Yesterday’s Thoughts

June 4, 2005

Obedience Equals Death?

Are you more likely to die if you do what you are told?

Wired has a story from Gary Wolf about a report from the National Institute on Standards and Technology called Occupant Behavior, Egress, and Emergency Communications. The whole report doesn’t appear to be on line and the site is down, so I looked at the Google cache of the Draft for Public Comment of the Executive Summary.

The report looked at the surviving occupants of the World Trade Center Twin Towers. Although up to 2,180 occupants of the Towers died in the attacks, 17,400 ± 1,180 successfully evacuated.

They randomly surveyed about 800 survivors, asking what they knew and did before, during and after the attacks. They asked if the survivors had participated in a fire drill, knew the location of the exit stairwells, had used the stairs, what instructions they were given over the PA and by guards, what other communications they had had, etc.

The executive summary does not say this, but according to Wolf, if you want to survive, “disobey authority. In a connected world, ordinary people often have access to better information than officials do.”

Wolf’s conclusion is provocative:

Anybody who has been paying attention probably suspects that if we rely on orders from above to protect us, we’ll be in terrible shape. But in a networked era, we have increasing opportunities to help ourselves. This is the real source of homeland security: not authoritarian schemes of surveillance and punishment, but multichannel networks of advice, information, and mutual aid.

I can’t imagine anyone who has ever stood in a line at a security checkpoint, or listened to the lies of an airline employee would be inclined to mindlessly obey instructions in what seemed like a life or death situation.

But Wolf’s conclusion is a speculation based on N=1 reasoning. Yes, in the case of the Twin Towers on the morning of September 11, 2001, it was probably better to disobey orders. But even as it was, the odds of surviving were pretty high. According to the executive summary, “only seven occupants (of WTC 2) who started evacuating below the impact region were known not to have survived,” and “only 11 occupants initially below the 78th floor were killed.” There doesn’t seem to be much to discriminate between Wolf’s hypothesis and its converse. