A poll, commission by the ruling United National Movement (UNM) party, shows that UNM has a big lead over Bidzina Ivanishvili-led Georgian Dream opposition coalition with 46% to 24% among 1,500 voters surveyed in early August.

According to the poll commissioned to the U.S.-based research and strategic consulting firm Greenberg Quinlan Rosner (GQR) and fielded by Tbilisi-based polling firm ACT, ruling party’s support increases further to 55% when allocating those who are undecided, don’t know, or refuse to answer.

Georgian Dream’s support increases to 33% among “allocated likely voters”, which, according to GQR, is estimated by asking a number of other additional questions to those likely voters who are undecided, do not know, or refuse to answer. According this poll 20% of all surveyed respondents were undecided, did not know or refused to answer vote preference question.

Christian-Democratic Movement (CDM) and New Rights Party have 5% and 3% support, respectively, among allocated likely voters, according to the poll. Margin of error for total sample is 2.8%; and the margin of error on results for likely voters is 3.7%.

• GQR Poll Commissioned by UNM – pdf

Greenberg Quinlan Rosner (GQR), which was first hired by the ruling party for the January, 2008 presidential elections, said that its survey results were “broadly consistent” with polling commissioned by National Democratic Institute in June and by International Republican Institute in late June and early July, both of which showed the UNM with a double-digit lead.

GQR also said that the only survey showing a different result was the one by Penn Schoen Berland (PSB).

According to PSB poll, which was commissioned by Ivanishvili-affiliated organization, conducted in the first half of July, Georgian Dream had 42% and the ruling party 41% among 1980 likely voters.

PSB said that in order to address “concerns about voter intimidation in the country” it used a ‘secret ballot’ method where voters were asked to mark a paper ballot on the question of their political preferences, seal the ballot in an envelope and insert it into a bag  so the interviewers wouldn’t know the respondent’s answer to that question.

Comparing its poll with the one commissioned by the National Democratic Institute in which UNM led over Georgian Dream with 36% to 18% among likely voters with 16% refusing to answer, PSB said that if ‘refused to answer’ were to be combined with Georgia Dream, “which is not an unreasonable assumption in a country with high levels of voter intimidation, the NDI poll would also have resulted in a statistical tie.”

But GQR argues that it’s not credible to assert that all of those who refuse to answer are actually hidden Georgian Dream supporters. GQR said that in Georgia’s 2008 presidential and 2010 local elections its allocation of likely voters, who refused to answer the vote question, produced “accurate predictions” of the actual election results and “in most of those cases, more of the likely voters who refused to answer the vote question split toward Saakashvili and the UNM.”

GQR also said that PSB’s assertion, that result of its poll was different from others because it was carried out under the secret ballot methodology, was not credible either. GQR said that as an experiment, it also asked half of respondents to indicate their vote preference by a secret ballot, but it produced no statistically significant difference in the vote margin between those questioned verbally and those who responded through secret ballot.