Thursday 10 October 2019

The Merry Mystery of the Missing Marchers

Another month another march. The group calling itself AUOB (All Under One Banner) that has been organising marches in support of independence recently held an event in Edinburgh. Claims for attendance have been all over the place from those taking part but with one common theme, they are rather ridiculously high.

The Background
It is pretty much standard practice for people organising marches, particularly political ones, to overstate attendance as it makes their cause seem more important and probably massages their ego too, but AUOB seem to be taking this approach to extremes. Often their numbers are quoted by the press as if they are factual accounts rather than pure guesswork, grossly over-optimistic delusion or indeed intentional big fibs (pick your interpretation) This rather unquestioning acceptance of the exaggerated figures seems to have emboldened AUOB to increase the estimates to new and ridiculous levels.

AUOB have been around a while, for much of the last year or so they seemed to be a slightly chaotic group of debatable competence.  There were questions being raised by people from both outside and within about how they operated, in particular where donations from marchers went and what they were used for (Refs 1,2,3). But after a split between some of the main players things seem to be running slightly less chaotically (although not completely, as a recent 'see you in court’ response to a bill from Aberdeen council shows! [Ref 4]).

Attendance figures at most of their marches have been questioned at some point, in 2018 they claimed 100,000 in Edinburgh but the Council later estimated that there was around 20,000 (Ref 5). More recently those who oppose their separatist message have started to look a bit more closely at the marches and indicate where observations and evidence don’t seem to support the supposed attendance figures. The two most recent examples being the March in Aberdeen in August 2019, as examined by Roger White (Ref 6), and the March in Perth where a couple of videos of the whole March passing through the city  suggested it to be much smaller than the organisers would have had you believe (Ref 7).

The Edinburgh March
Saturday the 5th October 2019, AUOB have been trailing this event well in advance suggesting that 100,000 were expected to attend, and then as if that wasn’t enough upping their predictions as high as 250,000! Newspapers gladly repeated the predictions (Refs 9 & 10).
AUOB also told of another plan by some activists, to do something ‘record breaking’; “the worlds biggest human saltire” no less! All it required, apparently, was 44,000 folk to gather in Holyrood park and stand in a formation marked on the grass (Ref 11). This was going to be interesting….
The result was somewhat underwhelming given the amount of promotion it had received on social media, one couldn’t help but be unimpressed. It was also perhaps the first real sign that actual attendance might be far short of what had been touted.

The 'people's saltire'  not exactly a record breaker...


Saltire aside, eventually they were all lined up on the road and ready to go. The bikers all came down Queens Drive from the west and cruised off up Horse Wynd past the parliament,and not long after the crowd sets off behind a pipe band and a couple of transit vans.

There are plenty of videos doing the rounds taken at street level but images or footage that show the extent of the march are few and far between, the best of them is probably one taken by Twitter user Derrick Farnell which was also shown on the BBC website (Ref 12, also SEE NOTE 1 AT END). It is taken from Salisbury Crags overlooking Holyrood and clearly shows the march progressing towards the parliament as it starts off, you can see the two transit vans at the head of the procession appear from behind the trees on Horses Wynd showing where the march (fronted by pipe band) begins and the video also shows the back of the march further along Queens’s Drive almost, but not quite, to the trees west of St Margarets Loch; a length of around 600m. However there are also a large number of people standing around the parliament, many of whom probably join the march and no doubt some people also join on route so while the video probably shows the bulk of the march it is unlikely to show everyone.

UPDATE: Derrick's video was mysteriously deleted (Note 1), however there is some drone footage showing the march just about to start that shows the same view, i.e. Horse Wind/ Qeens Drive Roundabout back to the trees West of the Loch (Ref 20)

My first glance estimate was that there was probably around 10k of people lined up, and allowing for maybe another 5k joining on route and same again going straight to The Meadows that might equate to around 20k attending.

Other people ffered estimates based on some pictures taken from the same vantage point, some were much lower. While reading one of these I discovered a website called Mapchecking.com that has a crowd size estimator, so I had a go. However the difficulty is in choosing the appropriate crowd density, it’s difficult and has to be a guesstimate, although you can use various figures as a guide and there are some helpful websites from experts that explain crowd density in different situations (eg. Ref 13). On reflection the crowd density in Holyrood park is probably reasonably high as most folk are standing waiting to move, however the people who are moving  (i.e. on Horse Wynd) are likely to be at a lower density as they are moving and will naturally spread out causing the crowd to slowly stretch as it sets out, and in fact you can see this clearly in the video (Ref 12 , also SEE NOTE AT END). Hence what begins as a crowd of length 450-500 meters could easily stretch out to double or much more and end up being a kilometre or more long, plus add ons.

