Background of “Integrated Analysis of the Final Phase of 2016 Legislative Council Election Survey” Back

Press Release on September 2, 2016

Whenever there are large-scale elections, Public Opinion Programme (POP) of The University of Hong Kong would conduct various election surveys to track opinion changes. The “2016 Legislative Council Election Rolling Survey” is conducted between July 30 and September 2, 2016, to track the voters’ latest propensity to vote and opinion changes in each geographical constituency. The rolling survey this year was co-sponsored by HK01, now TV, Cable TV and Power for Democracy, by the order of their confirmation date. The “Integrated Analysis of the Final Phase of 2016 Legislative Council Election Survey” announced by Power for Democracy today comes from the rolling data of 12 consecutive days starting from August 21 to September 1, with a sample size of no fewer than 5,000 cases. Combining a number of consecutive rolling samples together for conducting in-depth analysis is common in academic research.

The sample size of merged rolling survey data of 12 consecutive days is larger than that of the 5-day rolling sample, and thus can produce a more accurate analysis with smaller sampling errors. However, since the survey period is relatively longer, the short-term fluctuations are not easily detected. POP never objects any organization to carry out other integrated analysis on top of the 5-day rolling poll analysis. In fact, all the sponsoring organizations of the rolling survey this year have agreed to fully open the original dataset for public consumption. As a result, some people may have already carried out various kinds of integrated analysis using the dataset. POP welcomes all these efforts.

Power for Democracy has explained publicly that their objective for sponsoring this rolling survey is solely to provide reference and original data to the civic society including the general public, all media and political parties, for free. POP welcomes its decision. As long as the sponsoring organizations agree to our service policies, which include giving POP full autonomy in its operation and releasing all survey data for public consumption, POP would not consider the background and stance of the sponsoring organizations, and will treat all organizations on an equal footing. Besides, POP had accepted the sponsorship from civic think-tanks in the previous Legislative Council election rolling surveys. Therefore, POP believes that it is appropriate to accept the said sponsor. The list of all the sponsoring organizations has already been published on POP’s website, everything is open and transparent.

Power for Democracy first wanted to sponsor a large-scale survey one week before the election exclusively, with a sample size of no fewer than 5,000 cases. However, since POP would like to first consolidate the system of rolling survey, we advised Power for Democracy to use its resources to sponsor our rolling survey instead and to increase its scale, especially for the sample size of the final phase. This proposal not only fits the original idea proposed by Power for Democracy, it also allows the survey data to be shared with other sponsoring media and the civil society. Power for Democracy accepted POP’s proposal, and thus the sample size of the rolling survey was enlarged. POP hereby acknowledges its support with great thanks.

POP will release the final result of the 5-day rolling survey through its sponsoring organizations tomorrow. Please note that the data of the last day (September 2) is excluded in the “Integrated Analysis of the Final Phase” released by Power for Democracy today.

Media enquiry: Frank Lee   Tel: 3917 7724   Email: [email protected]