Threat levels: doing the numbers
Being interested in world event, and also a student of statistics, I'm always intrigued by the language of politicians when it comes to threat or risk.
For instance, this week we reached the dizzying heights of a 'CRITICAL' terrorist threat level, which is distinguishable from 'SEVERE' simply by the imminent nature of the threat. However, when we were downgraded to 'SEVERE' this morning, the questions started to surface. Recently politicians haven't said much of the terror threat level, but in the last few days I'm constantly hearing about how a threat is 'highly likely' now we are at the SEVERE level.
What does that mean? It seems it can be misleading in two distinct ways:
- How likely is highly likely? It suggest a very high probability that an attack will happen, but what is that? 0.5? 0.7? 0.99? More than likely it is an extremely low probability figure if we talk of an attack in the near future, maybe less than 0.01. Certainly, the impact of such an event could be catastrophic, which may justify using such language in a relative sense.
- Once we retreat from the immediacy of an event (ie the CRITICAL level), we have no reference to the timespan of these threats. We are expected to assume the level is describing the 'climate' in which we live. But does this mean an attack is actually highly likely? The language used does not make sense if we consider the intended message..
Perhaps I shouldn't think so much about it. Maybe they should publish estimated probability figures with 95% confidence intervals.. That would satisfy pedants like me.
14 Aug 2006 15:04
| 
Tags: Terrorism Uk
| 
Comments (4)
| 
Report a problem
Loading…
Surely, by definition, terrorist attacks should come without warning. This is demonstrated by the fact that last July’s bombings occurred the morning or the morning after a downgrading of the alert levels. If the security services know where these people are, then they apparently have a very good track record of jumping on them thus far. For John Reid’s department to say that they have all the known plotters, but they don’t know how many unknown plotters there are simply moves the situation into the surreal. Which, according to my first sentence is about where I came in….
14 Aug 2006, 15:30
Everybody knows it’s not about the actual risk level, it’s about justifying scaring the bejeesus out of little voters.
14 Aug 2006, 16:21
I don’t know what it’s supposed to mean, but I don’t believe it can be a probability because it seems implausible that one could have a model sufficient to generate a meaningful probability. I think Olivia got it right.
14 Aug 2006, 16:37
I think what they mean to say is that “something may happen somewhere in the world at some time”... shifty eye…
as for the probability, Scott, I don’t know what they could or should use…
I kinda why I feel it is stupid to use the highest level at times like this for as long as they did, they stopped it, use critical when the P=0.99 or something close, when they are pretty confident that an attack will happen on that day and they don’t reasonably think they can stop it… then at least people could stay indoors. If we were in a situation where an attack was a virtual certainty on that day how would we up the level?
14 Aug 2006, 22:13
Add a comment
You are not allowed to comment on this entry as it has restricted commenting permissions.