A new study contends that resentment of first class leads to air rage in coach.
It found that passengers in economy seating were 3.84 times more likely to have an incident of air rage if they were on a plane that had a first-class section. They were 2.18 times more likely to have an outburst if they had to walk through first class to board the plane, as opposed to boarding in the middle of the plane, directly into the economy section.
American Airlines Boeing 787 Economy
That’s the claim, anyway, and how the study is being reported. The underlying research is in the May 2 issue of Proceedings of the National Academy of Sciences and I decided to spend the $10 for two days of access to the article so I could see what’s actually being claimed.
- Data is circa 2010 from just a single airline, and the airline isn’t identified. We do know that the airline in question isn’t United because researchers weren’t able to evaluate whether extra legroom economy seating had any effect since “this feature was implemented by the airline in specific planes—not in entire plane models” over the period of years surround 2010 data being looked at.
Since American didn’t announce Main Cebin Extra until March 2012, and Delta’s announcement came a year earlier, it seems somewhat more likely that the data in question is Delta’s.
- The study acknowledges that longer flights and delayed flights correlate with air rate incidents. They just suggest that first class cabins are a bigger factor. They aren’t able to show that first class cabins drive air rage independent of flight distance. Remember that while there are certainly long flights on single cabin aircraft (and single cabin regional jets can feel even longer still), longer flights are more likely to have first class cabins.
- They find that boarding from the front of the plane correlates with greater instances of air rage. They also find that domestic flights are more likely to have instances of air rage than international flights for this one U.S. airline. Of course, smaller domestic narrowbody aircraft (737s, Airbus A320s) are more likely to board from the front of the plane.
American Airlines Boeing 737 Economy
- Since the only data they’re looking at is “flight number and date” as well as “seating class, gender, and incident type (e.g., belligerent behavior or emotional outburst)” they don’t have the ability to even look for or test alternate causes like alcohol or drugs, medical conditions, or extreme emotional disturbances passengers are bringing onto the plane. They didn’t have data on the load factor of flights with and without incidents.
The presence of a premium cabin didn’t cause a 24 year old passenger to attack another passenger — an overdose did. A premium cabin didn’t cause an Air India passenger to urinate in the aisle on his flight — anti-depressants and whiskey did. This American Airlines passenger urinated at his seat also because he drunk. Here a passenger melted down due to delays. After Blac Chyna’s meltdown she was arrested for drug possession.
The authors here want to “posit that the modern airplane is a social microcosm of class-based society” and that “inequality can trigger antisocial behavior” and thus inequality is bad. Unfortunately they’re using travel for their own purposes, not really trying to understand travel.
All they’re able to do here is correlate disruptive behavior with the presence of a first class cabin, without rejecting alternative hypotheses. That doesn’t get us very far — especially when they don’t seem to understand the subject matter.
They isolate air rage to domestic flights with first class, and suggest this is of growing importance because “first class cabins claiming an increasingly large share of total space” which is of course false. Take American’s retrofit of their legacy US Airways Airbus A319 aircraft, chopping those down to two rows of first class (and replacing MD80s with four rows of first with those same A319s with two). Similarly, American has been cutting down the number of business class seats as they’ve improved their seating.
Air travel is increasingly affordable — as prices are so much lower than before deregulation, and society is wealthier. Ultra low cost carriers make travel more affordable still. And planes run full. Airlines bring together diverse people from across society, it’s no longer the province of the wealthy and (mostly male) busines travelers. And it puts them together in close quarters. Sometimes with alcohol and drugs. And yet we think that it’s first class seats that’s he primary problem?
It’s usually the free loaders who don’t get their upgrades that have most of the rage.
Hey Gary, no need to pay the $10 next time. If you have a GMU.edu email address, you should have access to pretty much any academic journal.
@Mark I’d agree. Me purchasing a DL First Class buy-up for $40 causing the Plat to miss their upgrade is a pretty likely culprit.
Thanks DL!
And I believe they did control for flight length. Each of the three regressions presented in the paper has a control for “Flight distance in miles,” which is statistically significant, but it appears that there is still an effect from the existence of first class (see Model #1).
I’ve never understood this.
Everywhere else in life you can choose to pay more to get something better, and everyone is fine with that. Cars, houses, clothing, food, furniture, and services like postage, repairs, car rentals, seats at a concert and so on.
If you pay more to get a bigger seat on a plane people go crazy.
Hmmm…does road rage increase in the presence of Cadillacs, Lincolns and Corvettes?
@Jimmy H – they found a correlation for flight length but aren’t able to distinguish long flights from short ones to separate out first class as a factor, and even if they did there isn’t enough data precisely because longer flights have first
Great analysis Gary! I’m also surprised you don’t have library access due to your job.
There are 3 ways to publish in PNAS. The easiest is to have your friend (who’s a member) submit it on your behalf. This is why it states “Edited by Susan T. Fiske.” Not surprised it was edited given the shaky experimental design! #Science
Wow, you are seriously jumping to a lot of conclusions based on your own biases. Where did the research findings EVER state or infer that first class seats are the primary problem. Stating the results of things that FACTOR into air rage is not condemning one thing or the other.
