Please respond to the following survey question.
On-line polls are:
a) Meaningless.
b) Spinach.
c) Hello.
d) Quarks.
e) Perhaps.
It makes no difference to the usefulness of the result which answer you picked, even though “meaningless” was closest to the correct response. On-line polls, like call-in votes run by some radio stations and news organizations, are pretty much bogus.
I bring this up because of a survey that began, with great fanfare, a week or two ago on the MSNBC Web site. It purports to measure attitudes on cybersex.
Which reminds me, it’s time for another question in my own meaningless survey.
Cybersex is:
a) Not nearly as much fun as the real thing.
More Top Picks Best Golf Balls For Distance
b) Better than none at all.
c) Not tonight, sweetheart; the server is down.
Now, pollsters have gotten a somewhat bad name in recent years, for reasons not entirely their fault. We’ve seen the way political candidates conduct polls and then use the data to come up with positions on the issues. Incessant campaign-season polling seems to have deadened many voters’ civic sensibilities–after all, we begin to feel, why vote when you already know who’ll win? And poorly designed polls cast public doubt on the entire profession.
But survey research is a discipline with genuine roots in social and statistical sciences. Properly conducted surveys, querying random samples of a population with questions that aren’t loaded to create a specific response, can tell us important things.
A perfect survey includes everyone in the population you want to understand. The closest thing to that may be the Census “long form,” which goes to every sixth household every decade. Millions of responses, crunched by powerful computers, are the mother lode of who-we-are data without which our understanding of America would suffer.
Time for another question.
I fill out on-line polls because:
a) They excite me.
b) I get bored and restless if I have nothing else to do.
c) I like parachute jumping.
One of the most entertaining meaningless polls takes place every four years, when presidential candidates gather in Iowa for a cattle-call rally where Iowans cast ballots in the otherwise charming straw poll. Journalism organizations waste entire forests and megabucks worth of satellite time reporting the results.
A straw poll–the cybersex survey is just a high-tech version–has no scientific validity except in the room where it’s held. Participants are self-selected. The cybersex respondents, for example, aren’t representative of the general public or even on-line users; they’re only a sample of the on-line users who found out about the poll, knew how to find the Web site and then took the time to fill out the form.
Because too many people don’t understand that such a poll is basically bogus, it can influence opinions in the wider population.
By some reckonings, these polls can actually end up meaning something. In the case of the political straw poll, the news coverage it generates can give candidates sufficient publicity to become more credible in events where voters cast genuine ballots that elect delegates to state or national conventions.
Survey researchers cringe at this sort of thing. They’re right.
OK, it’s time for our final question.
People who take on-line polls seriously are:
a) Poorly informed.
b) Brain dead.
c) Looking for love in all the wrong places.