Research on Forecasting for the Manmade Global Warming Alarm
Testimony to Committee on Science, Space and Technology Subcommittee on Energy and Environment
on “Climate Change: Examining the processes used to create science and policy” – March 31, 2011
Professor J. Scott Armstrong, University of Pennsylvania,
with Kesten C. Green, University of South Australia,
and Willie Soon, Harvard-Smithsonian Center for Astrophysics
The validity of the manmade global warming alarm requires the support of scientific forecasts of (1) a
substantive long-term rise in global mean temperatures in the absence of regulations, (2) serious net
harmful effects due to global warming, and (3) cost-effective regulations that would produce net
beneficial effects versus alternatives policies, including doing nothing.
Without scientific forecasts for all three aspects of the alarm, there is no scientific basis to enact
regulations. In effect, the warming alarm is like a three-legged stool: each leg needs to be strong. Despite
repeated appeals to global warming alarmists, we have been unable to find scientific forecasts for any of
the three legs.
We drew upon scientific (evidence-based) forecasting principles to audit the forecasting
procedures used to forecast global mean temperatures by the Intergovernmental Panel on Climate Change
(IPCC)—leg “1” of the stool. This audit found that the IPCC procedures violated 81% of the 89 relevant
support!regulation!regarding!the!protection!of!polar!bears!from!global!warming —leg “3” of
the stool. On average, the forecasting procedures violated 85% of the 90 relevant principles.
The warming alarmists have not demonstrated the predictive validity of their procedures. Instead,
their argument for predictive validity is based on their claim that nearly all scientists agree with the
forecasts. This count of “votes” by scientists is not only an incorrect tally of scientific opinion, it is also,
and most importantly, contrary to the scientific method.
We conducted a validation test of the IPCC forecasts that were based on the assumption that there
would be no regulations. The errors for the IPCC model long-term forecasts (for 91 to 100 years in the
future) were 12.6 times larger than those from an evidence-based “no change” model.
Based on our own analyses and the documented unscientific behavior of global warming
alarmists, we concluded that the global warming alarm is the product of an anti-scientific political
Having come to this conclusion, we turned to the “structured analogies” method to forecast the
likely outcomes of the warming alarmist movement. In our ongoing study we have, to date, identified 26
similar historical alarmist movements. None of the forecasts behind the analogous alarms proved correct.
Twenty-five alarms involved calls for government intervention and the government imposed regulations
in 23. None of the 23 interventions was effective and harm was caused by 20 of them.
Our findings on the scientific evidence related to global warming forecasts lead to the following
1. End government funding for climate change research.
2. End government funding for research predicated on global warming (e.g., alternative energy;
CO2 reduction; habitat loss).
3. End government programs and repeal regulations predicated on global warming.
4. End government support for organizations that lobby or campaign predicated on global
Corrections 4/3/11: Version R22
Knowledge of Roman vineyards in Britain and Viking diary farms in Greenland together with plots of
temperature proxy data over hundreds, thousands, and hundreds-of-thousands of years provide evidence
that the Earth’s climate varies, so the existence of climate change is not a matter of dispute. Global
warming alarmist analysis is concentrated on the years from 1850, a period of widespread direct
temperature measurement, increasing industrialization, and increasing concentrations of carbon dioxide in
the atmosphere. As with other periods, during this period one can retrospectively identify upward trends
and downward trends, depending on the starting and ending dates one chooses. Over the whole period that
we examined, 1850 through 2007, global annual temperature proxy series constructed for the
Intergovernmental Panel on Climate Change (IPCC) show a small upward trend of about 0.004°C per
year. There is some dispute over the veracity of the proxy temperature series (Christy, et al. 2010). For
our analyses, however, we treat the data as if they were correct. In particular, we use the U.K. Hadley
Centre’s “best estimate” series, HadCRUt31 as described in Brohan et al. (2006).
