Public Opinion Polarization Within Population and Inter Groups: An Analysis of Hong Kong

Pundits criticize that Hong Kong is “splitting”. This phenomenon can be conceived as opinion polarization. Topic of opinion polarization is studied confining to U.S. context. Outside the U.S., this topic has been largely ignored. This study aims to fill in the blank by analyzing public opinion polarization in Hong Kong in the past two decades.

Opinion polarization is conceptualized as dispersion and bimodality of within population opinion distribution and measured by variance and kurtosis, as well as between-group differences among subgroups and measured by difference of means. We propose the research question as follow:
Do the approval rate of Chief Executive Leung Chun-ying, public opinion regarding political issues, economic issues and livelihood issues appear to be polarized in terms of within population and between sub-groups?

Method
The present study relies on the public opinion data from Public Opinion Program of the University of Hong Kong. The program conducts systematic and recurring survey regarding many aspects of social and political issues. We choose seven attitudinal items of the surveys for analysis.

Results
We first calculate variance and kurtosis for the whole sample in order to capture the distributional property of the population. Then we regress variance and kurtosis on time respectively to capture to trend of polarization. A larger variance or smaller kurtosis indicates greater polarization. Results show that public opinion about Leung is sharply polarized. Another attitudinal item that behaves significant polarized is evaluation of government performance. Two attitudinal items regarding political issues turn out to be polarized in the dimension of variance but not kurtosis. People’s evaluation of present political condition is the only item that behaves a depolarized trend. For attitude regarding economic condition, public opinion is stable. Livelihood issue opinion is also polarized in terms of variance.

In the next step, we analyze between-groups polarization using rating of Chief Executive. The subgroup of gender, in terms of male versus female, and the subgroup of age, in terms of people under 40 (the young generation) versus people in 40 above are analyzed.

We calculate the difference of means for each pair of subgroups in each time, and regress them respectively on time. Results show that the differences are both increasing in both pairs of subgroups, which indicate a trend of polarization in both subgroups. Besides, the gap between the young and the old generations is increasing much faster than the gap between male and female.
Within each subgroup, kurtosis is decreasing in the similar pattern in each pair of contrasting groups, indicating increase of bimodality in each subgroup’s distribution. The results also suggest the phenomenon of “parallel publics”. Although intergroup differences are increasing, opinion in each subgroup is moving in a similar pattern.

Conclusion
With all the evidence, we can say that overall public opinion in Hong Kong is polarized increasingly in political and livelihood issue domains, but not in economic issue domain.