German researcher Elizabet Noel-Neuman has described a phenomenon she called "the spiral of silence." On controversial issues, people tend to fall in with majority views rather than working out an opinion of their own; they are even less inclined to side with minority opinions.

So how can people be made to vote with their hearts? In other words, how do we determine who has the majority, even before the elections? By using the services of polling agencies, of course.

They claim that the margin of error in their forecasts is 2-3%.

The only problem is that there are many different pollsters. Consequently, there are many forecasts, and they frequently vary two-fold or three-fold and not by 2-3%.

A sneaking suspicion arises that these pollsters aren't researchers at all, but rather some kind of party propaganda workers: their regular publications represent support ratings and a kind of campaign activity.

For example, the Public Opinion Foundation (FOM) said that if the Duma elections had been held in September, United Russia would have got 19% of the vote; but the ROMIR Monitoring agency gave United Russia 30%. ROMIR disagreed with the National Public Opinion Research Center (VTsIOM, the non-Levada version) about the Communist Party (CPRF): ROMIR gave it 11% support, while VTsIOM put it as high as 27%.

The Ekspertiza Foundation ("Expertise") has even set up the National Political Forecasting Agency (NAPP) to average out the results produced by various polling services.

According to the averaged results of the NAPP, if the elections had been held in August, the CPRF would have won 156 Duma seats, United Russia 138 seats, the LDPR 29 seats, YABLOKO 23 seats, and the Union of Right-Wing Forces 21 seats.

But the situation changed rapidly once the campaign got underway in September, giving United Russia 185 seats, the CPRF 135 seats, YABLOKO and the LDPR 21 seats each, and the Union of Right Forces 22 seats.

The only question is whether the NAPP figures represent the golden mean or the sum of errors.

Whatever the case, it remains a puzzle why United Russia's rating suddenly climbed in September, while the CPRF dropped. Surely there must be some objective reason?

The NAPP also averaged out other polling agencies' assessments (or errors) regarding the power of the election campaigns of various parties - with the following results.

In September, the campaigns of the CPRF and United Russia were almost equally active, while in August United Russia was trying even harder than the Communists. Another question arises: if the publicity efforts of both parties are similar, why did United Russia pull ahead in September, although in August - when United Russia was promoting itself more intensively – it obviously trailed behind the Communists?

Either the researchers are using flawed methods, or they are deliberately lying.

There would be no need for pollsters to make up any figures. The desired results can be obtained simply by using a properly-designed set of questions.

If the first question asked which party was most noticeable in its campaign, the poll would favour United Russia. If the first question concerned military problems and military reforms, the outcome shifts in favour of the Union of Right-Wing Forces. If the first question referred to concerns about housing and utilities, the outcome is favourable for YABLOKO.

Polling agencies are usually commissioned to do their research by those interested in the results. The piper pays and calls the tune.

Obviously, there is no proof that any particular polling agency is a secret division of any particular party. However, there doesn't seem to be much point in seriously discussing the scientific validity of opinion polls, or the objectivity of forecasts and ratings.

Of course, the propaganda effect of ratings should not be over-estimated. Another point is more dangerous.

In Russia, the winner of the elections is announced by the Central Electoral Commission the following morning. There is no way of double-checking the honesty of the Central Electoral Commission's judgement. It is possible to challenge the voting results for a specific polling station, or even a whole district; but no way of verifying the count for an entire region or territory. The broad range of opinion poll figures is precisely what gives the election organizers some room for maneuver. If any questions are asked, they can always declare that the result is accurate - after all, it was predicted by the FOM, VTsIOM, or ROMIR.

See also:

State Duma elections 2003

Novaya Gazeta, October 23, 2003