We approach the issue of alarm over dangerous manmade global warming as a problem of
forecasting temperatures over the long term. The global warming alarm is not based on what has
happened, but on what will happen. In other words, it is a forecasting problem. And it is a very complex
To address this forecasting problem we first describe the basis of the scientific principles behind
forecasting. We then examine the processes that have been used to forecast the onset of dangerous
manmade global warming and the validation procedures used to demonstrate predictive validity. We then
summarizeour validation study.
We limit our discussion to forecasting. Those who are interested in the relevant aspects of climate
science can find summaries in Robinson, Robinson and Soon (2007) and in Idso and Singer (2009).
Based on our analyses, especially with respect to the violations of the principles regarding
objectivity and full disclosure, we conclude that the manmade global warming alarm is an anti-scientific
political movement. In an ongoing study, we identified analogous alarms and report on the forecasts
behind the alarms and outcomes.
The basis of scientific forecasting
Research on proper forecasting methods has been conducted for roughly a century. Progress increased
over the past four decades, asresearchers emphasized experiments that were designed to test the
effectiveness of alternative methods under varied conditions. Forecasting research has led to many
To make this knowledge useful to forecasters in all domains, I, along with an international and
inter-disciplinary group of 39 co-authors and 123 reviewers, expert in various aspects of forecasting,
summarized the evidence as a set of principles. A principle is a conditional action, such as “forecast
conservatively in situations of uncertainty”. There are now 140 forecasting principles. The principles are
described and the evidence for them is fully disclosed in the Principles of Forecasting handbook
(Armstrong 2001). The principles are also provided on the forecastingprinciples.com site (ForPrin.com),
on which we invite researchers to contribute evidence either for or against the principles.
In practice, nearly everyone believes that their situation is different and that the principles do not
apply. I suggest to such people that they conduct experiments for their own situation and publish their
findings, especially if they contradict the principles, and by doing so advance the science of forecasting.
There can never be enough situation-specific evidence for some people but, given the evidence that many
common forecasting practices are invalid, it would be in unwise to reject the principles without strong
evidence for doing so.
1 Obtained from http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/annual; notes on series at
Conditions that apply in forecasting climate change
The global warming alarm is based on a chain of three linked elements, each depending on the preceding
element, and each element is highly complex due to the number of variables and the types of
relationships. It is much like a three-legged stool. Each leg involves much uncertainty (Idso and Singer
2009). The alarm requires:
1. a substantive long-term rise in global mean temperatures in the absence of regulations,
2. serious net harmful effects due to global warming, and
3. cost-effective regulations that would produce net beneficial effects versus alternatives such as
Effective policy-making requires scientific forecasts for all three elements. Without proper forecasts,
there can be no sound basis for making policy decisions. Surprisingly, then, despite repeated appeals to
global warming alarmists, we have been unable to find scientific forecasts for any of the three elements.
Of course, there have been many forecasts based on what we refer to as unaided expert judgment
(i.e., judgments made without the use of evidence-based forecasting principles). For example, in 1896 the
Swedish Nobel Prize winner in chemistry, Svante Arrhenius, speculated about the effect of increases in
atmospheric carbon dioxide (CO2) and concluded that higher concentrations would cause warming. His
conclusion was drawn from an extrapolation of observational data2. Arrhenius’s idea attracted little
attention at the time, perhaps because he expected benefits from warming, rather than an impending
As noted, the forecasting principles provide advice about how to forecast given the conditions.
Here the evidence yields a finding that is surprising to many researchers: use simple methods when
forecasting in a complex and uncertain situation. This was a central theme in my 1978 book, Long-range
Forecasting. Those involved in forecasting dangerous manmade global warming have violated the
“simple methods” principle.
Audit of methods used to forecast dangerous manmade global warming
Kesten Green surveyed climate experts (many of whom were IPCC authors and editors) to find the most
credible source for forecasts on climate change. Most respondents referred to the IPCC report and some
specifically to Chapter 8, the key IPCC chapter on forecasting (Randall et al. 2007).
Kesten Green and I examined the references to determine whether the authors of Chapter 8 were
familiar with the evidence-based literature on forecasting. We found that none of their 788 references
related to that body of literature. We could find no references that validated their choice of forecasting
procedures. In other words, the IPCC report contained no evidence that the forecasting procedures they
used were based on evidence of their predictive ability.
