Politics

Why it's Stella Creasy, not Jeremy Corbyn, who is at the centre of Labour's Twitter network

Before the election, Jeremy Corbyn was the most "central" member of the PLP. Now, Stella Creasy has overtaken him. Why?

October 20, 2017
Data taken from twitter reveals that it is Stella Creasy, not Jeremy Corbyn, who is at the centre of the PLP. Photo: PA/Prospect composite
Data taken from twitter reveals that it is Stella Creasy, not Jeremy Corbyn, who is at the centre of the PLP. Photo: PA/Prospect composite

It is hard to argue with the fact that Jeremy Corbyn’s position as leader of the Labour Party was strengthened by his performance in the June 2017 election. However, social media reveals a different story.

An analysis of the way Twitter is used within the Parliamentary Labour Party reveals how a Corbyn-agnostic network has coalesced around a group of women within the PLP.

The most central figure in the PLP is no longer Jeremy Corbyn—at least according to this analysis—but Stella Creasy.

242 of the 262 Labour MPs have Twitter accounts, which means the social media platform is a unique roadmap to the way the PLP communicates. It is not simply about who tweets the most and to whom, but about degrees of closeness.

Between the May 2015 General election and the June 2017 election, Labour leader Jeremy Corbyn was the central figure in the parliamentary Labour party—a position confirmed by the activity of Labour MPs on Twitter. For two years, Corbyn was directly tweeted to, or retweeted by, more Labour MPs than any other Labour MP.

Not only that, but Corbyn was also the glue connecting the parliamentary party together. There is a statistical measure of this kind of proximity: from May 2015 to June 2017, Corbyn was ahead on a metric of the most “shortest paths.” Like Kevin Bacon in the famous game, Corbyn was closer to more MPs than any other member of the PLP.

However, this rapidly changed after the 2017 election.

Data collected from Twitter since then reveals Corbyn isn’t at the centre of his party’s online discussions any more. Now, Walthamstow MP Stella Creasy and Chair of the Labour Women’s PLP Jess Phillips are at the fore of a group of women MPs who are communicating with a broader section of the party, more often.

Their network may not have a name, or the profile of warring groups within the PLP like Momentum, Progress or Compass. But its members communicate so successfully across the party, they are the effective centre of the PLP—at least according to Twitter.

The elections of 2015 and 2017 boosted the number of women Labour MPs, from a third of the PLP before May 2015 to closer to a half today. This may have laid the groundwork for a new network, but the data shows these new connections spread across the party, beyond any particular clique or grouping.

Below are four tables, as well as a graphic to show centrality. The data comes from tweets sent between 9 June 2017, the day after the General Election, and the morning of 9 October 2017.

Who tweets who?

  MP Constituency MP connections Volume of tweets Volume/connection measure
1 Jeremy Corbyn Islington North 110 443 220.75
2 Angela Rayner Ashton-under-Lyne 70 168 108.44
3 Stella Creasy Walthamstow 67 215 120.02
4 Keir Starmer Holborn and St Pancras 56 103 75.95
5 Anna Turley Redcar 52 69 59.90
6 Diana Johnson Kingston upon Hull North 39 61 48.77
7 Jess Phillips Birmingham, Yardley 38 177 82.01
8 Jonathan Ashworth Leicester South 34 79 51.83
9 Emily Thornberry Islington South and Finsbury 32 62 44.54
10 Carolyn Harris Swansea East 29 74 46.32
Table 1: the number in the final column is purely relative, to create a scale for comparisons.

The table above shows the number of MPs connecting to the named MP by tweet/retweet. Jeremy Corbyn’s MP connections score of 110 means 110 MPs tweeted/retweeted him. The volume of tweets score is the number of tweets-to and retweets. The final score uses an algorithm to combine and balance volume of tweets with number of MPs doing the tweeting. Corbyn tops every measure. (Prolific tweeter Jess Phillips jumps up the rankings once the volume of tweets is considered).

In this first table, Corbyn is ahead: he was tweeted-to or retweeted by the most MPs. But his high standing is misleading.

The tables turn

Table two looks at how MPs reach out through twitter to the rest of the parliamentary party. Corbyn is not even in the top 10. Anna Turley is: she has tweeted to 65 fellow MPs.

After the election, Corbyn falls to 36 in the PLP; he is not a prolific tweeter (or, more likely, his office isn’t).

But as leader he has other ways to keep in touch with his parliamentary party. In this table, many of the MPs are known rebels with a strong anti-Corbyn agenda. These are the resigners, like Jo Stevens and Anna Turley, and critics like Wes Streeting, who lead the charge that Corbyn is soft on anti-Semitism.

Understandably, rebel MPs needed to find a platform to speak over the party gatekeepers. This table suggests they found it on Twitter.

