Differing methodologies were the reason why two major market survey companies in Estonia had different results on which of the two main political parties received most support in their latest surveys, spokespersons for the two firms said on Thursday's broadcast of ERR politics discussion show "Otse uudistemajast".
Tõnis Stamberg of pollsters Turu-uuringute, whose latest survey puts the coalition Center Party ahead of opposition party Reform, while with at least two other market research companies the reverse is currently the case, rejected claims of representing any particular interests and noted that Reform's support rating was not out of the ordinary.
Stamberg was echoing claims he made earlier, that Reform's support is more-or-less what it normally would be and that the party had enjoyed a boost early on in the summer, at 33 percent for June, compared with 23 percent for August.
There is a hole in ratings for July with Turu-uuringute, since they did not conduct a survey due to company vacations; Stamberg has also said earlier that both public and media interest in politics tends to wane in July and August, only to pick up once summer is over.
For Reform in particular, few leading figures have been in the spotlight recently, with the exception of Jürgen Ligi, Riigikogu state budget committee chair, who is spearheading opposition to a controversial deal between the Estonian government and U.S. law firm Freeh, Sporkin & Sullivan.
Stamberg also said on "Otse uudistemajast" that if one takes into consideration error margins, Reform's rating is not as bad as it might at first seem.
Other surveys, including Norstat's weekly poll, retain Reform in first place, ahead of Center, as well as rival pollsters Kantar Emor, who keep Reform in first place on 30 percent of support in their latest rating.
Also appearing on "Otse uudistemajast" was Kangar-Emor chief Aivar Voog, who explained the difference between his company's and Turu-uuringute's results in terms of methodology – his company's had remained unchanged since 2017 whereas Turu-uuringute's had changed, he said.
Face-to-face surveys often favor conservative parties more than online ones
Some changes were necessitated by the coronavirus crisis, which meant that face-to-face interviews had not taken place during the peak of the pandemic and had barely picked up since then, which, in Kantar's cases, was compensated for by more online surveys, Voog said, an approach which he said was more effective in any case.
Stamberg of Turu-uuringute said that face-to-face and online results in surveys did not generally differ radially, though leaned a little bit towards the right when conducted in person.
"What we've noticed is that when we go out in to the field, such as directly to a person's home, the results we get in tend to be somewhat more conservative," Stamberg said, adding that support for the Reform Party is always higher in online surveys.
Reform is generally liberal both in economics and on social issues. The Social Democratic Party (SDE) is also largely socially liberal, along with the non-parliamentary Estonia 200, and some sections of the Center Party. Isamaa, Conservative People's Party of Estonia (EKRE) voters are, like their parties, conservative or conservative-nationalist.
Cost was also a factor in face-to-face versus online surveys, both men agreed, with in-person polling being more expensive, by up to 10 times according to Voog. Other factors include members of the public, households and in some cases even entire apartment houses sometimes barring entry to face-to-face pollsters, though Stamberg said that the method remained the bread-and-butter of polling, and one which he would not want to lose entirely.
Political parties on the whole were tending to set less store in the results of such polls anyway, Stamberg said, than they had in the past; Voog said that this was because communications channels are more multifarious than they used to be, with the flow of strategic information reaching parties' command centers via various different means, and more quietly than published surveys would offer.
Kantar Emor sets an age limit, Turu-uuringute does not
Another difference between the two companies is age limit, or lack thereof; Kantar Emor does not poll anyone above 84, but Turu-uuringute has no age limit. Aivar Voog said that when his company had experimented a few years ago with maximum and minimum age requirements, this did not affect results unduly.
Some observers have said that placing an age limit plays more into Reform's hands, than Center's, given the latter traditionally polls a sizable chunk from the "gray vote", often assumed to not conduct surveys online either. Turu-uuringute (no age-limit, smaller proportion of online surveys) as noted put Center ahead of Reform; Kantar Emor's chief said this difference had become even more noticeable since the current Center-EKRE-Isamaa coalition was born in April 2019.
Methodology can work in some party's favor in other ways too, Voog noted, referring to non-parliamentary party Estonia 200, in fourth place according to both Kantar Emor and Norstat, ahead of both the Social Democratic Party (SDE) and Isamaa, and in fifth place by Turu-uuringute's reckoning, despite having no Riigikogu seats.
Stamberg argued that this was because surveys were often presented as a list of parties or candidates, rather than the more open question of "who would you vote for?". In the latter case, respondents are more likely to pick the higher profile names and parties, which have stuck in their minds, he said.
Both panelists agreed that polls ought not to be banned in the period immediately leading up to an election. This happens in some countries, as a way of encouraging the public to vote (on the grounds that if they see their party doing particularly well, or particularly badly, right before election day it might act as a disincentive to vote).
Voog said that, however, it was better to always be informed, adding that as there are different research companies, these need not have a one-way effect – indeed the case with the two companies' differing results for Center and Reform.
Online is here to stay, both interviewees agreed, not only for cost reasons but also due to people being more accessible online.
As has sometimes been argued in favor of e-voting in Estonia, where votes can be changed ahead of election day, online surveys have the advantage of people being able to change their choices later on if, for example, someone else had viewed their responses on-screen while they were compiling them, causing the respondent to modify their actions.
Of the polls ahead of November's U.S. presidential elections, Stamberg noted that while Joe Biden may be ahead of incumbent Donald Trump at present, polls were not the same as an election result, in particular given the nature of the American electoral college system and the fact that senatorial elections take place at the same time. As in 2016, a few key states may determine the result, he said.
The traditional televised debates in October could also change things, Voog said.
Editor: Andrew Whyte