We then conducted an audit of the forecasting procedures using Forecasting Audit Software,
which is freely available on forprin.com. Kesten Green and I independently coded the IPCC procedures
against the 140 forecasting principles, and then we discussed differences in order to reach agreement. We
also invited comments and suggestions from the authors of the IPCC report that we were able to contact
in hope of filling in missing information. None of them replied with suggestions and one threatened to
lodge a complaint if he received any further correspondence. We described the coding procedures we
used for our audit in Green and Armstrong (2007a).
We concluded from our audit that invalid procedures were used for forecasting global mean
temperatures. Our findings, described in Green and Armstrong (2007a), are summarized in Exhibit 1.
Based on the available information, 81% of the 89 relevant principles were violated. There were an
2 See description on Wikipedia and original paper at globalwarmingart.com/images/1/18/Arrhenius.pdf.
additional 38 relevant principles, but the IPCC chapter provided insufficient information for coding and
the IPCC authors did not supply the information that we requested.
Exhibit 1: Audit of the IPCC forecasting procedures
Principles were: IPCC Chapter 8
Apparently violated 12
Properly applied 17
Insufficient information 38
Total relevant principles 127
Much of the problem revolves around the use of computer modelers’ scenarios as a forecasting method. As stated
correctly by Trenberth (2007), a leading spokesperson for the IPCC researchers, the IPCC provides scenarios, not
forecasts. Scenarios are not a valid forecasting method (Gregory & Duran 2001), but simply descriptions of their
authors’ speculations about what might happen in the future.
Warming forecasts and polar bears
We also examined two forecasts that were developed to support proposed policy changes. The reports assumed that
there would be global warming as predicted by the IPCC. We examined the two reports that presented forecasts in
line with the stated goal, mentioned on the first page of the report “to support US Fish and Wildlife Service Polar
Bear Listing decision”—which we coded as a violation of objectivity. Our procedures were similar to those in our
audit of the IPCC forecasts except that we also obtained coding by a climate scientist who has published papers on
climate change in the Arctic. On average, these two reports violated 85% of the 90 relevant principles. For example,
long-term forecasts were made using only five years of selected data (
Armstrong, Green & Soon 2008).
Exhibit 2: Audit of forecasting procedures used in two papers on polar bear populations
Principles were: Amstrup (2007) Hunter (2007)
Violated 41 61
Apparently violated 32 19
Properly applied 17 10
Insufficient information 26 15
Totals 116 105
One key violation was that they did not provide full disclosure of the data in their paper, and they
refused our requests for the data. They also refused to answer our questions about key aspects of their
procedures, which were not fully described in their papers. They refused to provide peer review of our
paper prior to publication. At our request, the editor of the journal invited them to provide commentary.
They missed the deadline and our paper was published with commentary by other authors and with our
replies to the commentaries. We were surprised when their commentary appeared in the journal some
months later without us having being offered an opportunity to respond. In their commentary, the polar
bear scientists claimed “every major point in Armstrong et al. (2008) was wrong or misleading.” You can
read their commentary in Amstrup, et al. (2009) and form your own opinion.
Tests of predictive validity by global warming alarmists
For important problems, it is important to test the predictive validity of the forecasting methods used.
Validation tests are normally done by simulating the conditions involved in making actual forecasts
(called ex ante forecasts) by, for example, withholding some data and forecasting what that data will be.
Thus, if one wanted to test the accuracy of a method for forecasting 50 years from now, one would make
a series of 50-year-ahead forecasts using the method of interest and one or more competitive alternative
methods, in order to compare the accuracy of the forecasts from the different methods.
We were unable to find any ex ante comparisons of forecasts by the alarmists.
In the spirit of doing a systematic evaluation of forecasts, in 2007 I invited former Vice President Gore to
join with me in a test as to the whether forecasts by manmade global warming alarmists would be more accurate than
forecasts from a no-change model. Each of us would contribute $10,000 to go to the winner’s favorite charity. The
period of the bet was to be 10 years so that I would be around to see the outcome. Note that this is a short time period,
such that the probability of my winning is only about 70%, based on our simulations. Had we used 100 years for the
term of the bet, I would have been almost certain to win. Mr. Gore eventually refused to take the bet (the
correspondence is provided on theclimatebet.com). So we proceeded to track the bet on the basis of “What if Mr.
