The data shows that Brexit was a major factor in voter's decision at the ballot box.
The argument
The evidence for this explanation lies in the composition of the campaign's discourse over time. Early in the 2019 campaign, social-media posts ranged across the issues Labour wanted to fight on — the NHS most prominently, a subject on which the party performed strongly. As the campaign progressed and polling day approached, that mix changed: posts became steadily less focused on those issues and steadily more focused on Brexit. The inference rests on what the closing weeks of a campaign do. Voters decide late, and the subjects dominating discussion as they decide are the subjects they weigh. A party being talked about on its strongest ground going into polling day is in a very different position from one whose strongest ground has dropped out of the conversation. By this measure Labour spent the decisive stretch of the campaign being discussed on the issue its position was hardest to campaign on, and not on the issues where it led. Read that way, the pattern in the data is not one contributing factor among several but the explanation itself. It matches the timing of the collapse; it shows the substitution of one subject for another rather than a general loss of attention, which is what a vaguer story about a poor campaign would predict; and it identifies the specific subject that displaced the rest. Rival accounts of the defeat — the leadership, the manifesto, a long-term drift from traditional voters — would not predict a discourse shift of this particular shape in these particular weeks. A decisive Brexit effect would, and that is what the record shows. On this argument the data supply the clearest available indication that it was the party's stance on Brexit, and that alone, which produced the devastating result at the ballot box.
Context
Artificial intelligence analysed millions of social media posts surrounding the 2019 General Election. They found that the Labour Party’s Brexit policy was a key factor in the party’s loss.
Premises
Counter-arguments
Social media conversations are by no means an entirely accurate reflection of the British electorate at large. Using social media posts as a data set would skew the data towards younger, more left-leaning voters. Because many older, more right-wing voters are not on social media, or do not engage with it as frequently, it is impossible to draw conclusions about voting patterns based on social media posts alone.
Rejecting the premises
[Rejecting P1] Social media discussions do not indicate the priorities and opinions of the entire population.