You can tell that there are
national elections on the way, simply because of the avalanche of public
opinion polls that are being released on an almost daily basis, to inform us
about what we're supposed to be thinking with regard to the various political
parties, the individual party leaders, or even the economic abilities of the
different candidates for the future chancellor of the exchequer. Companies such
as Angus Reid, YouGov, Gallup, Harris, Ipsos Mori, ICM, Populus and Opinium are
the main polling companies that carry out such statistical exercises in the UK;
often in a very similar manner, but sometimes not.
Typically, most of these
companies will collect their data through either online polling, or via the
telephone, rather than using face-to-face interviews, which would be far more
expensive and might sway opinions amongst those respondents who dislike such
direct interviewing techniques. As a rule, most pollsters attempt to recreate
the demographic mix of the population at large, in terms of age, gender, social
class, length of residency, work status, whether they've taken a foreign
holiday in the previous three years, and geographical location, as well as
other incidental measures such as the newspapers that they read, which party
they have previously supported, which party they are leaning towards, etc.
Where respondents state that either they failed to vote in the last election,
or consider themselves unlikely to vote in upcoming elections, then they are
generally weighted down, or discounted entirely from the final calculations.
Studies in the United States has
found some evidence that although necessary for politicians, businesses and
policy makers generally, both telephone and internet polls can often have their
own inherent weaknesses, to the extent that sometimes a successful outcome, in
successfully matching polling evidence with actual outcomes, is just as much a
case of luck, as it is anything else. It is sometimes the case that telephone
polls are not truly representative of an entire market or electorate, for the
simple reason that the less economically well off, the poorest section of
society, are far less likely to have access to a telephone, or indeed a
computer. Similarly, other studies in America have found that elderly people
are far less likely to respond to a telephone poll, assuming that they bother
to answer the phone at all, whilst younger households are reportedly less
likely to have a fixed telephone line in their homes, than say their
middle-aged or elderly counterparts. These are just a few of the potential
problems and issues that virtually all public polling companies face, when
trying to find out what ordinary people think about a range of products,
policies, or whatever else the survey happens to be about. So when certain
polling organisations try to give the impression that they have successfully
identified every single possible group, whether by age, by gender, social
class, residency, work status, foreign holidays, or even by the number of cars
that they own, then it's perhaps worth being a little bit suspicious of their
claims, because there are any number of in-built social and economic issues
that make such data collection not quite so simple as they make it seem.
It is also worth pointing out
perhaps that one analysis of the now much more common internet poll, suggests
that by their very nature they are completely useless as a means of identifying
the random view and opinions of random voters, ostensibly because there is
absolutely nothing random about the people who participate in them, as the vast
majority of those being questioned are being so through personal choice. For at
least one leading American analyst, those who actively put themselves forward
to be polled, are doing so because they have a personal interest and/or
involvement with the subject at hand, which in his view makes them completely
unsuitable to participate in what is supposed to be a random survey on the
issues themselves, not of the preconceived or firmly held beliefs of the
respondents. In the same way, where is the sense in only offering respondents a
partially prompted choice of possible political parties, when other unprompted
ones clearly exist for people to choose from? Having recently participated in a
number of online surveys, the writer of this blog knows it to be true that he
chose to involve himself in the activities of a particular polling company; and
now knows how to avoid the rather obvious question of selecting either Labour,
Conservative or Liberal Democrats when posed the question of which one he would
vote for. The point being that the poll questionnaire is deliberately designed
to encourage people to choose one of the three traditional parties, rather than
selecting any one of the minor political parties, which purposefully skews the
outcome. There's nothing random in that, it's a flawed result, brought about by
a flawed concept.
Of course, there is an argument
that such questionable polling can often become a "self fulfilling
prophesy", in that they help to convince wavering voters not to vote for a
party, simply on the basis of the percentage that the polling companies have
published for them, a vote for a minor party is going to be a wasted vote, thus
encouraging some people to cast their ballot for one of the traditional
mainstream parties instead. Additionally, where deliberately leading questions
are posed to the respondent, this can often affect both their answer, as well
as their overall perception of other specific groups, to the extent that the
same question posed in a variety of ways can elicit a variety of answers,
depending on how random the polling company was the results to be.
The reality is of course that
significant numbers of people do want to genuinely participate in the decisions
that directly affect their lives; and as a result willingly enrol themselves
into polling panels, focus groups and online communities, with the intention of
having their voices heard, which is admirable. Unfortunately for most, the big
polling companies have been at their jobs a great deal longer than the average
respondent and probably know the psychological makeup of the average voter,
better than the individual knows themselves. In common with our modern media,
it is a mistake to believe that today's pollsters are following the political
developments of the day, because in all likelihood they are actually involved
with shaping the political landscape that controls our lives. Polling companies
reflect public opinion? Don't make me laugh, poll the other one!
No comments:
Post a Comment