Gore had taken the bet” by using the IPCC 0.03ºC per-year projection as his forecast and the global average
temperature in 2007 as mine. The status of this bet is being reported on theclimatebet.com.
Claims of predictive validity by alarmists
The claim by alarmists that nearly all scientists agree with the dangerous manmade global warming forecasts is not a
scientific way to validate forecasts. In addition, the alarmists are either misrepresenting the facts or they are unaware
of the literature. International surveys of climate scientists from 27 countries, obtained by Bray and von Storch in
1996 and 2003, summarized by Bast and Taylor (2007), found that many scientists were skeptical about the predictive
validity of climate models. Of more than 1,060 respondents, 35% agreed with the statement “Climate models can
accurately predict future climates,” while 47% percent disagreed. More recently, nearly 32,000 scientists have
disputed the claim of “scientific consensus” by signing the “Oregon Petition”3.
Perhaps in recognition that alarmist claims of predictive validity cannot sustain scrutiny, expressions of
doubt about the alarm are often parried with an appeal to the so-called precautionary principle. The precautionary
principle is an anti-scientific principle designed to silence people who have reached different conclusions. Alarmists,
such as James Hansen of NASA, have even suggested publicly that people who reach different conclusions about
global warming have committed crimes against the state (reported in Revkin 2008). Such attempts to suppress
contrary evidence were ridiculed by George Orwell in his book 1984: The Ministry of Truth building was inscribed
with the motto “Ignorance is truth.” For a closer examination of the precautionary principle from a forecasting
perspective, see Green and Armstrong (2009).
Experts’ opinions about what will happen have repeatedly been shown by research to be of no value in
situations that are complex and uncertain. In 1980, I surveyed the evidence on the accuracy of experts’ judgmental
forecasts and found that experts were no better at forecasting about complex and uncertain situations than were
novices (Armstrong 1980). Bemused at the resistance to this evidence, I proposed my Seer-sucker Theory: “No
matter how much evidence exists that seers do not exist, seers will find suckers.” More recently, Tetlock (2005)
presented the findings of 20 years of research over the course of which he obtained over 82,000 forecasts from 284
experts on “commenting or offering advice on political and economic trends,” which represented complex and
uncertain problems. Consistent with earlier research, he found that the experts’ forecasts were no more accurate than
novices’ and naïve model forecasts.
Our validation test of IPCC forecasting model
We conducted a validation test of the IPCC forecast of 0.03°C per-year increase in global mean temperatures. We did this
starting roughly with the date used for the start of the Industrial Revolution, 1850. As it happens, that was also the start of the
collecting of temperature from weather stations around the world. We used the U.K. Met Office Hadley Centre’s annual
average thermometer data from 1850 through 2007. Note that the IPCC forecast had the benefit of using these data in
preparing the forecasts. Thus, it had an advantage over the no-change model.
To simulate the forecasting situation, we needed unconditional (ex ante) forecasts. We obtained these for the
years from 1851 through 2007. The period was one of exponentially increasing atmospheric CO2 concentrations,
3 See petitionproject.org for details.
which are the conditions that the IPCC modelers assumed for their “business as usual” model forecasts of 0.03°C per-
year increase in global mean temperatures. We used the process of “successive updating” to obtain a total of 10,750
forecasts for horizons from 1 to 100 years ahead starting with forecasts for 1851 through 1950, then for 1852 through
1951, and so on. Relative forecasting errors are provided in Exhibit 3.
