Comment Re:Worse than that... (Score 1) 770
Not so much. The classic example is the minimum wage argument.
A person in a voluntary employment contract at $10/hour that roughly nets his employer a profit of say $10/hour, should he raise his demanded wage to $50/hour (or any arbitrary wage where the employer would no longer be profitable), would certainly become unemployed, either by being substituted with another worker, or by putting his employer out of business. Thus one clearly sees raising a wage causes unemployment. The large size of the raise helps make it clear that this would happen.
In practice, minimum wage and employment effects are nearly impossible to determine, because there can be all kinds of slop in the economic measurements (employers may not lay off immediately and may defer hiring, or may raise prices, or some other factor such as strong economic growth may offset the wage increase). So, we must reject empirical data that says marginal changes of a few percent don't obey the same laws as large changes.
There must then be further study to determine what factors may play into the apparent disagreement between the collected data and the presented argument. But the argument, being clearly true, can't be wrong in this case, so the data must be incomplete.
The climate science parallel is the disagreement between the apparent lack of warming in the past 11 or 17 years and the models. Given the disagreement, investigators had to determine why the model didn't fit the data, indicating a problem in one or the other (both, maybe).