When an explanation is not an explanation

Source (WaPo 09/16/04): The full court press is on: Virtually every day the editorial pages are hammering home the accusation of “flip-floppery.” It is looking like this, more than anything else, will define the presidential campaign this year. Most conservative columnists seem to be splitting their time equally between the accusation of flip-floppery, castigation of CBS, and discrediting of Kitty Kelly’s unflattering gossip about the Bush family.

An adequate logical analysis of these accusations of “flip-floppery” exceeds the space we have here. There are a number of questions that need to be answered in order to understand and evaluate the meaning of, the evidence for, and the validity of the implicit arguments based on this accusation: What is a flip-flop? When is a change a flip-flop? What is the argument in each instance of putative flip-flopping? Is flip-flopping bad? Are the arguments constructed on the basis of this accusation fallacious or unfallacious ad hominem arguments? Perhaps over the next six weeks we will be able to clarify some of this, but for now the focus must be narrower.

Margins of error rarely stop pundits from drawing conclusions from polling data. The transformation of subtlety, nuance, ambiguity, and probability into simple certain opinions is the modus operandi of the many of the more ideological pundits. The political value of certainty, whether expressed by our leaders or our pundits, is being demonstrated once again in this election season.

In today’s Washington Post, Charles Krauthammer takes it upon himself to explain Bush’s lead in the polls by John Kerry’s flip-floppery.

If the election were held today, John Kerry would lose by between 88 and 120 electoral votes.”

To be fair to Krauthammer this kind of prediction is quite common and even perhaps useful with significant qualifications (see electoral vote predictor). Nonetheless it is profoundly misleading in this context. Currently 74 electoral votes for Bush and 42 for Kerry are within the margin of error of all polls (<2-3% difference for the most part). This might still leave Bush with 60-70 point spread, but only on the condition that the election involved the disenfranchisement of around a quarter of the country, that is, with leaving 136 electoral votes uncounted.

To put it bluntly, we simply do not know who would win these states if the election were held today and any claim to know on the basis of current polling involves fallacious reasoning: In this case, probably a fallacy of hasty conclusion. It may, in fact, be that the conclusion is true (If the election were held today. . ..), but the polling data on which this conclusion implicitly relies is not an adequate justification that the conclusion is true, or in this case, even that it is more probable than not.

The volatility of the poll numbers does not chasten Krauthammer. If Krauthammer had written this column a week ago, he might have written: “If the election were held today, John Kerry would lose by 2 electoral votes.” (source). Has Kerry significantly changed his strategy or his positions in the last seven days? Is there any plausible explanation for this wild oscillation in the poll numbers, except a fundamental and inherent uncertainty of political polling? Or perhaps right after the Demmocratic convention he would have written: “If the election were held today, John Kerry would win by 57 electoral votes (source).

Predictions based on statistical differences within the margin of error of the sample cannot reliably be used to draw the conclusion that Krauthammer wishes.

Now this is only the first sentence of Krauthammer’s editorial. Krauthammer wants to explain this supposed electoral victory by Kerry’s flip-floppery.

Explanations proceed from a well-established fact and provide an account of its causes. Thus, they also involve arguments which proceed from well-established premises and argue to a conclusion. In the case of explanations they argue to the conclusion that the explanation is, at least plausible, and ideally the best one available. But if the fact that is being explained is not a fact, then there is nothing to explain and the the project of explaining becomes fatuous, even if the facts that provide the explanation are completely true!

By definition an argument with a false conclusion and true premises, however, is invalid.

  1. Bush’s central vulnerability is the war in Iraq.
  2. Kerry cannot credibly speak on Iraq, because of his flip-floppery.
    • Various contradictions in his speeches on the subjects.
  3. John Kerry would lose the election, if it were held today.
    • recent polling data from the various states.
  4. Therefore, Kerry’s flip-floppery is the reason for (best explanation of/cause of) his current losing of the election.