Ratio of errors in IPCC (2007) forecasts to errors in “no change” model forecast from 1851 through 2007
Forecast horizon Error Ratio # of Forecasts
Rolling (1-100 years) 7.7 10,750
1-10 years 1.5 1,205
91-100 years 12.6 305
Note that the errors do not differ substantially in the short term (e.g., forecasting horizons from 1 through 10
years). As a consequence, the chances that I will win my 10-year bet with former Vice President Gore are not
overwhelming. The IPCC model forecast errors for forecasts 91 to 100 years in the future, however, were 12.6 times
larger than those for our evidence-based “no change” model forecasts.4 In an extension, we also examined a no-
change model that used ten-year periods (instead of annual data) to forecast subsequent ten-year periods, updating
this to make a forecast each year. The results were quite similar to those in Exhibit 3.
Exhibit 3 shows relative errors, but it is also important for policy makers to look at absolute errors. Absolute
errors for the no-change model are presented in Exhibit 4. The accuracy of forecasts from the no-change model is
such that even perfectly accurate forecasts of global mean temperatures would not provide much help to
policymakers. For example, the mean absolute errors for 50-year-ahead no-change forecasts averaged only 0.24°C.
The alarmists claim that validation tests cannot be done because things have changed. Such claims are
commonly, but illogically, made by people who believe that their situation is new or so different from other
situations, and cannot be related to the past.
4 Note that, had adjustments been made to reflect the heat island effect, the shifting base of weather
stations, unsubstantiated revisions in historical temperature records, the error ratio of the IPCC forecasts
(relative to our no-change model) would have been much higher.
Exhibit 4: Forecast errors for the no-change model
Forecast horizon: Years in the future
Mean absolute errors
For forecasts of no change in global average temperatures
vs Hadley temperature data, by forecast horizon
Conclusions from our analysis of the procedures used to forecast alarming manmade global warming
Global warming alarmists have used improper procedures and, most importantly, have violated the general scientific
principles of objectivity and full disclosure. They also fail to correct their errors or to cite relevant literature that
reaches unfavorable conclusions. They also have been deleting information from Wikipedia that is unfavorable to the
alarmists’ viewpoint5 (e.g., my entry has been frequently revised by them). These departures from the scientific
method are apparently intentional.
Some alarmists claim that there is no need for them to follow scientific
principles. For example, the late Stanford University biology professor Stephen Schneider said, “each of
us has to decide what is the right balance between being effective and being honest.” He also said, “we
have to offer up scary scenarios” (October 1989, Discover Magazine interview). Interestingly, Schneider
had been a leader in the 1970s movement to get the government to take action to prevent global cooling.
ClimateGate also documented many violations of objectivity and full disclosure committed by some of
the climate experts that were in one way or another associated with the IPCC.
The alarmists’ lack of interest in scientific forecasting procedures6 and the evidence from opinion
polls (Pew Research Center 2008) have led us to conclude that global warming is a political movement in
the U.S. and elsewhere (Klaus 2009). It is a product of advocacy, rather than of the scientific testing of
anthropogenic-global-warming/ and http://wattsupwiththat.com/2010/10/15/another-wikipedia-editor-has-been-
Forecasts of outcomes of the manmade global warming alarmist movement
Using a process known as “structured analogies,” we predicted the likely outcome of the global warming movement.
Our validation test of structured analogies method was provided in Green and Armstrong (2007b).
Global warming alarmism has the characteristics of a political movement. In an ongoing study, we have
been searching for situations that are “alarms over predictions of serious environmental harm that could only be
averted at great cost.” We have searched the literature, contacted various researchers -- especially those who believe
in the global warming alarm. We have also posted appeals on email lists and on websites such as
publicpolicyforecasting.com. We repeat this appeal here.
To date, we have identified 26 analogous alarmist situations in the past. Kesten Green and I independently
coded the alarms. We coded them for:
1. Forecasting method.
2. Did the proposed action involve substantive government intervention?
3. Accuracy of forecasts was rated on a -1 to +1 scale
(-1 =wrong direction, 0=no, or minor, effect; +1=accurate)
4. Did substantive government intervention take place, or not?
5. Outcome of government policies to date on the value of their net benefit on a -1 to +1 scale
6. Persistence of government policies, to-date, on a 0 to 2 scale
(0=reversed; 1=no or little change; 2=strengthened)
We will be preparing descriptions of the analogies that will include the following elements and references
to sources of information:
1. Forecasts of impending catastrophe
2. Methods used to forecast the catastrophe
3. Actions called for (actions by government or by others)
4. Salient endorsements of the forecast by scientists and politicians
5. Challenges to the forecast
6. Outcomes of each conflict over the alarming forecast and calls for action, including forecast
We have posted full disclosure of our procedures at publicpolicyforecasting.com, and have sent announcements to
websites and individual requests to people to comment. Thumbnail descriptions are available for nine of the 26
situations (indicated by italics in Exhibit 5) at publicpolicyforecasting.com.