Taking a conservative estimate of 2 people per square meter average across the whole area in Derrick’s video (i.e. slightly higher in the stationary crowd, but lower in the moving segment) Mapchecking.com calculated around 16k.  If we allow for the extras seen milling about on the crags and in Holyrood park and around the parliament plus a few late arrivers maybe that would rise closer to 20k actually marching 
I would also assume a good few people joining the march while it’s underway, counterbalanced by a few peeling off, However I suspect a net gain, probably adding a few thousand at any rate, if we generously suggest an extra 20% we get to an estimate of around 24k marchers.

Certainly by the time the march is heading up the Royal Mile the density has dropped considerably. A video shot from above shows the crowd density at this point (Ref 14):  It looks dense when seeing the crowd coming towards or walking away but when you see it straight down from above you realise that it’s actually reasonably loose as most moving crowds usually are.

Now perhaps the static crowd at Holyrood was more dense than 2 per sq m, but realistically maybe not much more, however even if we doubled the density we’d be looking at 32k which with add ons en route might reach 40k of marchers. That still leaves us massively short of the 200k being claimed by many.

Others have tried to calculate the size of the crowd by other quantitative means, such as how long the march took to pass and the sped at which they would have to move, and many of them come out with conclusions in the same ball park as mine (I.e. 15-30k)

However the more optimistic claims for supporters of the marchers tend to revolve around those of the organisers who after the event said there was over 200k in attendance (Ref 5), or was it just 100k (ref 15)? The reports are confused, but perhaps not as confused as MP Joanna Cherry at the front of the march who Tweeted that there was 250k behind her (Ref 16).

Of course maybe the other 60-210k (depending on who you believe) just showed up at the Meadows, interestingly I haven’t seen any pics of the Meadows once they all arrived, more people than Glastonbury, that’d be impressive. (Actually I have seen some Meadows footage now, and it's not very conclusive, but in the high four figures or low five figures at best.

It might also be worth noting that a few days later Scotrail Tweeted that they had 8000 passengers go through Haymarket and Waverley between 11am and 2pm. So it sure looks like most of the alleged 200k didn't come by train! (Ref 18).

All aboard the bandwagon
For what it’s worth I don’t think that the organisers or Joanna Cherry can’t count, instead what is going on here is a deliberate attempt to overstate support for independence to try and encourage the bandwagon effect to do its thing.
The bandwagon effect is a phenomena whereby people are inclined to join a particular group or cause simply because they believe everyone else is. There are a number of psychological explanations for this that you can read about elsewhere but it is a real effect and has been widely studied by social scientists. It has it’s origins in 19th century US politics and the context of this march is almost exactly the situation from which the effect got its name (Ref 17).
It's one reason that political parties get so animated about opinion polls, particularly changing ones, they'll either vehemently deny their accuracy or laud over them depending what the polls say. A rise in our poll rating is portrayed as evidence that people are starting to come to us, and you should too. A falling one is evidence that the polling company got it wrong. It's all about claiming 'momentum'. Political activists always like to portray their support as on the rise and it's in this context that we should view exaggerated march numbers. A march no bigger than last time shows no momentum, and even if there are more folk than last year why not say there's even more, because it makes you look like you've got even greater momentum, and if you've got away with it a few times and the media are reporting your exaggerated figures then hell why not just keep inflating them? Also if you’ve been making ridiculous predictions before the event, to admit you had fewer turn up is going to be embarrassing so in some way AUOB may have painted themselves into a corner.

These numbers matter, if people lie and exaggerate it can have real world effects, which is why they need to be called out and rationally examined. After all, lies and exaggeration are not a good thing to base any momentous and essentially irreversible decision on (look at Brexit!), even if both sides claim that's always what the other side does. So yeah lots of people marched and it was impressive, but nowhere near as many as some would like you believe. 




Note 1
When I was finishing off this post I discovered that Derrick Farnell's Twieet  had disappeared, along with his video! how bizarre... Perhaps he was a supporter of independence and suddenly realised that his video exposed the exaggerations around attendance at the March so removed it (and all of his tweets it seems). The embedded tweet on the BBC website also disappeared. Maybe there's nothing sinister going on, but it sure seems like an attempt to erase the evidence. You can see that the tweet with video was still appearing in Google's Cache, but clicking on the link just resulted in a "that page does not exist". 



Note2:
Roger White has also written about the exaggerated numbers, but has used a different method calculate a broadly similar figure. He published his while I was writing this. (Ref 19)


Ref 7: https://www.youtube.com/watch?v=4fheK-hCAyU (AFFG video of Perth March)