-- AN  ONLINE  PUBLICATION --



                  John M. Grohol, Psy.D.

                    Associate Editor

Americans are an odd lot. We thrive on the media, largely 
television, for its entertainment and informational value. 
Gone are the days of every city having two major newspapers 
competing for headlines and local news. Instead, most city 
newspapers are owned by large media companies, who repackage 
and cross-promote "news" items across not only their print 
media holdings, but online, on billboards, and even on local
television stations' evening news. More than half the time, 
the "news"  is not actually news, but repackaged bits of 
consumer information, opinion, or entertainment gossip which 
run on a constant basis (see the twice-annual sweeps months 
to understand the bizarre stories which get dragged out on 
your local television station's newscasts). Actual news is 
relegated to a few minutes in each national broadcast.

We have become a society which not only tolerates this type
of mediocre media coverage, but which actually thrives on it, 
allowing the media to define what is and isn't newsworthy.

Whenever a "news event" breaks, the media are on the phones 
to their polling partners to get some polls going to track 
"public opinion." This often occurs when the government is
going to take some drastic action, or something really big 
is about to occur in politics (e.g., bombing Iraq, the 
impeachment of the President). Then the media show over the 
intervening weeks a careful plot of how "public opinion" has 
risen or fallen with regard to the "news event." Spin is then 
applied, according to whichever way the media company wants 
to portray the poll results.

Americans watch or read the news, and see the results of 
the polls. Their opinions are then influenced by the polls 
and how the media portray the results. When new polls are 
taken, the new polling finds these influenced opinions. An 
opinion snowball effect can be created by this regression.

The polls themselves are probably not at fault, although 
very little is said about the fallacies of such polls and 
how polls' quality varies greatly, making the generalizability 
of their results questionable. What is problematic is most 
media's lack of insight into this phenomenon. The media move 
from simply reporting events which take place to trying to 
shape and influence their outcome. Perhaps the effect is so 
subtle the media don't realize what may be occurring. Or 
perhaps, more cynically, the media realize exactly what it 
is they're attempting to do (shape public opinion in their 
own likeness) and have no moral or ethical dilemma in 
doing so.

A perfect example of this phenomenon cropped up in December 
1998, as the hearings of the U.S. House Judiciary Committee 
wrapped-up and they voted to send the impeachment charges 
of President Bill Clinton to the full House for a vote. 
While most polls continued to show strong support for 
stopping the impeachment process, news organizations began 
spinning the results of one poll's trend data after gathering
only one additional data point. This data point showed that 
there was a 7% increase in the number of people who wanted 
to see the President impeached. Yet the results of the poll 
still showed the vast majority of Americans favoring no 
impeachment of the President. The news organizations 
conclusion from this one data point? The tide was turning 
on the President and public opinion was shifting in favor 
of impeachment hearings.

Was the poll's data accurate, or simply a result of the media
blitz and attention focused on this singular issue (despite 
most Americans' apparent disgust over the amount of coverage 
given it, much like the American public tired over the daily 
O.J. Simpson trial updates a few years ago)? The question is 
a difficult one, but one which needs to be answered. If the 
data, as I suspect, are only indicating a self-made media 
trend, then the data are worthless in the sense of accurately 
portraying most Americans' attitudes toward the impeachment 
process. If the media are _making_ news by influencing public 
opinion in this manner, they need to be more aware of it and 
work to stop it. The American public expects the media to 
report news, not make it. 

But more interesting is the question of whether polls 
which ask about arcane political processes can ever 
reflect anything. Do most people even understand the 
impeachment process? Do they understand that every 
minute the Congress spends debating this issue, is a 
minute lost to governing the U.S.? I don't believe most 
Americans realize the consequences of the complicated 
process which is impeachment, and there is no way for 
the polls to accurately reflect this. If you ask a person, 
"Would you like to see well-known person Z 'dougafied'?"
and the media kept focusing on 'dougafying' this person, 
you might very well say "Yes," without a clue as to 
what "dougafying" means or entails. 

If the poll were to ask, however, "Would you like to see 
Bill Clinton impeached, understanding that in doing
so, the government will come to a virtual stand-still, 
all other government issues will fall by the wayside 
(including fixing Social Security, reforming healthcare 
and campaign spending, etc.), and the end result may be 
that six months from now, after millions or more of 
your tax dollars were spent pursuing this issue, nothing 
may come of it?" That is a much more fair question, 
giving a lot more information about the consequences 
of the word "impeachment" in this context. Yet a poll 
would never ask such an admittedly biased question. 
But polls could be designed to ask questions where the 
definitions of the words and their likely consequences
were well-understood by those being polled. Without that 
understanding, we might as well ask people what they 
feel about "dougafying" others.

Polling and its effects on public opinion need to be 
better understood, if not by the media who commission 
and report the poll results, then by laypeople who 
hear or read the results. Badly designed or conducted 
polls are just like any other bad research -- their 
results are meaningless and can usually be ignored.
The more people question how polls are conducted, who 
was sampled, and what specific questions were asked 
(and in what context), the more likely polls will 
begin to show their true colors. They are media 
creations in this media-driven age, where instant 
feedback on any issue facing Americans is not only 
becoming an expectation, but a demand for what the 
public thinks or feels on any issue.

The psychology and the dynamics behind polling and 
its effects on its subject need to be more closely 
researched. Until that time, polls should be taken 
with a healthy grain of salt.