Exhibit 5: Analogies to the alarm over dangerous manmade global warming
(Thumbnail descriptions available for italicized analogies)
Population growth and famine (Malthus)
Timber famine economic threat
Uncontrolled reproduction and degeneration (Eugenics)
Lead in petrol and brain and organ damage
Soil erosion agricultural production threat
Asbestos and lung disease
Fluoride in drinking water health effects
DDT and cancer
Population growth and famine (Ehrlich)
Global cooling; through to 1975
Supersonic airliners, the ozone hole, and skin cancer, etc.
Environmental tobacco smoke health effects
Population growth and famine (Meadows)
Industrial production and acid rain
Organophosphate pesticide poisoning
Electrical wiring and cancer, etc.
CFCs, the ozone hole, and skin cancer, etc.
Listeria in cheese
Radon in homes and lung cancer
Salmonella in eggs
Environmental toxins and breast cancer
Mad cow disease (BSE)
Dioxin in Belgian poultry
Mercury in fish effect on nervous system development
Mercury in childhood inoculations and autism
Cell phone towers and cancer, etc.
Exhibit 6 provides an example:
Exhibit 6: Example of a thumbnail description of an analogy to the global warming alarm
Title: DDT and cancer
Date: Started in 1962
Forecast of impending disaster: Based on a book, Rachel Carson’s Silent Spring, DDT was claimed to be
a dangerous cancer-causing chemical. Publication of the book was followed by what some called a
national hysteria. The alarm over forecasts of DDT’s harmful effects combined concerns about the health
and wellbeing of people with concerns about other species. Papers by scientists purported to demonstrate
harmful effects on people from DDT exposure.
Forecasting method: A scenario based on the author’s speculations from various pieces of information
about the effects of DDT. There was no direct evidence that DDT harmed people.
Actions called for: Governments were asked to ban exports of DDT and World Bank loans would be
banned to countries that used DDT.
Endorsements of and challenges to the forecast: Leading scientists from institutions (such as Stanford
University), politicians (such as Senator Al Gore,) and a report by a commission appointed by President
Carter. The reports of the dangers were widely covered by the mass media.
Outcomes of the conflict: The U.S. Environmental Protection Agency (EPA) banned the use of DDT
following an 80-day hearing in 1972. Europe and Africa, under pressure from international agencies, did
too. No actual harmful effects on humans have been found to result from DDT. Millions of people have
died from mosquito-born diseases such as malaria. The EPA decision was based on two studies of
animals: the first could not be replicated and the second used a flawed experimental design.
Sources: Edwards (2004); Waite (1994)
Here are our preliminary findings. None of these alarming forecasts were correct. Twenty-five of them
called for government intervention. In the 23 cases where interventions occurred, none were effective. The policy
changes caused harm in 20 of the cases.
The findings will change as the project progresses and as we identify new analogies, provide more and better
description of the analogies, and obtain codings from others, especially from experts in the various areas.
We were not surprised by the outcomes, as none of the alarms were based on scientific forecasts. They
typically began with stories and progressed from there with appeals to scientific support. Another reason that we were
not surprised is that others had anticipated our findings. For example, after compiling a list of analogous situations In
1990, Julian Simon said, “As soon as one predicted disaster doesn't occur, the doomsayers skip to another... why don't
[they] see that, in the aggregate, things are getting better? Why do they always think we're at a turning point—or at
the end of the road?” And considerably earlier, in 1830, Thomas Babington Macaulay concluded, “On what principle
is it that when we see nothing but improvement behind us, we are to expect nothing but deterioration before us?”