Logic does suggest that walking through a first class cabin to your crummy economy seat would be one more thing that could piss you off. I mean, envy is as old as humankind. But is it statistically measurable? Beats me. It would also seem to be far less important than other factors. You wouldn’t know it from the headlines, but I’m sure rich people are less likely to “cause trouble” than poor people (they should have less to compain about, right?). So airlines that cater to the hoi polloi — aka, discount airlines without first class cabins — would seem more likely to have air rage incidents. You know, things like drunk bachelor parties gone wrong (I think I saw that headline this week). You do always read about over-entitled rich idiots behaving badly on airplanes, but I kind of assume that’s a man-bites-dog story. It’s more unusual, and more interesting, so it gets more attention.
I’m not anywhere suggesting that it couldn’t factor into some air rage incident somewhere, just that the authors went looking to find this but their study doesn’t appear to support the claim.
I’m also not suggesting that the authors believe it is the only cause, they do not make that claim.
Lets assume for a moment that the study is valid, and that, indeed, Y pax on AA and UA are more likely to go into air rage than their counterparts on Spirit and Allegiant.
I suggest there may be an alternate causality – Y pax on legacy carriers have greater built in expectations than their counterparts on no F cabin LCC’s who damn well know to gate check their expectations. Legacy carriers are therefore more prone to disappoint Y pax.
@iahphx I wonder what the results would look like if we excluded Conrad Hilton from the data
http://viewfromthewing.com/2015/02/03/conrad-hilton-arrested-went-berserk-london-los-angeles-flight/
I’d suggest that if booze were eliminated, many problems would be solved. Even if people arrived on the plane after having “one too many”, they wouldn’t be able to add to their condition if booze were not served on board.
Didn’t read the article, but trying to understand the critique that the authors did not control for flight length. Jeremy H says that they did by including a control variable measuring flight distance in miles. Gary then says “they found a correlation for flight length but aren’t able to distinguish long flights from short ones…” If in fact the authors did control for flight length, then I don’t understand Gary’s response. Is it that the authors should have controlled for flight length differently?
There is a simple solution here — have the peasants enter from the rear and place a steel door between coach and first so they cannot enter first class! This would also solve the first class lavatory usage problem.
They are saying that flight length is a factor, they aren’t demonstrating that first vs no first class is a factor independent of flight length. Two very different things.
Class warfare in the air makes snakes on a plane look like a Merry-Go-Round. There’s been a Hollywood script in development ever since lunch.
I should probably just read the paper, but what is the regression that they run? If they ran regressions of whether a flight had an incident on whether that flight had a first class cabin and the flight distance, and they found a significantly positive coefficient on whether the flight had a first class cabin, then they are demonstrating that first class vs first class if a factor independent of flight length (putting aside all other likely issues with the paper).
How did the authors define “air rage”?
Here are the regressions:
https://docs.google.com/drawings/d/1QNK2_48UUpVZZjGd8V6D2ykf77gKmz0fflU332G7Hs8/edit?usp=sharing
I’ve circled the control for flight distance. Now of course, none of this implies causation, and their are potentially many important omitted variables. But it is inaccurate to say that they didn’t control for flight distance. They’ve done so in the most conventional way possible, and “first class present” still shows up as statistically significant in Model 1 and “boarding from front” is still statistically significant in Models 2 and 3.
They definitely have distance in the model. But the study isn’t able to show that first class cabins drive air rage independent of flight distance.
While there are a handful of longer domestic flights on single cabin aircraft those are few and far between. Nearly all long flights have a first class cabin. The data is exclusively from a single airline that does have a first class product, either Delta or American (so Southwest isn’t a factor).
I don’t quite know what you mean by “independent of flight distance.” I think you are using “independent of” in a different way than it is used in econometrics.
Are you saying you would like to see a similar regression for a subset of the data, using only short flights, say, those less than 1,000 miles? Would be interesting, but I doubt you’ll get a different result.
I think the best way to respond to this result is to find a better control that is easily measurable.
Here’s a good response:
http://andrewgelman.com/2016/05/03/ahhhh-ppnas/
Let’s assume we’re talking about American data. Doesn’t matter, this would be the same if it’s Delta or United.
American’s 8000 mile flight, Dallas – Hong Kong and indeed all their transoceanic flying offers premium cabin seats.
Their 2000+ mile transcon flights all have premium cabin seats.
Their 1500+ mile midcons all have premium cabin seats.
Nearly all 1000+ mile flights in the US have a first class cabin. There’s a handful of outliers, like Phoenix – Des Moines at 1149 miles. But most 1000 mile regional jet flights, even, are on larger planes with premium seats.
So what does it even mean to control for distance at that point? That it doesn’t much matter whether a flight is 60 minutes or 90 minutes, the effect is the same?
My point is that claiming to “control for” flight distance when considering the effect of a first class cabin on air rage incidents does not mean anything other than offering a rhetorical dodge.
Perhaps some humor can be found in the fact The Proceedings of the National Academy of Sciences website is “pnas dot org”
Now how can you believe any research published there?
