Skip to main content

Opportunities for government to listen

As part of the Reimagining participatory democracy discovery at the end of last year I became really interested in how social media, specifically Twitter, is a space where people talk a lot about government policy. At the same time, it’s often hard for government to get good response rates to consultations, specifically from diverse voices. There seems to be an opportunity there - how can government (without being creepy) listen to what people are saying? Also, how can people lead the conversation?

This thinking is not new. Audrey Tang, the Digital Minister in Taiwan, has been part of a government that demonstrated a “scalable listening process.” Taiwan found that listening and deliberation were often missing during policymaking, yet they are a critical step in bringing diverse groups or polarised thinking together.

I’ve also being reading Carl Miller’s work on the Rise of Digital Politics. His UK research talks about how social media is changing the way that people participate in democracy. It’s engaging some of the most disengaged. Young people, who are least likely to vote, say that they feel more engaged politically when they participate in social media for political purposes and that they’d be more likely to vote because of it.

Testing out methods for listening on Twitter

To test out whether the assumption that people are talking about government policies on Twitter in reasonable volumes is true, I asked Jay Gattuso for help. Jay has deep knowledge and skills with this type of work. His earlier efforts around the Kaikoura earthquake and the General Election 2017, are really insightful, and he also has the tools and theoretical frameworks for analysis at his fingertips.

After talking we decided to crawl a hashtag on Twitter that people consistently use to talk about politics: #nzpol.

4 hashtags were found in an initial crawl of the #nzpol hashtag that looked like they deserved further analysis:

Each of these appeared 48 times, except #OIA which appeared 49 times. The 48 instances were actually one tweet by @domesticanimal and 47 retweets.

The tweets were analysed through a number of lenses: word frequency, tweet length, tweet popularity and sentiment analysis. These approaches were used to show some of the range of insight that can be sought from a corpus of tweets.

Sentiment analysis is a particularly interesting method that may offer some genuine insight into the emotional response of tweeters commenting on a topic. We used two different models to give some common grounding to the technique:

Using two methods provided a more robust way of analysing the data. We wanted to identify trends in behaviour. Is the person normally negative? Or is it just about political issues? This information could be used to moderate the data – how do you interpret many tweets from one person vs one tweet from someone else?

Conclusions and more questions

There’s value in doing some more deep dives into the data to see if government can do scalable listening in an open way. We can imagine using language- and geographical-based metrics to further refine the dataset, perhaps following individual tweeters to see who they are interacting with in their open discourse. We could baseline the techniques by comparing tweets of other comparably-sized and technically-savvy nations. It would also be valuable to investigate if people are happy to be shifted from a discussion on the social media platform to participating in government decision making.

We need to keep in mind that:

  • There is the potential for people to game or manipulate what we do.
  • Volume levels in NZ are low compared to overseas, which makes sentiment analysis more difficult.
  • Statistical/data crunching skills are needed to produce robust insights.

I’d be really interested to hear insights from other government agencies on how they’re using social media other than as a broadcast communications channel. If you’re working on something or just thinking about it, get in-touch, it would be great to hear from you.

Utility links and page information

Was this page helpful?
Thanks, do you want to tell us more?

Do not enter personal information. All fields are optional.