By John Coleman

This can be done either by amateurs, in which case we can usually spot the fraud a mile away, or by very clever professionals who employ cutouts and firewalls to prevent all but the most discerning from knowing what’s really going on. For the latter, consider the pharmaceutical company that shops its testing around until it finds a friendly and qualified “outside” investigator to perform clinical trials. Until 1962, this was done as a matter of course. Finally, that year congress beefed up the FDA’s oversight and gave it authority to approve and validate all clinical trials of drugs and cosmetics. More than a few of us can remember in the early 1960s testimonials from time to time citing medical “findings” suggesting that the adverse health claims for tobacco were overstated. Guess who funded those docs and those studies?

The process of polling itself is quite interesting. In school, the typical example of focused or biased polling is a poll about gun control that showed an unusually high percentage of people who were polled as being in favor of gun control. The lead-in question that biased the results was: “John F. Kennedy, Martin Luther King, Jr., and Robert F. Kennedy were all shot and killed with unregistered firearms. Do you believe in gun control laws?” While a bizarre example of biased polling, this is not very far from some of the pop polling that goes on today, particularly in the area of drug control.

Consider the following: “The US incarcerates more people for drug violations than any country in the world; or The US has more African-American males in jail for drug crimes than in college or universities; or The US has almost three-quarters of a million people in jail for marihuana violations, most of whom were convicted of possession of the drug (sometimes the word “simple” is inserted before possession).”

Given any one or more of the above, the punch line is usually something like: “Therefore, do you believe the war on drugs is successful or not?; Or, Do you believe that the US drug laws are racist? Or, do you believe we should reform our marihuana laws so that we avoid the expense of incarcerating so many of our young people and thereby give them and us a chance for successful lives?”

Statistically, polling is a legitimate mathematical process for estimating population norms from sampling but the validity of the process depends very heavily on several things, including the randomness of the sample (so that it is representative of the entire population), the sample size (this can be mathematically determined for each population being studied), and the impartiality of the questions used (if polling is done by survey), the setting in which the polling takes place, and several other exogenous influences. While some of these can be policed better than others, it takes only one inherent weakness or flaw to collapse the integrity of the polling and cast doubt upon the results.

Politicians often use “private” pollsters to perform special polls, the results of which may not always be disclosed for obvious reasons. It might be better for a politician to know, for example, what percentage of people in Peoria doesn’t like him or her, rather than what percentage does like him or her. In practical terms, they already have the support of the latter so it makes no sense to spend time and money finding out how many people like you anyway. Of course, the latter polls are done and are useful for publicity purposes, hopefully to move swing voters toward the candidate.

Not surprisingly, the druggies use the latter technique when they poll people on how dissatisfied they are with present drug policies. Although a seemingly “neutral” sounding question, this one is loaded with political, cultural, and personal biases. Add the above lead-in statements or questions and you overload the bias. We should not be surprised by the results and no professional policy maker who has studied statistics 101 in high school should be swayed by them.

The more appropriate question for a poll on drug control might begin with a simple and less biased question such as, “Do you support the regulation of some drugs by federal and state governments?” There should be a range of answers so as to measure the respondent’s specific response.

Admittedly, such polls are very complex and extremely difficult to design. This is particularly true when polling on a single subject, such as drug control. I think we get a better measurement of population norms when the topic “drugs” is contained within a larger, less focused poll. For example, during the election season lots of polls are taken of voters’ general concerns, and the questions are usually posed in a way that allows the respondents to rank their concerns from 1 to 10, for example. Using unbiased, one-word definitions such as, “economics, war, national-security, drugs, abortion, civil-rights, guns, abortion, etc.,” people are less influenced by the poll itself and are more likely to respond truthfully. In these polls, we find that “drugs” usually ranks quite high, often near or at the top of many such polls.

George Gallup, jr., son of the founder of the Gallup Organization, and is CEO the Gallup Organization, the world’s largest and oldest polling outfits, is a long-time personal friend of mine and we’ve have many discussions over the years about some of these issues. He and his His organization have has developed the “science” side of polling to an extremely precise range, but the technique itself remains vulnerable on several fronts to some of the influences I mentioned above. For those who would dismiss outright the practice of polling, or its scientific legitimacy, another George (Will) once quipped: “If you don’t believe in polling, the next time you go for a blood test, have the doctor test it all!”

Author: John Coleman
Source: Northwest Center for Health & Safety
Published by: Editor BRAHA - editor@braha.net

ATTENTION: The publication of the material in this site is intended as a source for research and consulting by serving as a source of information for society and therefore has no commercial objectives.