Thanks for exposing the junk science behind this “report.” Now if we could only do something about this global warming hoax…
The authors also concluded that banks cause bank robbery, homes cause burglary, living people cause murder (okay, I guess they do), and women who dress provocatively were asking for it.
@Gary Leff — Yeah, the Conrad Hiltons of the world are an interesting angle to this. Is air rage really much of a problem in first class? I certainly haven’t seen it, but then I’ve never witnessed ANY air rage in the couple thousand flights I’ve taken. As noted above, my guess is that the celebrities-behaving-badly stories get publicity because celebrities get publicity — not because such activities are common. The only caveat to my theory is that I could believe that free somewhat-unlimited booze MIGHT cause first class passengers to behave worse than you’d otherwise expect.
As far as travellers go, though, I will tell you that I prefer fancier hotels to less fancy hotels not so much because I love the amenities, but because fancier hotels tend to be quieter than cheaper hotels. I can tell you with 100% certainty that people who pay more for their hotel rooms are quieter and less annoying than people who pay less. Maybe this dynamic doesn’t work in the air, but I’d be a little surprised if it didn’t.
Correlation does not imply causation. This is a significant flaw in most studies such as this.
@ The Masked Poser- LOL- you want to see air rage, cut me off from any drink on an 17 hour transpacific flight! People gonna be breaking into their duty free, stealing from first class, business class, etc…
Gary,
Your comments are right on. As an author/researcher of several dozen published medical papers in refereed journals, it is so important to “go through the weeds” in a publication like this. It would be so easy, based on this paper, to jump to a quick (and incorrect) conclusion. I honestly didn’t think the paper was worth $10 to go beyond the abstract, but I was able to view the SI (supporting information) for free. This gives an overview of the methodology and lists the odds ratios of a number of variables (class, day of week, length of flight, etc.). There are a hundred reasons why the results must be taken with a grain of salt-many have been pointed out by readers. The authors note that the threw out the results from 59,000 flights because of missing data. Wow!. While it’s good that these 2 researchers went to the effort to look at this issue, the ONLY thing that can be said about the results is: “More research is needed to corroborate or refute these results and to better understand the multi-factoral nature of air rage incidents.”
Clearly, the paper does not support that the presence of premium seats cause air rage incidents. However, all of this reference to “controlling” for flight length raises a question in my mind – are there any long distance or international flights that don’t have a premium cabin? Even Condor and Norwegian have business class on all international flights, don’t they? So how do you control for a variable that doesn’t exist? (Also, I’d just be curious whether there are any.)
I would just say that every really bad incident I have had on a flight over the years has been in international economy and due to a fellow passenger. Business class did not cause that problem, but getting into business class sure has reduced it. Of course, that is an anecdote, not data.
Good point, Arthur. Walking by the comfy biz class seats may slightly annoy someone, but it isn’t very likely to make you fly off the handle. But that insanely annoying pax sitting in very close proximity to you in coach very well could provoke an incident. Up front, unless you encounter The Most Annoying Person in the World, there’s usually enough space to tune out annoying behavior. In coach, such behavior could literally be in your face, and I could see how folks would be more likely to snap.
PNAS is a legitimate science journal. But, jokingly referred to by us scientists as P-NAZ. Get,it?
Class welfare and the air rage between passengers make the conflict and the bitch fight between United Flight Attendants (Sub UA) on the LAX-SYD route a merry-go-around.
From my experience working for UA as a Flight Attendant for years and also doing travel journalism and flying around the globe as a real paid passenger, I have never seen such psychotic people in my life that can get into cat fights, about if they should serve the pre departure champagne. or if they should skip the first beverage service and do a combine Beverage/Meal so they can get more rest times!
Maybe they should study the UA Flight Attendants and see why the job attracts so many clinically bi-polar mean spirited Flight Attendants.
UA Flight Attendants out of rage would spit in passengers meals, drop it on the floor, step on it and serve it, they put minis in each other’s bag try to get each other fired. They write false flight reports to get passengers arrested out of hate and other Flight Attendants to get fired.
All the care is about the break time and the fact is the United Flight Attendants are on sleep meds the whole flight and they are practically in Coma when they wake up from their breaks as the affects of sleep meds do not wear off that quick and I can not imagine if they have to evacuate an airplane. %99 percent will be useless. They bicker and bitch and complain and they gossip the whole flight and god knows if a passenger pisses them off..Why are the UA Flight attendants are always pissed off and get into arguments and air rage? They should study their behavior! Air rage among United Flight attendants is very common!
I suspect alcohol is a significant factor. Flights with a first class cabin are more likely to offer free alcohol and are more likely to be longer, meaning much more time for passengers to drink and get drunk.
Reminds me of the Seinfeld episode where Jerry imagines that the flight attendants are giving looks that say ‘well, if only you worked harder…’ when they close that first/business class curtain….
Lol. This kind of study is made to order. The order in general is like this: the result is A, and the reason is B. Now please, dear researcher kindly make a paper to justify it. I will sponsored the study, as long as the final conclusion in accordance to my previous instruction.