No, you didn’t read that wrong. That’s really what our lovely government has decided, courtesy of its system for keeping tabs on people who take flights: “’Suspect’ requests likely to cause innocent holidaymakers to get ‘red flags’ as potential terrorists include ordering a vegetarian meal”.
You could smile and move on to read about something else.
You could get angry and move on to read about something else.
You could wonder quite what a paranoid conspiracy theorist would have to come up with these days to exceed what our government is actually getting up to.
Or you can do something about it.
It’s not just about our civil liberties, it is also about our safety. Any government that has got to the point where it is spending serious amounts of time and money worrying about whether or not someone eats lentils isn’t doing a good job protecting us from the seriously dangerous people – those with bombs, not lentils.
For all its flaws, we do live in a democracy and one that often listens to public opinion. So you don’t have to just sit there and let this happen. You can do something about it. Yes, you. No, don’t look over your shoulder as if there is someone else that was addressed to. You.
You and the computer you are sat at are an incredibly powerful tool. Campaigners through the centuries would have given their right arm, and quite a few other limbs, for publicity possibilities that you’ve got sat at the computer, right now.
So what can you do?
1. Share this post with your friends. Whether it’s Facebook, Digg, email, Twitter or printing it out and gluing it to your shirt – the more people who know about the story, the better.
2. Join the Facebook group and share it with your friends on Facebook.
3. Write to your MP. It only takes a moment. And you can do it right now, for free – at http://www.writetothem.com.
4. Back the Liberal Democrat Freedom Bill campaign. It’s true that the Bill doesn’t explicitly contain a clause protecting the rights of lentil eaters, but it is about rolling back the sort of crass and ineffective Big Brother policies of which this is the silliest extreme. Go to http://freedom.libdems.org.uk/petition/ and sign up. Already signed? Then email 10 of your friends to encourage them to sign. It’s as simple as that.
5. Make a donation. We’d love it if you gave us some money – as the bigger and better this site is, the more people will see stories like this one – but you could instead donate to the Liberal Democrats or a good cause such as Liberty.
6. Bookmark this page and come back and do another one of the above five steps later in the week.
After all, once they’ve started coming after the lentil eaters, it is getting serious isn’t it?
Follow @LibDemVoice on Twitter
Like us on Facebook
Subscribe to our feed
Sign-up for our daily email digest
Anyone try the West Cliff Lift?
Dropping the 1p tax rise has political costs too
Matthew Sweet on conspiracy theorists and the media
A blob-chart way of dissecting Britain's prosperity failure
A matter of trust
New housing planned for Fishmore Road - Environment Agency wants more information
Michael's weekly ward surgeries today! #dundeewestend
Ewan Hoyle
James Baillie
Tom
Bob Hale
53 Comments
Not quite sure why ordering a veggie inflight mile = raised chances of being a terrorist! is it some sort of fear of green militancy, or a misguided belief that islamic extremists don’t eat meat?
I think we should be told.
For the record, I usually opt for veggie in flight meals because i suspect there’s less chance of getting food poisoning from one!
* is so reminded of Angry Magrat in Lords and Ladies by this post *
I never trusted veggies and that bloke in the video’s obviously up to no good – more power to Jacqui, I say.
I always suspected my sister was a terrorist, and now I know.
“Join the Facebook group and share it with your friends on Facebook.”
I thought you wanted people to DO something.
Actually I can think of a few in-flight veg bakes that could have tipped me over the edge.
Anthony Bourdain in ‘Kitchen Confidential’ made some reference to “Vegetarians, and their Hezbollah-like splinter faction the Vegans, …”
Phrase stuck in the mind!
Cheers LDV, Next time I fly between caves, God Willing, I shall order the Lamb. No Infidel vegetarian choice for me!
Lovely quote Morus. Will remember that one!
It would be flagged up because moslems would not eat meat in case it is not halal.
Altho, if you were about to blow up the flight, ordering a halal meal might not be high on your todo list…
“It would be flagged up because moslems would not eat meat in case it is not halal.”
I suspected there would a good reason for it.
Now don’t the people who were making such a fuss about this feel a little bit sheepish?
Except Vicky that on many, if not most, airlines you can specify a Halal meal.
Ah, but terrorists wouldn’t specify a Halal meal for fear of giving themselves away. It all makes perfect sense if you think about it …
“Muslim” and “terrorist” aren’t interchangeable, you know!
Terrorists wouldn’t specify an El Al meal because ….
Though if I’m on a long haul flight and the person next to me orders a dish of lentils I can see good reasons for moving seats…..
First anonymous:
I was the second member of the Facebook group. I have a lot off vegetarian friends, many of whom are now members. The groups has on its frontpage some links to writetothem and ways of helping people actually campaign.
If all you do is join the group, then you acheive nothing, but if you join the group and do some of the other thigns Mark has suggested, like spread the word, write to your MP, etc, then it is worth it.
Facebook, on its own, does nothing. But using Facebook as a way of organising a campaing? That can be effective.
Anonymous1. No, not at all. This Govt has been labelling vegetarians as possible extremists for some time, as did the previous one–Vicky is wrong, that’s not the reason. Animal rights extremists are labelled under UK law as terrorists. Animal rights extremists tend to be vegetarian or vegan.
Ergo, if most animal rights extremists are vegetarian, it follows that vegetarians are animal rights extremists. It’s just as stupid as assuming all muslims are a threat, but what do we expect?
Sounds like vegetarians have so much scope for subversive connections that they’re well worth keeping a close eye on.
Isn’t this the same person exposed as having a collection of suspicious photographs a few weeks ago?
https://www.libdemvoice.org/why-mark-packs-awaiting-a-visit-from-special-branch-12277.html
Added to which of course he is the sort of person likely to be involved in fraud
I know a few people who’ve joined Liberty only to leave on finding that they don’t really /do/ anything.
Personally, I think that if you’re going to give your money to a civil liberties campaign, it’ll probably achieve more practical aims in the hands of NO2ID.
“Personally, I think that if you’re going to give your money to a civil liberties campaign, it’ll probably achieve more practical aims in the hands of NO2ID.”
?
“Practical aims?”. Do you really think a Tory government is going to introduce ID cards? Or that Labour is going to come out of the next election with an overall majority?
But Liberty is going to be very necessary whoever wins the next election. The recent willingness of the Lib Dems to sacrifice the rule of law to their short-term populist agenda has only confirmed that.
OK, so do we believe there should be spot checks on people entering or leaving planes in order to check against terrorism, drug smuggling etc?
If so, do we think the selection of those checked should be completely random, or should it be based on factors where there is a statistical link with the problem?
If it is not at complete random, should it be using mathematical algorithms which use the data in a dispassionate way, or by an officer doing it using rules which exist in his or her head?
Are there any Liberal Democrats who understand such mathematical concepts as logic and statistics, because what is written here so far suggests there aren’t.
Matthew: I agree that using data to steer crime-fighting is sensible. But it’s got to be good data analysed with sensible tools.
One of the key figures here is not “what proportion of terrorists eat vegetarian meals on planes?” (although the evidence I’ve seen on this is that the answer is ‘very few’) but rather “what proportion of those who eat vegetarian meals on plans are terrorists?”, to which the answer is a absolutely miniscule number, and it therefore makes this a very poor tool to use.
It’s something we see in plenty of other areas of policing: using a measure that generates overwhelmingly false positives doesn’t assist good policing, it hinders it.
So if you are a terrorist, make sure you order the standard ‘meat’ meal and just don’t eat the meat! After all, if you are intending to blow the plane up fairly soon, it’s not like you’ll get hungry!
Or will we now have airlines getting suspicious if you don’t clear your in-flight plastic containers? There’ll be a lot of us to arrest in that case!
Mark,
The article mentioned says nothing about the weighting given to ordering vegetarian meals or how it is combined with other factors, so we cannot make the conclusion from it that you make, that it generates large amounts of false positives. It may do, it may not. The article does not say, contrary to much of the comment here, that merely ordering a vegetarian meal makes one a “terrorist suspect”.
Given that the article is from the Daily Mail, one cannot expect any standards whatsoever of numerical or logical reasoning. Not that the other newspapers are that much better. It would be nice if newspapers employed journalists who had some sort of scientific background so that the many stories which have a science or statistics background get sensible reporting instead of sensationalist nonsense or embarrassing innumeracy.
It wouldn’t be amiss if stories reporting some scientific finding, such as the “X causes cancer” ones came with a reference to the paper where those interested could find out what the details actually are. After all, with half the population now meant to be going to university, most readers of the quality press ought to be familiar with citing references.
“The article mentioned says nothing about the weighting given to ordering vegetarian meals or how it is combined with other factors, so we cannot make the conclusion from it that you make, that it generates large amounts of false positives.”
You’re really arguing that vegetarianism could be a useful indicator of the likelihood that someone is a terrorist, whatever algorithm you stick it in, or whatever weighting you apply to it? Frankly, that’s just bizarre.
What I find mildly depressing, though, is that Mark seems to be arguing against this purely on the grounds of efficiency. Shouldn’t we be arguing that it’s just plain wrong for the security forces to be keeping tabs on the eating habits of the entire population, let alone attaching “red flags” to people on the basis of such patently ridiculous criteria – with whatever consequences that may imply.
You’re really arguing that vegetarianism could be a useful indicator of the likelihood that someone is a terrorist, whatever algorithm you stick it in, or whatever weighting you apply to it? Frankly, that’s just bizarre.
No, there is just one mention of vegetarian meals in an article from a newspaper which is notorious for publishing articles which show gross lack of understanding of any mathematical and logical argument.
I’m supposing this is some sort of data mining algorithm in which you just plug all the data you have into it, plus some training examples, and see what correlations it picks out. In which case, there would have been no human decision to pick out ordering vegetarian meals as a factor.
There are security checks at airports. People are picked out and their baggage searched. It has happened to me and when it did I could sort of guess the factors that might have led them to think I was worth looking at in more detail.
As I have said, should this be done at random, or should those picked out for more detailed checks than the average passenger be on the basis of some algorithm which uses factors shown by mathematical analysis to have a correlation with terrorism/drug-smuggling/ etc activity?
Personally, I think I’d rather like to know the scientific facts and mathematical reasoning behind this before jumping to conclusions based on the notion 1) you can believe what you read in the Daily Mail and 2) not only is what is in the Daily Mail true, but you can jump to all sorts of further conclusions based on assumptions that aren’t actually supported by the words it uses.
Matthew
Well, if you’re saying that the Daily Mail may have made up the claim that vegetarianism is a factor, that’s a completely different question.
But what I’m asking is whether you believe vegetarianism could be a useful indicator of whether someone is likely to be a terrorist, whatever algorithm is used, or whatever weighting is given to it.
If you want to suspend judgment on the basis that there may be some unrevealed “scientific facts and mathematical reasoning” behind this, that’s your affair – but don’t hold your breath while you’re waiting for the details to be made public!
I think a little common sense has already given most of us the answer.
But, as I suggested, I would have expected most liberals to be primarily concerned about the civil liberties aspect of this story, rather than whether this is an efficient use of resources.
But what I’m asking is whether you believe vegetarianism could be a useful indicator of whether someone is likely to be a terrorist, whatever algorithm is used, or whatever weighting is given to it.
What is being done here is data-mining. I do know a bit about that. It is clear you and others commentating here don’t.
Matthew
This is like pulling teeth.
So – with the benefit of whatever it is you claim to know about “data mining” – do you find it plausible that vegetarianism could be a useful indicator of whether someone is likely to be a terrorist? Or not?
Data-mining means you put in the data, true and false cases as training, and you see what correlations the algorithms give you. In this case, the algorithm seems to have come up with a correlation with ordering vegetarian meals. I have no idea why that should be. I note it only as a statistical fact rather than make any sort of value judgement about it. As anyone who knows about this sort of thing would realise, odd correlations often happen. One needs to look further to work out why they happen. Not having access to the data and experimental results, I am hardly in a position to comment any further than that.
“Not having access to the data and experimental results, I am hardly in a position to comment any further than that.”
Unless you’re willing to use a bit of common sense, like the rest of us.
I see both sides of this argument going on above.
No, from the outset, being a vegetarian does not mean that you are a terrorist, causing terror makes you a terrorist.
however, as pointed out above there are some useful indicators as to the types of people that are, or may be terrorist.
firstly, the animal rights issue was pointed out above, most, if not all animal rights activists do indeed tend to be vegetarian or vegan
there for when looking at a profile of 100 “terrorists” we see that of those 33% are vegan, 66% are vegetarian and 1% are omnivores.
therefore there is a greater statistical correlation that says that terrorist are more likely to be vegitarians,
when you board a flight, there is no check box that says are you a terrorist, (that form comes at immigration once you’re off the plane at the other end). and thus you have to approach the whole situation backwards.
In this case they may be looking at your name, area of residence, previous flights taken, and yes, eating habits, because it’s one factor in a number of other statistically significant other factors that *MAY* indicate that you’re a terrorist.
by no means is anyone saying that all vegetarians are terrorists, and nobody is saying that all terrorists are vegetarians, just that there are statistically more vegetarian terrorist than there are terrorists that eat meat. (supposedly).
When Matthew takes about weighting, what it means is that everything is given a score. then that score is compared against a median average, and if it is unusually high then action may be taken…
much like this.
consider the example.
person A
ethnicity: White British (score: 3)
sex: female (score: 1)
home town: stow on the wold, (in the middle of the Cotswold area) (score: 1)
previous flights: 7 business trips to America. (score: 3)
taking
meal choice: standard (score: 0)
total (8/50)
Person B
ethnicity: British Asian (score: 6)
sex: male (score: 8)
home town: Finsbury, north london. (score: 3)
previous flights: 2 personal flights to Pakistan (score: 7)
meal choice: veggi (score: 2)
total score 26.
so person B gets stopped as they scored more than half marks,
because statistically there are more city dwelling Asian men, who are vegetarians who are actually terrorists, than there are white business women living in the countryside who are terrorists.
yes, sexual/racial/class profiling is wrong. but it is the only tool available to try to guess who may be terrorists, and who may not be.
“by no means is anyone saying that all vegetarians are terrorists, and nobody is saying that all terrorists are vegetarians, just that there are statistically more vegetarian terrorist than there are terrorists that eat meat. (supposedly).”
“Supposedly” is the word.
Just out of curiosity, have animal rights activists ever blown up or hijacked a plane?
“Just out of curiosity, have animal rights activists ever blown up or hijacked a plane?”
(do you know? thought not, in which case I’ll say a thousand?)
I’d hazard a guess that none have… (at least very few)…
however, how many pane hijackers have blown up animal research labs? sent death threats to researchers and any company dealing with research labs, how many send packets of icing sugar to make people think that them and their family have received anthrax in the post?
I’d say equally as few, but it doesn’t make the terror that they cause any less, they are still terrorists.
how many climate protesters have delayed planes and set up dangerous runway obstructions?…
now you’re getting it…
You don’t need to have blown up a plane before to be considered a terrorist, but for some people there is always a next step along. -which is why you still get violence at peaceful protest.
Are you suggesting that terrorists of any kind are all innocent until they hijack the plane? send a death threat, blow up a lab etc… regardless of the intent to perhaps do so?
in this case, do we let the crime happen and put whatever pieces of the person are left on trial? or do we make effort to detect and stop the person before they board said plane? approach said building?
I said supposedly for the simple reason that everyone knows that for this country at least (historically) there are possibly far more white meat eating males of Irish decent that have proven to be terrorists than there are young Asian chaps. -who seem to be the current focus of anti-terror police.
would you suggest that we should stop all Irish people travelling as historically speaking in this country there are far more terrorist attacks committed by this demographic. or would you agree, that rather than living in the past you need to look at current and future threats, and that may well be young Muslim men, who happen, to be vegetarian, or order veggie meals, that it may be animal rights activist who again happen to be vegetarian?
in any case, being stopped and searched/questioned at customs when going either outbound or inbound is of no real consequence, personally I’ve been stopped and searched 3 times in my life, and been with a travelling companion who has been search once before as well. it was no hassle. five minutes out of my day. half an hour at most.
now, I usually hate the argument “if you’ve got nothing to hide” but seriously, if you’ve got no guns/knives/drugs/bombs etc then what’s the problem with having five minutes out of your day for a chat with, (in my experience usually very friendly) customs officers?
it’s not a hardship. it’s a safety feature.
back to what you said though…
it doesn’t matter that there have to date been probably no animal rights activist attacks on planes. the question is what if the people who are threatening other companies find out that supplies for the research labs at Oxford university are transported by air, by say BAA. now there is known history, and empirical evidence that these activist groups after failing to stop the new labs there being built after terrorising the university staff, and then the building contractors, have started targeting suppliers and others who do business with the department within the university. so like I say. how long might it be before someone says BAA brings the supplies [they use] to this country so we’ll threaten them and their business.
(yes, I used the word might, because it’s not a certainty, same as that word supposedly). the point is it could happen. it’s not unrealistic.
now, we go back to the point that this veggie weighting is likely a very low weighing in this likely complex system of statistically rating likely terrorist suspects.
anyway, that’s completely off subject.
the point is that there are vegetarian terrorist, there are specific terrorist organisations that appeal to vegetarians, and the specific weighting given to meal choice, (unless you choose spaghetti made of fuse wire as your meal) is likely to be extremely low.
more importantly… that word supposedly was used because I, sitting here in my chair, don’t have relevant figures to be able to say for sure.
perhaps (since I doubt anyone else has either) you should re-read all these posts and mentally add that word to everything else as well?
of course, that doesn’t appeal to you. you say “just use common sense”.
I did use common sense. I did back up my common sense with specific arguments, and you choose to ignore them and just think that all vegetarians are nice people who wouldn’t hurt a thing, well certainly not an animal… (supposedly)-and I say supposedly this time in a mocking fashion since there are statements above that can be backed up with news stories regarding these nice wouldn’t hurt a fly veggies actually hurting threatening and intimidating, (we’ll wrap all that up in the word terrorising) other people.
I fear that in writing so much my argument has been lost.
so I’ll state again, as others have above.
it’s all about the statistical likelihood. and a weighted marking score being given for each attribute of a person.
being called Abdul, might give you a score of 97, which being a veggie might only add 1 to your score count, and then we see that eating habits, whilst being counted, are weighted so low as to be almost insignificant.
a points based scoring and weighting system is the same way that spam filters work for email as well. and we all know that they are not perfect either. some systems are better than others.
if you can’t see that there are terrorists that are vegetarian then you’re just ignoring facts
if you can’t understand the concept behind statistical weighting and probability then there are better places to learn about it than comments fields on web blogs.
Dan
In case you didn’t notice, the article we’re discussing was about assessing the risk posed by air passengers.
I see…
so you decide to ignore facts and change the subject then?
Might have known… this is a political blog after all.
or are you saying that certain terrorists should be allowed to travel freely whilst others should be restricted?
Dan
We’re not talking about “terrorists being allowed to travel freely”. We’re talking about a system that’s meant to assess the risk posed by air passengers. On your own account, animal rights activists don’t appear to pose much of a risk as air passengers.
If you’re advocating a wider strategy of giving everyone a risk rating and restricting the movements of people whose rating is too high, then that’s not what this thread is about, and it’s not something I would expect liberals – or anyone in their right minds – to support.
“If you’re advocating a wider strategy of giving everyone a risk rating and restricting the movements of people whose rating is too high,”
no you;re making it painfully obvious that you don’t understand the point of a weighting system.
yes, everyone is given a risk rating already, that’s exactly why this veggie option, is there everyone already has a risk assessment, those with a higher risk assessment are singled out for further processing.
(if you want to campaign against that you;re already too late)
we’ll still go with the example of animal rights activists, once upon a time, people wouldn’t have dreamed of liberating animals from research centres, -then it happened,
but people didn’t dream of destroying research buildings, -then it happened
then people didn’t dream that researchers, their family, and suppliers to their business and their family would be terrorised, have their property vandalised, have their kids harassed outside of schools -oh but then it happened
now you can’t dream of an animal rights activist blowing up a plane. -i wonder how long till it happens.
everything that was once considered impossible of implausible eventually happens (whether you’d like to admit it or not).
and I would expect every known terrorist to be double checked before they were allowed on a plane, regardless of whether plane hijacking was their thing, or just something that they may or may not decide to do one day.
and as such, I’d suspect anyone who may be profiled as a terrorist by statistical correlation of known facts to be given special attention, we’re not talking about cavity searches for all veggies here, we’re talking about people having a slightly deeper look, possibly without the passenger knowing into their flight plans etc…
I don’t understand the problem here,
you seem to understand that before flights that passenger lists etc do need to be checked.
you understand that there can be a given, possibly significant correlation between a particular eating habit and terrorist activity.
you understand the idea of statistical weighting, and the fact that this might be the least significant part of a much larger picture of information that is looked at.
but you’re still choosing to go with the sensationalist headline, like somehow all veggies are going to be denied air travel… when that’s simply not the case.
Dan
Why not try to condense your point of view into two or three sentences of correctly written, comprehensible English?
If you can do that, it may be worth responding to.
We’re trivialising a serious issue here.
You can argue, if you like, that for civil liberties reasons we should not keep any data at all on air passengers, and we should just rely on catching terrorists at the airports. You will then have to make your appeal for votes in the wake of the next terrorist outrage.
Or you can argue that we should keep all the data we can get hold of, and trawl it minutely to optimise your risk ranking and mount surveillance on the people who you predict are those most likely to be terrorists. Now there are various arguments against doing that, but, making a silly joke about vegetarians isn’t one of them.
As somebody commented, a much more reliable risk indicator would be the name Abdul. That’s bound to be much better correlated with terrorism than ordering a blinking veggie meal. So that’s what to use…. or is it?
…But if we go and hassle all our Muslim passengers, we are going to cause some very justifiable resentment, aren’t we? It’s just like the offence of “driving while black”. And yet, it will probably help catch a few of the right people. That’s where the hard choices come in.
Anonymous1
“Not having access to the data and experimental results, I am hardly in a position to comment any further than that.”
Unless you’re willing to use a bit of common sense, like the rest of us.
I’m noting that what is being used here is mathematical reasoning. Mathematical reasoning is what is used to design a plane that flies. Common sense tells us a heavy thing like a plane can’t fly.
Mathematical reasoning here, so far as I can work out from what is reported in the Daily Mail, and a little bit of search on relevant websites (for obvious reasons they don’t say much, but they confirm it is data-mining they’re using) has noted a correlation between ordering vegetarian meals and being a passenger who is dangerous. The correlation may be very slight, I’m sure that would be the case. It does not mean, as some in this thread have suggested, that every passenger who orders a vegetarian meal is a terrorist suspect. The correlation may be a false one which has come about from some aspect of the training data, and with more data it would go away.
Some have advanced possible reasons as to why this correlation might have been thrown up. I have not, I have simply noted that the correlation seems to have come from analysis of the data, there has not been a human decision to put this as a factor in picking out passengers for further search. So when you are asking me whether this is plausible, you are asking me for my opinion on whether the algorithm used is a good one or not. That is why I cannot answer, because I don’t have any details about it.
Dan
Why not try to condense your point of view into two or three sentences of correctly written, comprehensible English?
If you can do that, it may be worth responding to.
frankly, I find that offensive. -all my points have been written concisely and in plain English. the last bastion of someone loosing an argument is that they will go on to insult the other person…
Generally I felt the need to expand the points since your one line answers of, well I use common sense, just rather smacks of ignorance, you;re neither capable nor willing to listen to another point of view and understand it, even when evidence it put before you.
anyway. as you wish. bullets points.
>>There are vegetarian terrorists, and terrorist organisations that appeal to vegetarians more.
>>whilst said organisations have not targeted air travel yet, that does not mean that they won’t (as I said there is always a bigger better target when your point isn’t heard the first time).
>>regardless of the usual mode of terrorist attacks, no known terrorist should be allowed to board a plane without first being screened more closely. (I’d suspect an interview would be the best course)
>>second to known terrorists being screened more closely, I’d expect that statistically likely suspects would also be screened more closely, that may be vegetarians for the reasons given, (over and over and over again). though you would likely call these soft suspects, and additional screening might only go as far as a more thorough x-ray scanning of their baggage.
>> lastly, since the point has been made, Abdul, is quite a popular name, yet there are probably only a handful of famous terrorists with this name, perhaps a few thousand less notorious terrorists with this name, I wonder what the statistical likelyhood (world wide) of being called Abdul and being linked to a terrorist organisation is. compared to being Vegetarian and being linked to a terrorist organisation?
>> and I wonder what the stats look like for that in the UK, especially when you consider that organisations like PETA, have knowingly funded terrorists.
http://www.fsc-watch.org/archives/2007/05/31/_Legality___SmartWood_style
ass have the World wildlife foundation and their support charities.
(all charities supported by lots of UK people, all having known terrorist contacts, and fund terrorist activities).
>>that is to say that being a vegetarian is probably a better statistical indicator of being a terrorist than being called Abdul.
“all my points have been written concisely”
Oh, come on! By no stretch of the imagination could the inane verbal diarrhoea you’re coming out with be considered concise.
Matthew Huntbach
For someone who is “hardly in a position to comment any further”, that’s quite a lot of comments.
But I note that you say “The correlation may be very slight, I’m sure that would be the case” and “The correlation may be a false one … and with more data it would go away”, by which I presume you mean it may not be statistically significant.
Call me an old fashioned liberal, but I think as a bare minimum these security bods ought to demonstrate that it is statistically significant before they ask for permission to store this data and use it in giving people “red flags” as potential terrorists.
And even if that were shown to be the case, I think there should be the most careful and sceptical consideration, and – perhaps a dangerous thought, this – searching public debate before they are allowed to do so.
The original article suggested there had been a deliberate human decision to collect data on who orders vegetarian meals and to use that to flag those people as potential terrorists.
A brief look at what was actually being done is that ALL data used when buying plane tickets was collected, and fed into a system which sees if it can spot any correlations. If all data is collected, then it will include meal ordering data, but that was not singled out to be included. There is a very brief mention which suggests a correlation was found with ordering vegetarian meals. This may have been an early case, data-mining tends to throw up odd correlations which go away as you give it more training cases.
Therefore the claim that there was some deliberate targeting of vegetarians was ridiculous, and anyone with a bit of scientific sense and ability to check a few things on Google ought to have been able to see this.
The wider question is whether this data should be collected and put into state ownership or not. I did not venture an opinion on this, my only point was to note that the suggestion there was a deliberate targeting of vegetarians was wrong.
I do not know what it means to “red flag” someone as a potential terrorist, and that is not stated in the article. I know some people on travelling are picked out for more detailed search than others – it has happened to me. I think this is done through some sort of system which identifies possible signals of problem cases. The problem cases are much more likely to be drug smugglers than terrorist, and a further check of the Daily Mail item and other information on this system found on the web does indeed show the system isn’t aimed just at detecting potential terrorists. Smuggling of food items which could cause an outbreak of animal diseases, or of material obtained from endangered species is also an issue.
It seems to me to be worth asking do you agree with picking out passengers for extra search, and if so, should there be any use of factors which are shown to be correlated to problem cases. If you still say yes, then comes the question about whether you agree to the collection of data which would make this procedure more accurate.
“I know some people on travelling are picked out for more detailed search than others – it has happened to me.”
How many people have been charged or convicted of terrorism offences purely following such stop & searches?
Matthew
If you read the article, you’ll see i’s _not_ just about extra searches.
The article talks about people being “red flagged”. It does not say what is meant by that. I was assuming it meant being chosen for further search.
However, the point I was making was not on whether some system like this is good or bad, or on whether keeping and using the flight-booking data was good or bad. It was only to note there was no human decision made, so far as I could see, to flag-up vegetarian meal ordering as a problem issue. This was something thrown up by the analysis algorithm and not but a human being saying “veggies are trouble makers, let’s ‘red flag’ them”.
“At the heart of the system is a highly classified computer algorithm designed to pick out people to be searched, questioned by security staff or barred from flying.”
Matthew: throw enough data at an algorithm and it will find spurious correlations which pass tests for things such as statistical significant but don’t actually have any predictive power. Overlooking this point is a moderately common statistical error, so who knows if that’s what has happened in this case, but if the maths is being done well, then there certainly does need to be the intervention of humans at some point to make judgements about what may or may not be spurious and what statistical models to test.