At issue is the election of the chief executive in 2017. At most there will be three candidates, vetted by a 1,200-member nominating committee dominated by pro-establishment figures.

The local government began the reform process vowing to win over the public. The use of polling was to be its key strategy to convince sceptics, but it has found itself confronting divisive and sometimes disappointing results.

According to the rolling poll by three universities, opposition and support for the proposal has been neck and neck, although in the University of Hong Kong's latest survey released yesterday, 51 per cent of people called for Legco to approve the bill.

Still, few expect any of the 70 lawmakers to base their final decision on such public sentiment.

First, public sentiment seems to matter less than the convictions of the lawmakers, analysts say. As one pro-democracy lawmaker said on condition of anonymity: "Even if the universities' poll say there's more support than disapproval, we can't vote yes - that's not in line with what we have been asking for all along."

Second, the poll results at the core reflect the status quo of a deeply divided society. Even most polls commissioned by Beijing loyalists show at best a support rate of 60 per cent. After factoring in the margin of error and the response rate, the level of support is unremarkable.

But these findings also differ from the latest releases by the tri-university polls - regarded as one of the most authoritative - which show that opposition had briefly overtaken support, reaching what the pan-democratic lawmakers cheerfully describe as a "golden crossing".

Some 43.4 per cent said they did not back the proposal, compared with 41.6 per cent supporting it, in the poll conducted between June 4 and June 8 by HKU, Chinese University and Polytechnic University. However, the level of support again surpassed opposition in the most updated figure, released on a daily basis.

On Friday, Chief Executive Leung Chun-ying raised eyebrows when he responded to the findings by saying one should pay attention to whether the pollsters exhibited "strong political inclination" and how professional they had been. He added that the public should compare different polls as questions were asked differently.

A day later, he found at least one survey useful. He drew the media's attention to an HKU poll in collaboration with RTHK which found that 50 per cent of respondents said Legco should pass the reforms, against 33 per cent who said otherwise.

Like Leung, pro-establishment politicians argue there is a difference between asking whether one approves of the proposal and whether Legco should pass it.

The latest HKU poll - funded and commissioned by Liberal Party lawmaker James Tien Pei-chun with a "supersize" pool of respondents of 5,000 plus - covers both. Forty-eight per cent supported the proposal, while 38 per cent opposed it.

When asked which button lawmakers should press, 51 per cent supported Legco passing the bill, against 37 per cent who disapproved.

The latter finding is in line with previous surveys conducted by pro-establishment groups that asked whether Legco should approve it: there would be more support than opposition with a support rate of above 50 per cent.

Take a mid-May poll conducted by the Hong Kong Policy Research Institute commissioned by the Democratic Alliance for the Betterment and Progress of Hong Kong, the Beijing-loyalist party with the most lawmakers. Some 61.9 per cent of 1,070 respondents chose the first option, against 32 per cent who opted for the second.

But the way the questions were drafted was a "classic example of leading questions", said Dr Fu King-wa, an expert in statistical journalism at HKU.

In the poll, the DAB asked respondents: "Do you think the Legislative Council should: (1) pass the proposal, in order to allow universal suffrage of the chief executive by one person one vote, or (2) vote down the proposal, at the expense of the political system stepping on the same ground in 2017?"

Fu says: "It is too negative to include such terms as 'at the expense of' and 'stepping on the same ground' for a professional poll."

In a poll by the New People's Party in April and May, 51.3 per cent supported the proposal.

The poll was conducted on the street by the party itself - a party with only two Legco seats. Fu says the problem with a poll like this is the willingness of passers-by who do not support the party to be surveyed.

Party chairwoman Regina Ip Lau Suk-yee admitted the poll was not "weighted" and not citywide. "Poll findings are a matter of trust at the end of the day," Ip said as she unveiled her findings.

Fu stressed the importance of "weighting" a poll, a technique to adjust answers to account for over- and underrepresented groups according to census statistics.

Also on the pro-Beijing side, a more consistent monthly poll conducted by the Hong Kong Research Association found 60 per cent support for Legco to pass the plan over time.

In its conclusion for the latest release yesterday, the association, a favourite pollster of pro-Beijing groups and parties, does not hide its political inclination and "appeals to pan-democratic lawmakers to … consider supporting the proposal".

Professor Ma Ngok, a political scientist at Chinese University, says the joint-universities poll is the only credible, non-commercial one and if lawmakers were to make a decision based on a poll, this ought to be it.

"Without a majority support, a government cannot claim that it is a reasonable policy," Ma says.

After the vote, all eyes will be on whether the pan-democrats will suffer in the District Council elections this year and the Legco election next should they vote down the proposal and, in a way, disenfranchise the public. But Ma doubts this will happen, given such an evenly divided public.

"The original plan for the government was to blame the pan-democrats. I'd say that no longer works," Ma says.