'A lot of people said that polls were done with a political objective in mind, such as when we asked people to evaluate Deng Xiaoping's deeds. Some conservatives also regarded the polls as tools to create waves of opinions on an issue,' Mr Chung said.

Doubts surrounding the neutrality of poll findings came to a head during the chief executive contest last year, as critics charged that the frequently-publicised approval ratings for the candidates played a part in fostering support for candidates and eventually the victory of Tung Chee-hwa.

Mr Chung, who has been involved in the work of the centre since its formation in 1985, said poll results should be regarded only as a neutral frame of reference.

'Results should be used to provide a source of information. We should not be using the findings as a pointer to lead society. They should only be used as reference for objective discussion,' he said.

Pan Zhongdang, who lectures on public opinion studies in the Department of Journalism and Communication at the Chinese University of Hong Kong, believes the public may be deeply influenced by the findings of polls.

'Reporting polls is necessary for the normal operations of a democratic society, but the problem is in thinking that opinion polls are the same as public opinion itself,' said Mr Pan.

There are also concerns that some surveys have been masquerading as genuine indicators of public opinion. Among the myriad research bodies in Hong Kong at present, those held in the highest regard are probably the three centres run by local universities - the Social Science Research Centre of the University of Hong Kong, the Chinese University of Hong Kong's Institute of Asia-Pacific Studies, and Lingnan College's Research and Survey Programme.

Each conduct regular studies as well as doing research for other organisations. The centres, along with several other major commercial research companies, use computerised techniques to randomly draw respondents.

However, in recent years more and more polls - often by public pressure groups and media organisations - have been done by asking readers or viewers to ring a certain telephone number and key in their opinions. Some were even done in a straw-poll fashion when the public would be canvassed in the street for their opinions.

Mr Chung said such polls were open to manipulation to foster false opinion trends. He said the findings produced should be deemed irrelevant in measuring the sentiments and views of the populace.

'What I was worried about was that these sub-standard figures were flogged as real indices and actually managed to change the public mood. Then even professional polls done afterwards would reflect the same results. The false data had set society's agenda and become a self-fulfilling prophecy.' Both pollsters and academics also blame the media for improper reporting on opinion polls, as well as serving as an accomplice in spreading the findings of substandard polls.

Mr Chung expressed dissatisfaction over the media's performance in reporting poll findings, saying too much emphasis on particular findings distorted their original context.

'For example, we were quite disappointed during the chief executive race last year. Our work then was to evaluate the candidates themselves and the approval ratings were very much lower on our priority list.

'However, the media had already established an agenda of making it a horse race and they have taken one particular question [on approval of candidates] and based their reports on it,' he said.

Mr Chung said that instead of reporting the findings as released, newspapers should follow up on figures obtained from polls to ensure they were honest and unbiased.

Dr Pan was equally vocal on the media's handling of polls. 'My overall comment is that the media is doing a very poor job in reporting opinion polls. They neglect important information that could allow the public to make judgments on whether it was credible,' he said. Dr Pan said information such as the response rate in the poll, the interviewing method, the margin of errors and even the exact wordings of the questions should be included in any report concerning polls.

'It may be because journalists were operating under the pressures of deadlines, news holes, or got careless and neglected information they do not think is essential,' he said.

Dr Pan added that the media should also be sceptical in reporting polls which were done irregularly and may be unrepresentative.

'The worst category, however, is when the media themselves conduct polls. It will be mixing together two things, reporting news and making news.

'It would be fine if they used it simply as an instrument to report reality, but I am worried that this kind of trend is creating pseudo-reality to readers or the audience,' he said.

Hong Kong Journalists' Association vice-chairman Carol Lai Pui-yee said limited space and time restricted the media's balanced reporting.

However, Ms Lai also admitted that the media had been abusing polls as a means to generate news, adding that some of the journalists were not experienced enough in handling them.

'I admit that sometimes particular responses to some polls have been pulled out of their original context and the reports become quite illogical.

It might be because of the large turnover in this profession, which makes those concerned less experienced in handling the work.'