We are witnessing the most intense wave of public protest in the history of Hong Kong. Why the protests are happening is widely understood: sparked by opposition against a proposed bill to allow the extradition of Hong Kong citizens to mainland China, the citywide movement calls for more democracy and an independent investigation into police violence. But who are the protestors, and how might a survey of such a complex and evolving protest be carried out?
Exploring the meaning of the Hong Kong protests
As a student of social movements born and raised (and based) in Hong Kong, I have been studying the demonstrations at a close range alongside with a team of university researchers and student helpers. We have been conducting a series of onsite surveys starting on 9 June, when one million people from all walks of life took to the streets to demonstrate against the now shelved amendment bill, which would have allowed extraditions to mainland China, further weakening the city’s judicial independence.
Since then, there have been many more protests. So far, we have completed 24 surveys, interviewing over 13,000 protesters. None of us thought that the scale of the research would grow as big as this — just as no one could predict that the anti-extradition bill protests would have morphed into Hong Kong’s biggest political crisis.
Surveying protests is nothing new to me or to other members in my team. We have carried out research on the annual June 4 commemoration of the Tiananmen Square massacre, the annual July 1 anti-government rallies, and the Umbrella Movement in 2014. But in our experience, none of these are comparable with this summer in terms of the scale, frequency and intensity of protests.
There have been protests every weekend throughout the summer and an increasing number of them on the weekdays. Many of these protests were announced shortly before taking place, and it has become difficult to keep track of them. Some of them were massive in size, which are not unseen in the city but only happened once in a while back then. And most of them would descend into chaotic and violent confrontations between frontline militant protesters and riot police who came with heavy-handed responses, making it practically impossible to continue the onsite research.
Sample sizes and methods
These previously unseen conditions created a number of challenges for our surveys. Safety of our helpers (and ourselves) was one major concern — but another pressing issue from a research perspective is to collect a large-enough sample while ensuring that the sample is representative of the protesters. The spontaneity of the protests means that we had little time to amass a large team of well-trained survey helpers, which made it unwise to stick to the traditional paper-and-pencil approach (which is time-consuming and is likely to result in a small sample).
Although the availability of smartphone-based online surveys (given the high penetration of smartphones here) offers the opportunity to quickly reach a large number of respondents, it is likely to produce a convenience sample biased towards those who are more technologically-savvy and those who are less concerned about leaving digital footprints behind. Deciding how to strike a fine balance between these two conflicting objectives became a primary challenge for us.
Over the course of studying the cause of the protests and demographics of protestors, we learned that one way to ensure a large but representative sample is to combine different sampling methods. Our normal procedure was to assign survey helpers to different locations evenly distributed along a protest route or across a protest site. Survey helpers were instructed to select every ten protesters they encountered, who would in turn be presented with the survey questionnaire and then given the choice between the paper-and-pencil approach or the smartphone-based approach (through a QR code).
Unsurprisingly, the smartphone-based method would often end up collecting a much larger sample (and usually with younger respondents) than the paper-and-pencil one. But the latter remains useful because we believe that it is closer to the “true value” of the protester population (an assumption yet to be empirically tested). The final step is to weigh the sub-sample collected from the smartphone-based method by the distribution of those collected from the paper-and-pencil method.
Using flyers to survey protestors
We subsequently added a third sampling method starting from late July as the protests became more confrontational and began to be banned by the police. Initially out of safety concerns for the survey helpers in the July 27 protest in Yuen Long, where thugs indiscriminately attacked protesters and passers-by the week before, we dispatched a small team of experienced helpers and put them in a safe location to distribute flyers with a QR-code through which protesters could access the online questionnaire with their smartphones.
Certainly, this method was not ideal, since it all depends on whether the protester who was given the flyer would take the initiative to complete the survey on his/her own. But it did help to obtain a reasonably large sample without exposing helpers to excessive risks.
Distributing flyers at the Hong Kong protests has an additional advantage, as we soon discovered. It may help to reduce the selection bias caused by survey helpers, who may – either consciously or subconsciously – pick their respondents without adhering to our sampling scheme. For instance, survey helpers often select respondents of similar age and those of the opposite sex, or those who look more friendly. Helpers are less likely to pick and choose recipients when handing out flyers. Since August, we began to combine flyers distribution with the other two sampling methods, depending on the level of risk at the protest.
The issue of trust
Challenges did not only come from the unpredictable nature of the protests or the heavy-handed police responses. It also came from the protesters themselves. While the response rate is generally very high (likely because protesters are particularly willing to talk to researchers as they think they are rarely being listened to in the city), there are multiple instances in which our surveys were questioned for their authenticity.
In June, when we were surveying a group of young protesters outside the Chief Executive office, some protesters ‘reported’ our survey on the Telegram a messaging app with the suspicion that we were working with the police. Fortunately, an anonymous netizen quickly clarified matters on our behalf.
During a protest at the Hong Kong airport in mid-August, the same happened again on LIHKG, a popular online forum that works like Reddit. The post gained traction immediately and people began to ‘mess around’ with the online questionnaire, filling in answers that apparently made no sense. Eventually we had to cut short the survey and clarify our intention by posting on the same forum (which, interestingly, did not gain as much attention).
The protests continue
At the time of writing, the protests in Hong Kong still had no end in sight. Despite having withdrawn the contentious extradition bill, the government has refused to concede to protesters’ demands for an independent commission of inquiry and broader political reforms. Instead, the government has chosen the repressive path, endorsing the police to quell the protests with coercive tactics and to make mass arrests. From the survey findings, which the team has published elsewhere, this strategy seems to have little effect on stopping the protests.
Although violent protest actions have become common in this otherwise very peaceful city, protesters are still showing an extraordinary level of solidarity with one another and tolerance for radical tactics. For now, it is unlikely that there will be an endgame. Perhaps there will never be one. Protests may die down or fizzle out at some point, but the dissents are certainly here to stay.