A proper assessment evaluates the long-range, far-reaching, and hidden social as well as economic impacts of a new or proposed technology. Ideally it identifies policy issues and assesses impacts of alternative courses of action while leaving the final decision to policy-makers and executives. Simultaneously it should inform the general public of what is in store.33 Technology assessment provides "the systematic identification, analysis, and evaluation of the real and potential impacts of technology on a wide array of societal and natural systems and process."34
Like technology forecasting, technology assessment is not so much a method as an area of inquiry that uses a wide range of other methods. Technology assessor generally believe that there is no prescribed or precise methodology, but that each assessment should vary depending on the situation. Although it must be rigorous, it is in the end an artful and creative process.35 For a description of how technology assessment was conducted at the U.S. Office of Technology Assessment, see chapter ___.
Precursor or Bellwether Analysis
Precursor or bellwether analysis is the identification of and analogizing from actors (typically government jurisdictions) that are early testers and adopters of new concepts to anticipate the timing and nature of developments in other jurisdictions or institutions. For example, the Scandinavian nations are often believed to be precursor jurisdictions, who are often first to adopt policies that are later adopted by other nations. Studying social policies in Sweden, for example, can aid in forecasting social policy in the United States.
Risk Assessment
Risk assessment is the identification and characterization of the quality and quantity of potential adverse affects of an event, such as an investment decision, a new technology, a natural phenomenon an action, or other event.36 Risk assessment is a new and rapidly developing science. The bulk of the work focuses on health and environmental risk, such as estimating the risk of lung cancer incurred by smoking or the risk of losing a species in an ecosystem given a certain level of pollution. "Risk assessment" can also focus on social or political risk, for example, the risk of a revolution in a particular country in which one is thinking of investing, although this specific type of risk assessment is usually specified as "political risk assessment." Often, "risk assessment" will refer only to the identification and quantification of health and ecological risk.
Risk assessment relies on a wide range of methods, which fill textbooks themselves. Risk assessments often involve extrapolations, and a fundamental problem of risk assessment is the validity of such extrapolations. For example, if laboratory experiments determine a specified dose of a chemical causes cancer in a certain percentage of rats, can we extrapolate this finding to conclude that a proportionate dose will cause cancer in the same percentage of humans? How do we determine what a proportionate dose would be? By body weight? By caloric intake? By some other measure? Will the percentage of rats or humans who get cancer rise and fall in sync with the dosage level, or are there threshold effects such that below a certain dosage level no one gets cancer, but above a certain level the rate of cancer rises very quickly?
Political risk assessments often look at various societal statistics as indicators of political stability or instability, such as income distribution, labor unrest, crime, etc.
Cost-Benefit Analysis
Cost-benefit analysis is a standard tool of policy analysis, business management, and other fields, as well as futures studies. In the abstract, the method is simple and unassailable: analyze and quantify the costs and benefits of a given decision, along with its alternatives, and choose the option with best cost to benefit ratio.
In practice, of course, this proves extremely difficult because so many costs and benefits are extremely difficult to quantify. How much is a human life worth? What is the value of maintaining a clear vista in a Western mountain range, rather than allowing some development? An extensive literature has developed that attempts to answer these and related questions, but the techniques remain controversial.
To many people the very concept of placing a monetary value on human life is morally reprehensible, yet policy-makers often must do so, at least implicitly, whenever they deal with health and safety issues. When health insurers decide whether to cover a new procedure, they are essentially deciding whether the expected number of lives saved justifies the cost. When regulators at the Occupational Safety and Health Administration (OSHA) create a new regulation, they are -- implicitly or explicitly -- deciding that the cost imposed on industry is justified by the human lives (or limbs) saved. Conversely, when they reject a proposed regulation, they are often deciding that the cost per life saved is too high.
Another issue with cost-benefit analysis is that frequently some costs are borne by people who do not reap the benefits. For example, when a factory pollutes a river, the factory's owners, employees, and customers reap the bulk of the benefits of the factory's existence, while people downstream from the factory bear part of the cost, for which they are usually not compensated.
The Systems Approach
[text to come]
Creativity Methods
[text to come]
Leading Indicators
[text to come]

Futures Research: Telling Good Work from Bad
Because futures research relies heavily on human judgment, it is often difficult to know which futures research projects are well-executed and which should be largely ignored. It is also important to realize the limitations of even the best-executed futures research.
Experienced futurists have developed certain caveats one must keep in mind whenever one engages in or uses futures research, guideposts for telling good work from bad, and rules of thumb for understanding why some futures research fails.