 
  MP Constituency MP connections Volume of tweets Volume/connection measure
1 Anna Turley Redcar 65 134 93.33
2 Stella Creasy Walthamstow 52 179 96.48
3 Jess Phillips Birmingham, Yardley 49 145 84.29
4 Jo Stevens Cardiff Central 46 125 75.83
5 Wes Streeting Ilford North 42 95 63.17
6 Rosie Duffield Canterbury 41 76 55.82
7 Stephen Doughty Cardiff South and Penarth 40 109 66.03
8 Lucy Powell Manchester Central 38 159 77.73
9 Luke Pollard Plymouth, Sutton and Devonport 34 115 62.53
10 Thangam Debbonaire Bristol West 33 105 58.86
36 Jeremy Corbyn Islington North 21 38 28.25
Table 2

 

The centre of the net

Table three evaluates the way a figure within the PLP links to other MPs, and allows us to calculate how central they are to the party: how they connect people, and how close they are to the entire network, not necessarily their own cliques. (Again, the figures are a relative scale for the purposes of comparison).

It shows the "betweenness" measures of the MPs – the MPs have ranked by the measure to show the balance of weight between connections and volume of connections. The betweenness measure in networks is a measure of influence within the group—in this case it is defined, for an MP, as the number of shortest paths between two other MPs that this MP lies on. The higher the score the more an MP is influential in linking parts of the network.

This is the key measure: it looks at centrality by measuring the “shortest paths” between MPs.

At once, Stella Creasy appears streets ahead of any other MP, with Jess Phillips next. Corbyn is a quite distant third. Moreover, Creasy isn’t speaking to a clique or single group, such as the  women’s PLP or the Welsh MPs who figure prominently in the tables.

She is an effective communicator and facilitator, and the data suggests her colleagues have already—consciously or unconsciously—recognised a figure capable of communicating across the party and placed her at its centre.

 
  MP Constituency Centrality on connections Centrality on tweet volume Centrality balance
1 Stella Creasy Walthamstow 5659.49 18395.78 11247.65
2 Jess Phillips Birmingham, Yardley 2675.62 17148.57 9409.42
3 Jeremy Corbyn Islington North 3404.74 8303.17 5698.13
4 Anna Turley Redcar 6034.06 5483.25 5290.4
5 Lucy Powell Manchester Central 1867.63 4079 3467.42
6 David Lammy Tottenham 1595.56 7223.25 3231.28
7 Jo Stevens Cardiff Central 1633.91 4477.5 2658.38
8 Luke Pollard Plymouth, Sutton and Devonport 1250.49 3867.58 2173.92
9 Carolyn Harris Swansea East 1804.61 2975.33 2111.7
10 Sarah Champion Rotherham 1263.05 3475.33 1924.58
Table 3

In comparison...

The fourth table shows the comparable data from between May 2015 and June 2017, when Corbyn was still at the centre of his party. This is the betweenness table for the Labour MP network between the 2015 and 2017 General elections. Owen Smith probably ranks so highly because he fought a leadership election in the period. The MPs have been ranked on the measure which balances number of connections and volume of connections.

 
MP Constituency Centrality on connections Centrality on tweet volume Centrality balance
1 Jeremy Corbyn Islington North 2122.75 16042.33 9200.83
2 Jonathan Ashworth Leicester South 1262.43 7417.67 3342.70
3 Angela Rayner Ashton-under-Lyne 908.57 4364.00 3028.67
4 Owen Smith Pontypridd 730.40 5325.83 2217.42
5 Jess Phillips Birmingham, Yardley 618.66 4244.33 1901.37
6 Wes Streeting Ilford North 1061.97 2003.50 1572.25
7 Michael Dugher Barnsley East 1158.69 2205.50 1310.42
8 Jo Stevens Cardiff Central 590.25 2637.50 1096.83
9 Kerry McCarthy Bristol East 383.90 1525.50 1058.67
10 Clive Lewis Norwich South 496.72 2379.00 1033.67
Table 4

Visualising the data

The graphic below shows the Labour MP tweet network. Each node represents an MP, and each connecting line represents a twitter connection between them (the direction of the connection has not been indicated to avoid cluttering the diagram). The size of each node is proportional to their betweenness score, with the top 10 being singled out in a different colour and labelled.

article body image



 

Notes:

Labour MPs sent 64052 tweets between 9 June and 9 October.

Mozdeh, created by the University of Wolverhampton Cybermetrics Research Group, was used for the data collection. It can be downloaded here—http://mozdeh.wlv.ac.uk/

The work of Dr Tore Opsahl, and his R package tnet, was invaluable in calculating the network centrality scores. A value of = 0.5 was used to calculate centrality measures. Further information is here—https://toreopsahl.com/tnet/