As with our other publications related to climate change, we have received no funding, so we expect this
study to drag on. The good news is that it will allow an opportunity for researchers to provide peer review and to
suggest further improvements in our study – or, better, to conduct independent studies of analogies.
To help ensure objectivity, government funding should not be provided for climate-change forecasting. Kealey (1996)
summarized evidence on the dangers of bias in government-funded research. The government should instead rely on
As we have noted, simple methods are appropriate for forecasting for climate change. Large budgets are
therefore not necessary. Private individuals have been willing to invest much time and effort in examining the global
warming alarm without external rewards. In fact, a number of them have engaged in research on the global warming
alarm at great personal cost. The cost has been at least in part because governments have almost universally
sponsored scientists who have supported the manmade global warming alarm and these scientists have, as a
consequence, attained considerable power over learned societies, journals, funding, and universities. With the power
has come influence over news media that, by nature, are attracted to stories such as environmentalist alarms that grab
the attention of audiences
The burden rightly falls on government to obtain scientific proof that a policy will lead to superior outcomes
before increasing the burden of laws and regulations. It is not defensible to use anti-scientific procedures such as
asking scientists or scientific organizations to “vote” on policy recommendations, even when the experts are provided
with excellent information. This is especially true, given the evidence that expert opinions are useless for complex
problems such as climate change.
Instead, government should look for strict standards of objectivity in the evidence. Thus, we suggest that
government should use information for each of the legs on the three-legged stool that underlies the global warming
alarm: warming, effects of warming, and outcomes of alternative proposed policy changes, including “don’t just do
something, stand there!” The following should be included for each leg:
1. evidence, rather than experts’ opinions,
2. research from scientists with diverse views,
3. research that involves testing of multiple reasonable hypotheses,
4. use of scientific (evidence-based) forecasting methods
4. full disclosure of data and research methods,
5. criticism, replications, and extensions, and
6. testimony from scientists who have nothing to gain from the acceptance of their evidence.
Amstrup, Steven C., et al. (2009), “Rebuttal of "Polar bear population forecasts: A public-policy forecasting audit"
Interfaces, 39 (4), 353-369.
Amstrup, S. C., B. G. Marcot, D. C. Douglas (2007), Forecasting the rangewide status of polar bears at selected
times in the 21st Century. Administrative Report, USGS Alaska Science Center, Anchorage, AK.
Armstrong, J. S. (1978; 1985), Long-Range Forecasting: From Crystal Ball to Computer. New York: Wiley-
Armstrong, J. S. (1980), “The Seer-sucker Theory: The value of experts in forecasting,” Technology Review, 83
Armstrong, J. S (2001), Principles of Forecasting. Norwell, MA: Kluwer Academic Publishers.
Armstrong, J. S., Green, K.C., & Soon, W. (2008), “Polar bear population forecasts: A public-policy forecasting
audit,” Interfaces, 38, No. 5, 382–405. [Includes commentary and response]
Bray, D. & von Storch, H. (2007). Climate scientists’ perceptions of climate change science. GKSS –
Forschungszentrum Geesthacht GmbH.
Brohan, P., Kennedy, J.J., Harris, I., Tett, S.F.B. & Jones, P.D. (2006). Uncertainty estimates in regional and global
observed temperature changes: a new dataset from 1850. J. Geophys. Res, 111, D12106,
Christy, J.R., B. Herman, R. Pielke, Sr., P. Klotzbach, R.T. McNider, J.J. Hnilo, R.W. Spencer, T. Chase and D.
Douglass, 2010: What do observational datasets say about modeled tropospheric temperature trends since
1979? Remote Sensing, 2(9), 2148-2169.
Edwards, J. Gordon (2004), “DDT: A case study in scientific fraud,” Journal of American Physicians and Surgeons,
9 (3), 83-88.
Green, K. C. & Armstrong, J. S. (2007a), “Global warming: Forecasts by scientists versus scientific forecasts,”
Energy and Environment, 18, No. 7+8, 995-1019.