First, some caveats:
1. Forecasts will be incomplete. As Herman Kahn once said, "The most surprising future is one which contains no surprises."37
2. No forecast that depends on what humans will do can be 100 percent accurate.
2. Futures depend on chance.
3. Accurate forecasts of some complex and nonlinear systems may be impossible.
4. Extrapolation is bound to be wrong eventually.
5. Forecasting and planning must be dynamic and able to respond to new information and insights.38
Roy Amara has suggested this list of do's and don'ts for improving forecasting and planning:
1. Don't be a vacuum cleaner, collecting every speck of information that comes across your field of view -- rather, construct a set of lenses or filters to avoid infoglut; 2) Don't substitute error for uncertainty: some variables are more uncertain than others, and this must be acknowledged; 3) At times, lean against the wind and question conventional wisdom or turn a trend on its head; 4) Hedge forecasts with possible low probability/high impact surprises; 5) Look for breakpoints and discontinuities; 6) Focus on underlying driving forces; 7) Look for clusters of drivers; 8) Translate environmental forecasts into forms that have direct meaning for the organizational functions; 9) Don't over- or underestimate the rate of adopting some technologies; 10) Keep asking "So What?"39
Joe Coates suggests the following as common problems of poorly executed futures research projects: unexamined assumptions, limited or misplaced expertise, lack of imagination, neglect of constraints, mechanical extrapolation of trends, and overspecification of solutions.40
Conclusion
Futures studies methods are an evolving set of concepts, tools, and approaches, many of which have not been long in use. Thus futurists, perhaps more than other researchers, must take extra care to examine the methods they use, to explicate and to question the assumptions they use, and to bring a multiplicity of perspectives and methods to be on the problems they address. When they do, futurists can provide invaluable service in helping
1Wendy Schultz, ____________
2We should analyze this data to see whether any individual respondents rely heavily on just one method, and what the general spread of methods is for the average futurist.
3Coates, An Overview of Futures Methods, The Knowledge Base of Futures Studies, p. 63-64.
4Harold A. Linstone, Trend Indicators, Enc of the Future, p. 944
5 (Masini at 102, citing Morrison and Renfro).
6James L. Morrison, Scanning, Encyclopedia of the Future, p. 814.
7Adapted from William P. Neufield, "Environmental Scanning," FRQ, Fall 1995, citing EPRI, 1985.
8Morrison, Ency. p. 815.
9Heidi Meeker, "Hands-On Futurism: How to Run a Scanning Project," The Futurist, May-June, 1993, pp. 22-26.
10Schwartz, Scenarios, Enc of the Future, p. 817
11 FSA 95, 13341, citing Ian Morrison, The Futures Tool Kit, Across the Board, Jan 1994.
12John B. Robinson, "Futures Under Glass: A Recipe for People Who Hate to Predict," Futures, Oct 1990, 820-42.
13Karl H. Dreborg, "Essence of Backcasting," Futures, 28(9), 1996, pp. 813-828.
14(Slaughter, Vol. 1, p. 278)
15Bezold, The Visioning Method, in The Knowledge Base of Futures Studies, p. 170.
16Bezold, p. 167.
17Jungk and Mullert, Future Workshops: How to Create Desirable Futures, Institu for Soc. Inventions, London, 1987, quoted in Dator, From Future Workshops to Envisioning Alternative Futures, in The Knowledge Base of Futures Studies, Vol 2, p. 162; See also, Bell, pp. 302-03; Dator, FRQ Fall 93.

18Dator, From Future Workshops to Envisioning Alternative Futures, in The Knowledge Base of Futures Studies, p. 164-65.
19Bell 1997, p. 262.
20Bell 1997, 262-63
21Bell, 264
22 FSA 93, 12124, citing An Evaluation of Delphi, Fred Woudenberg, Tech. Fore and Soc. Change, 40:2, Sept 1991, 131-50.
23 FSA 88-89, 9545, citing FRQ, Summer 87, 5-37. [Quotes need to be confirmed with original source.]
24Hines, The Futurist, Nov-Dec. 95, p. 21.
25 FSA 93, 12119, citing Amara, Futures, Jul-Aug 91.
26
27Bell, 273, citing Meadows and Robinson, 1985.
28Earl Lane, "Crises Forecasting Offers New Ways to Predict Natural or Human Events," Washington Post, Jan 2, 1998, p. A18.
29Nicholas Rescher, Predicting the Future: An Introduction to the Theory of Forecasting, 1998, p. 109.
30(Richardson, FRQ, Spring 1987, p. 5)
31Cole, Global Models, Data Bases and Goe Info Systems, in Foundations of Future Studies, Vol. 2, 153.
32Joseph P. Martino, Technological Forecasting: An Introduction, The Futurist, Jul-Aug 1993, p. 15.
33[X][S]
34Roger C. Herdman and James E. Jenson, "The OTA Story: The Agency Perspective," Tech. Fore & Soc Change, 54(2&3):131, 135, Feb/Mar 1997.
35Herdman and Jensen, p. 135
36FSA 96, 13941, citing Covello and Merkhofer, Risk Assessment Methods: Approaches for Assessing Health and Environmental Risks, 1993.
37Quoted in Gordon, UNDP "Integration," p. 4.
38This list draws on but significantly alters that put forth by Gordon, pp. 4-5.
39FSA 90, 10240, citing Amara, Futures, Aug 88, 385-401.
40FSA 95, 13338, Coates, A Chrestomathy of Flawed Forecasts, Tech Fore and S.C., March 1994, 307-311.