Symbolix Identity

So people rioted in Belfast over a decision to only fly the Union Jack over the city hall except on important days (like er, like the Duchess of Cambridge's birthday!).

Check out the pictures. Most colorful, if you like orange.

Now I would have thought that that reaction was fairly obvious but, that is a qualitative statement from having lived in Belfast and experienced some of the culture.

Is it possible we could have quantitatively predicted this result? I reckon so, and I'm gonna show you the numbers...

Do you have a flag?

So, there was a report done for the OFMDFM. You can try and read it here, though you might bleed from your eyes :-)

The kind folk at IEL let me have the original report data in the form of an ipseus file. There are 115 responses from across the community spectrum. The details of the make up are in that report.

I was interested in seeing if the flag thing jumped out of the results. Or better yet, could it have been predicted. Luckily there was a construct in the instrument used that directly references the issue.

Construct 9 opposed '...like flags and murals that reflect my traditions' against '...all flags etc should be removed'.

So my initial thinking was this construct should be fairly divisive and along traditional culture boundaries. First, I looked at the complete set of responses as one big group.

Below, the table of Construct results show an almost equal split between folks that believe flags are good (#57) and flags are bad (#58). Good and Bad should probably be surrounded in quotes there. From a group perspective 'flags are bad' comes out as 'good'!

'Structural Pressure' is kinda like a measure of conviction. A high positive value suggest a black and white view of a construct. A high negative value suggests that the individual/group say one thing but judge things in the opposite way; they don't walk the talk as it were.

A value around zero means that either the construct is irrelevant for making judgement (i.e. not significant) or that it depends on the situation (i.e. ask me at one time I'll say one construct pole is the positive one, ask me another time and the opposite pole could be the positive one). For a group, when there is a split in the group count favouring each pole, such conflicted constructs tend to mean that there are individuals that strongly favour one pole and others that strongly favour the other.

Ruling out the Rioters

So around the flag question we have two subgroups that we should look at in more detail. In ipseus you can set up some rules to automatically split up the full group of 115 individuals.

Here is the rule set I created to split up the group based on the flag question:

01	[FlagConstruct] 9
02	[@Flags] (-1 = Polarity( Rule( "FlagConstruct" ) )) AND StructuralPressure( "Secondary" ) < StructuralPressure( Rule( "FlagConstruct" ) ))
03	[@NoFlags] (1 = Polarity( Rule( "FlagConstruct" ) )) AND StructuralPressure( "Secondary" ) < StructuralPressure( Rule( "FlagConstruct" ) ))

When you create rules you can give them a name (just put some text inside [] at the start). If the rule name starts with an @ it means that you want that rule to become a criteria.

The “Flags” rule says that I want anyone that has endorsed the left pole of the construct (i.e. polarity equals -1) and they have some level of conviction behind that endorsement (structural pressure is positive and relatively large).

So I now want to create two sets of subgroups. One set splits folks up based on their view of the flags issue. Another breaks up the full group along the line of their traditional cultures (PUL - Protestant Unionist Loyalist or CNR - Catholic Nationalist Republican).
As I said, my initial thinking was that only the PUL folk would care about flags. However, that is a stupid bias based on the news images on the telly only showing Union Jacks (sometimes upside down!). In truth the flags thing, or cultural symbols to be more generic, should be important to some degree to both sides of the political spectrum.

For Flapping Flags?

So the next step was to intersect the Flags/NoFlags subgroups with the PUL/CNR subgroups. Running my 'Flags' and 'Traditions' rule had created new sets in he 'Criteria' section.

We can see that there are 31 individuals that like flags with some conviction, while the majority, 52, would prefer to ditch the symbolism. When we intersect the two criteria of interest we should see the flags question fairly split across the communities. And lookie here:

[NoFlags][All CNRs]	30
[NoFlags][All PULs]	22
[Flags][All CNRs]	11
[Flags][All PULs]	20

So, these numbers do suggest that the whole flag thing is a little more of a PUL issue than a CNR issue. Lets find out why.

Flag waving Prods

Now we have those subgroups we should look at the ipseus profile for each subgroup separately. The immediate one of interest is our rioting friends; the pro-flag PULs.

In order to see the make up of this subgroup I'm jumping to the Construct Summary section and looking for a stable pattern i.e. a good bunch of core/secondary convictions with a smattering of conflicted ones. Holy kak! The flag waving PULs are anything but. We actually get the opposite.

Our flag construct (#9) is only one of two core constructs. The second core construct (#16 local parades important) being very much related to that issue. A third related construct (#8 trouble at parades ok) also stands out. Everything else is is slipping into the conflicted, or don't know which way to swing, zone! In simple terms this pattern suggests that the subgroup has no identity other than the symbols they use to identify themselves; Flags, Parades with a little bit of religion slipping in there at the edge. Now, here is the crux of my argument thus far; if you have a significant minority of people with no cultural identity other than a few symbols, and you remove one of symbols, what the hell do you expect to happen!

Turning the tables

I want to now do two other things. First I want to flip the table and see if we would get the same reaction if it was an Irish Tricolor taken down from the roof of Belfast city hall by contrasting the flag waving PULs with the flag waving CNRs. Second I want to suggest a different policy that could be followed without undermining core struts to folks identity. So, here is the same Construct Summary screen for the flag waving CNRs:

Overall this CNR subgroup has a more stable profile. Few core convictions to be true, but less conflicted ones as well. While the 'flag' symbolism is still important, the parades issue (#8) is deemed not worth the hassle in opposition to their PUL brethren. From this I would predict that, while there would be deep upset over taken down their flag, there wouldn't be such a violent reaction. Construct #20 “...living peacefully by knowing and respecting each other” has some conviction behind it after all!

You didn't want to do that...

So the numbers support the notion, predictably, that dropping the Union Jack from the top of Belfast city hall was a bit of a stupid idea.

What about another option like flying a less prominent Irish flag along side it while educating everyone what the 'gold' in the green, white and gold stands for? There is contention in that idea and you would have to watch that the conviction around construct #15 “...CNRs gaining advantage” stays conflicted for both groups.

Trying to change ideas held with strong convictions just doesn't work without pain. So for a better way forward we need to focus on constructs that are conflicted for both groups, preferable ones that both groups agree on which pole is positive and which negative.

For example, taking construct #2 “...police better than paramilitaries” and construct #21 “...Irish sports, language and arts should be enjoyed by all sections of the community” together then the PSNI should be ramming “culture” down everyone's throats until both sides are sick of it and move on!
To be less flippant, this could mean police enforcement of flying both flags, police enforcing parades only with both sides of the community or having joint Gardai/PSNI rugby/GAA matches.