Polling Critic #55 - Looking Back – Forecast Proves Correct

Looking Back – Forecast Proves Correct ~ #55
By Alan F. Kay, PhD
© 2005, (fair use with attribution and copy to authors)
Oct. 13, 2005

What Column #54 Was Set-Up to Do. I, as a sponsor of public-interest polling, have been dedicated to determining fully and fairly what the public wants for governance in many issue areas. Poll findings released by my organization, Americans Talk Issues (ATI) Foundation have had to satisfy both the Republican and Democratic parties that its views are fully and accurately presented. Other than a handful of public-interest polling sponsors, none of the thousands of people in politics or working in government, including all the political pundits, ever in public explain fully the policy issue views of the majority of ordinary people (such as found in a large random sample).

The reader of the previous column, #54, may not have noticed that, in a way that the two parties would accept, I attempted to describe how, under similar constraints of financial and human capital, the two parties utilize their vast resources to promote all of the policies and actions that they think will first elect and later support in office their candidates and the horde of officials and staff to be appointed thereafter. This is true in off-year elections, but more so for presidential elections. A good presidential campaign is a huge and increasingly expensive task. Each quadrennial, the task is complicated by the greater range of motivations and varieties of actors working on it. These actors include the candidates themselves, prominent (face card) supporters, campaign managers and administrators, pollsters, issue experts, donors, fund-raisers, public and media relation managers, advertisers, event planners, candidate's personal and travel arrangement assistants, voter turn-out and vote-count control operators, not to mention all of the professionals, economists, lawyers, and accountants to help these people, all of whom collectively determine the next POTUS (President of the United States).

My task in column #54 was to describe how both parties did their best, and were quite successful at, controlling the entire campaign, the language and images used, and the viewpoints expressed on both issues the party was promoting and those it was ignoring. If I was successful in this task, success came from many years of working to see fairly and accurately both the Republican and Democratic sides of an issue, and the successful outcome would be both parties accepting my analysis as fair and reasonable, even though I made no mention of which party's campaign I was describing. Maybe another look at #54 (www.cdi.org/polling/54-democracy.cfm) will convince you that this succeeded.

Forecasts in Previous Columns. Many of all the 54 columns written in the last 40 months have suggested some likely major future developments, including my analysis, under various conditions of what would happen and what the final outcome would be. As time goes by, these have held up remarkably well. Doing justice to the complexity of the detailed analysis required to back-up that assertion demands the publication of a new book, not just another column. I hope to write that book some day soon. I can include here a single illustrative example.

A Forecast, So Far Correct, Contrary to Expectations of Most People. A remarkably persistent and robust forecast, first published July 1, 2003 in column #23, "Defeating Terrorism," led to a prediction in the last two paragraphs of the column, as follows:

"Based on the analysis of terrorist motivations presented in this column, it is reasonable to predict that we will not see extremely murderous attacks on the United States again in our lifetime. This prediction is contrary to what many people believe, and it has this policy consequence: the longer it remains valid, the more the American public will shift away from supporting the organization of our defenses based on fear of terrorism.

"I am less complacent than the average American and totally disagree with the behavior of all terrorists, domestic and global. Nevertheless, their motives, behaviors, and actions must be studied and understood to increase the likelihood that the United States and the world will thwart their actions. This fact is buried in the public view of top Bush officials, who ignore or vilify terrorist motives, while mainstream media seldom question such official attitudes. Do ordinary Americans recognize this [unmet need]? We do not know. Poll database searches reveal they have never been asked."

Click on column #23 (www.cdi.org/polling/23-terrorism.cfm) and note the July 1, 2003, publication date. CDI has a great editor, Theresa Hitchens, who would correctly not allow any edit of this column once posted, without the column being correctly updated. The above prediction, verbatim, was published over 27 months ago.

CBS News, often with the New York Times, repeated the following question in over 24 surveys in the four years from Sept. 20-23, 2001, to Aug. 29-31, 2005, providing a very good time series on the public's fear of imminent terrorist attacks:

"How likely do you think it is that there will be another terrorist attack on the United States within the next few months?"

The high point of fear occurred in Oct. 25-28, 2001, when 88% said "likely," of which a majority, 53%, said "very likely." Earlier in the same month, Oct. 8-9, 2001, the fear of attack had been almost as high (84% likely, 48% very likely). By December 2002, it had dropped to 73% likely, 23% very likely. Never in the four years did "likely" drop below 52%. Always a majority believed an imminent attack was "likely."

Let me dramatize the validity of the prediction in this way. Suppose I was a betting man (which I am not) and offered to bet, even money, that there would be no terrorist attacks in the next three months -- in my mind a 50-50 proposition, by no means a sure thing. An even money bet is fair for me and for whoever took the bet. Suppose my first bet on Nov. 30, 2001, was $1,000. Among the majority who believe an imminent terrorist attack likely (over a hundred million Americans according to the CBS surveys), many would be happy to take my bet. On Feb. 28, 2002, I have $2,000. I continue to bet my winnings and the original $1,000, the first of every one of the 15 quarters ending Aug. 31, 2005, thus winning a total of $32,768,000. There is no way these bets would change a majority of Americans’ thinking about the likelihood of imminent terrorist attack. There always would be an individual or pooled group, ready to step to the plate.

The prediction of America remaining free of terrorist attacks, valid so far, also has this "policy consequence: the longer it remains valid, the more the American public will shift away from supporting the organization of our defenses based on fear of terrorism."(Quote verbatim from column #23.) This is validated so far by CBS findings showing that "likely" levels, averaged year by year, dropped as follows: (1) 79%, 71%, 63%, 63%, 57%, respectively for: 2001(4^th quarter only), 2002, 2003, 2004, and 2005 (excluding 4^th quarter). The downward trend is significant, but not very fast.

A Bad Prediction. As I think back on all 55 columns, I do recall only one, column #2, published May 21, 2002, that included a bad prediction. I misjudged the significance of a new arms control agreement between President George W. Bush and Russian President Putin. The agreement replaced all previous arms control understandings by requiring each side to make an enormous two-thirds cut, across the board, in its deployed nuclear armaments. The agreement, only three pages long, omitted the details that were meticulously included in former U.S.-Soviet arms control treaties and, to Bush's credit, had important aspects favoring the United States. I thought that, when properly understood, the agreement would satisfy both hawks and doves in the United States. Previous studies of arms control agreements led me to believe that this agreement was a coup for Bush. I estimated 90% of the public would approve the agreement, but that support was soon proven irrelevant and never tested. In my enthusiasm I wrote: