Major political events, almost by definition, are the things that can move polls suddenly, whether they're speeches or party conferences or policy announcements or whatever. But sudden moves always need to be interpreted with care, for two reasons.
Firstly, sudden moves aren't always real. Events might provoke a disproportionate response from people with stronger-than-typical opinions on them, whether positive or negative. This is a short-term version of good old fashioned response bias.
Secondly, even if opinion genuinely changes, it may only do so temporarily. If you ask people how concerned they are about aviation safety the day after a major air disaster, you'd probably get a different response than if you'd asked the day before.
Whether short-term blips in polling are the first of these, the second, a bit of both or just coincidental sampling noise, doesn't really matter. First there's the event itself, then the reporting, then the reaction, then the reaction to the reaction, then maybe some "shock poll" front pages (and with fewer polls these days individual surveys get even more excessive attention than pre-2015), then things settle down and opinion may or may not have changed.
Normally it pays to wait at least a week to see if a move is real and sustained before concluding anything. Otherwise you can get the sort of situation Scottish pollsters found themselves in after the EU referendum, with many finding increased support for independence that magically disappeared months or even weeks later.
But with all those caveats in mind, what do we know so far? Sky Data have done some scientific snap polling and found that people favoured the May plan to leave the single market 51-39, with net support everywhere except London. On the questions that had also been asked in October, there were swings outside the margin of error in the PM's favour.
I'm aware of one BPC member firm being in the field already on this. With any polling that comes out, pay close attention to the fieldwork dates.
The Ipsos MORI release I mentioned yesterday turned out to be a fascinating global study of populism. It confirms what many had thought – that "2016" type trends are evident across much of the world.
Now if you'll excuse me, breakfast means breakfast…