NHS Debate on Twitter

During the period of Feb 22 — March 18 2016 I collected 670k Tweets relating the the NHS.

There is currently a debate about several important issues regarding the future of the NHS in the UK including the amount of privatisation and a contract dispute between junior doctors and the government.

The objective of this analysis is to understand the Twitter conversation on these (and other) topics in respect of:

1. Who the important Influencers are.
2. What topics of conversation are taking place.
3. What communities do people belong to.

The analysis is done first by automated machine learning technology and which outputs information to maps and dashboards for further analysis.

Here is a link for further technical information about the methodology used for this analysis.

Firstly let’s look at a single week from the conversation — 7th-13th March.

Here is the conversation map of the Twitter activity that week.

Click here for zoomable high resolution version.

In these maps each node is a User and the lines connecting them represent communication in the form of a Re Tweet, Mention or Reply. Densely grouped Users are ones that communication with each other. The colours also highlight groups of Users who tend to share common interests in other words communities of Users.

You can see that there are is a single very tight community in Blue and to the left of that a large group consisting of a few sub communities. Just by looking at the Users in each area you can get a feel for what the shared interests are.

Here I have illustrated four of the main communities and with some of the topics that have been detected automatically from the content of the Tweets within those communities.

This clearly indicates that there is a coherent set of interests detected by the analysis.

QUIZ Example

The word quiz at first seems a little odd in the context of a conversation about NHS Operations.

Using the dashboard we can isolate that word and analyse the content of the Tweets within the communities that contain it as shown below.

This is the most ReTweeted Tweet in this community promoting an NHS Choices public health campaign.

Notice also that the #OneYou hashtag was prominent in the hashtag list. Here is the search for that hashtag on Twitter.

Which helps explain why these guys crop up on the map:

The Save The NHS Campaign

Here is a closer view of that part of the map illustrating the group I have called the “Save the NHS” campaign.

If you look at the overall map you can see that this is by far the biggest and most dense part.

Also our analysis calculates the most influential Users in the overall conversation using four measures. Here are the tables for the week being analysed.

See here for an explanation of the four tables.

Highlighted are the Users explicitly associated with the Save the NHS campaign in the Top Overall table.

So, it appears that the Users tweeting about this issue are highly influential on Twitter and successfully promoting the objectives of the campaign.

However, there is a potential catch.

Echo Chamber Effect

Look again at the overall map.

The Save the NHS campaign is broadly fighting against the privatisation of the NHS.

In a public debate you usually see the various parties represented on Twitter. See our initial analysis of the Brexit discussion for an example.

Brexit Analysis

There is something missing from the NHS discussion. The pro privatisation side is not represented.

The allegation is that the government is conducting the privatisation. So you could say that the proponents of privatisation are represented by the government.

However, these Users are highly visible in the map because they are being mentioned a lot by others who are overwhelmingly being critical of them.

Here you can see that two of these Users Jeremy Hunt and David Cameron are mentioned thousands of times but have zero ReTweets and only two Tweets between them.

In short they are not engaged in this conversation.

Conclusion

If all these factors are taken into account there is a strong indication of an echo chamber effect taking place. This is evidence worth investigating further that the Save the NHS campaign on Twitter is making a lot of noise but the only people listening are the participants themselves.

If the objective of the campaign is to reach a wide and politically diverse audience of voters then it may be necessary to adjust the tactics employed on Twitter.

Next I will analyse the full four week data set with over 600k Tweets. This provide further evidence to support or contradict the theory of an echo chamber.