Selective Surveys

   Life is a series of if-then statements. That's why forecasting, economic and otherwise, can be very dodgy.
   Even so, this is the season for predictions of what's likely to happen in the coming year. But given the number of times forecasters are tripped up by reality, here are some things to remember.
   Surveys and polls are useful devices for measuring attitudes and opinions of a sample group on a given day or brief time period, then comparing these results with those in previous surveys.
   But.
   In any survey, there are several assumptions.

   First assumption: The survey is based on a valid, representative sample. That is, the responses accurately reflect the opinions of the population at large. Often, this is true, when the polling is done by independent, nonpartisan groups using scientifically valid statistically techniques.
   Moreover, the questions are carefully worded to avoid leading respondents to a politically desired answer.
   Therefore, the reputation of the surveyors is critical.
   Second assumption: The sponsors of the survey, and the pollsters themselves, do not have an interest in a specific outcome of the survey.
   Third assumption: The answers are truthful, and are not attempts by the people surveyed to please the pollster.
   Fourth assumption: The statistical sample will remain valid even if the opinions of the larger population base changes slightly.
   Fifth assumption: The survey base won't change drastically.
   
   Reality, however, has a way of stomping on assumptions. Pollsters know this and don't claim that their results will never change.
   Partisan pollsters, on the other hand, often insist that their survey results are immutable, and accurately reflect the opinions of the entire nation.
   Beware of absolutes.

   In effect, partisan pollsters deliver advance sales pitches for a politician with an agenda designed for a specific purpose.
   So how is the public to know the difference? Here is one of the duties of a free and independent press.
   Journalists will report the results of the polls, but will also note the affiliation of the pollster, if any.
   Otherwise, surveys and polls masquerading as valid and truthful representations of public opinion become little more than government propaganda.