Green, K. C. & Armstrong, J. S. (2007b), “Structured analogies for forecasting,” International Journal of
Forecasting, 23, 365-376.
Green, K. C. & Armstrong, J. S. (2008), “Uncertainty, the precautionary principle and climate change,” Available
on-line at publicpolicyforecasting.com and other web sites.
Green, K. C. & Armstrong J. S. (2011), “Effects of the global warming alarm: A forecasting project using the
structured analogies method,” Working Paper. Latest version available at
Green, K. C., Armstrong, J. S. & Soon W. (2009), “Validity of Climate Change Forecasting for Public Policy
Decision Making,” International Journal of Forecasting, 25, 826-832.
Gregory, W. L. & Duran, A. (2001), “Scenarios and acceptance of forecasts.” In J. S. Armstrong, Principles of
Forecasting. Kluwer Academic Publishers (Springer).
Hunter, C. M., H. Caswell, M. C. Runge, S. C. Amstrup, E. V.Regehr, I. Stirling (2007), “Polar bears in the
Southern Beaufort Sea II: Demography and population growth in relation to sea ice conditions.”
Administrative Report, USGS Alaska Science Center, Anchorage, AK.
Idso, C. & Singer, S. F. (2009). Climate Change Reconsidered: The Report of the Nongovernmental International
Panel on Climate Change. Chicago: The Heartland Institute.
Kealey, Terence (1996), The Economic Laws of Scientific Research. Hampshire, UK: Macmillan Press.
Randall, D.A., et al. eds. (2007). Climate Change 2007: The Physical Science Basis. Contribution of Working Group
I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK and
New York, NY, USA: Cambridge University Press.
Revkin, A. (2008). “Are big oil and big coal climate criminals?” The New York Times: Dot Earth, June 23.
Available from http://dotearth.blogs.nytimes.com/2008/06/23/are-big-oil-and-big-coal-climate-criminals/
Robinson, A.B, Robinson, N.E., and Soon, W. (2007). Environmental effects of increased carbon dioxide. Journal
of American Physicians and Surgeons, 12, 79-90.
Schneider, S. H. (1989). As quoted in an interview in Discover Magazine, October. Available at
Tetlock, P. E. (2005), Expert Political Judgment. Princeton, NJ: Princeton University Press.
Trenberth, Kevin E. (2007), “Global warming and forecasts of climate change”, Nature.com’s Climate Feedback:
the climate change blog. Available at
Waite, Donald E. (1994). “Myths and facts about DDT,” in D. E. Waite, Environmental Health Hazards,
Environmental Health Consultant; Columbus, Ohio.
Author and collaborators
J. Scott Armstrong (Ph.D., MIT, 1968), a Professor at the Wharton School of Management, University of Pennsylvania, is
the author of Long-range Forecasting, the creator of forecastingprinciples.com, and editor of Principles of Forecasting
(Kluwer 2001), an evidence-based summary of knowledge on forecasting methods. He is a founder of the Journal of
Forecasting, the International Journal of Forecasting, and the International Symposium on Forecasting. He has spent 50
years doing research and consulting on forecasting (details at http://jscottarmstrong.com). Dr. Armstrong has also published
over 30 papers on peer review and the scientific method. He can be reached at Armstrong@wharton.upenn.edu.
Contributions to this report were made by:
Kesten C. Green (Ph.D.) of the International Graduate School of Business at the University of South Australia is a Director
of the International Institute of Forecasters and is co-director with Scott Armstrong of the Forecasting Principles public
service Internet site (ForPrin.com). He has been responsible for the development of two forecasting methods that provide
forecasts that are substantially more accurate than commonly used methods. (Kesten.Green@unisa.edu.au)
Willie Soon (Ph.D.) is an astrophysicist and a geoscientist at the Solar, Stellar, and Planetary Sciences division of the
Harvard-Smithsonian Center for Astrophysics. He is also the receiving editor in the area of solar and stellar physics for the
journal New Astronomy. He has 20 years of active researching and publishing in the area of climate change and all views
expressed are strictly his own. (firstname.lastname@example.org)