Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Bill Clinton Scandal - SANITY CHECK -- Ignore unavailable to you. Want to Upgrade?

To: j_b who wrote (1963)	9/9/1998 9:21:00 AM
From: Zoltan!	Read Replies (1) \| Respond to of 67261

Could have been "push-polling", where the caller pretends to be taking a poll but is actually working for a party. There is no genuine polling happening, they just use propaganda to try to influence the callee's vote.

To: j_b who wrote (1963)	9/9/1998 11:35:00 AM
From: The Philosopher	Respond to of 67261

OT You may have heard it many times (that you can get any poll results you want by wording the question right and selecting your polling audience), but in this case the oft-repeated happens to be true. In one of my earlier jobs I was a professional poll writer, and I know. Nothing so obvious as hanging up on people who don't give you the "right" answer. A lot more subtle. For example, suppose you take the same questions and ask them of 100 subscribers to National Review and another group of 100 subscribers to Mother Jones. I guarantee you will get a different result just by who you asked. Or ask the same question in an upscale grocery store in Orange County and in a bogeta in Los Angeles. Again, same questions, same target group "selection" (100 grocery store patrons), very different results. Another example: Walk into the Republican caucuses in your precinct and ask: Which question comes closer to your view of Bill Clinton's job as President: a) he is doing a fantastic job, just about perfect. b) his performance leaves some things to be desired. Define the first question as job approval, the second as job disapproval. If you don't get at least a 90% "job disapproval" response, I don't believe you. Then go into the Democratic Caucuses and ask a) he is doing a rotten job and should be impeached immeditely, or b) he is doing an okay job and deserves to say in office. At least 90% "job approval" guaranteed. That's crude, of course. Pollsters have much more subtle and sophisticated ways of getting the results we want in ways that the untrained won't recognize. One example: if you give people a list of choices, most will avoid the end points. So if you say Clinton's job performance is "excellent, good, fair, poor" you will get more good and fair than excellent and poor. You then classify "excellent, good, and fair" as job approval, and bingo. You're almost guaranteed a positive result. But ask if Clinton's performance is "good, fair, poor, very poor." Some of the same people who said "good" in the first question will now move to "fair," and some of the "fair" answers in the first example will move to "poor." Identical group of people, different result. Since you will have seen only one of the questions, not both side by side, you won't know that I have manipulated the question to get the result I want. In fact, the real problem for an honest pollster is designing a question with the least possible internal bias. Believe me, as honest pollsters we used to spend hours debating a single word or selection list to try to ferret out and eliminate every possible bias. It's very challenging!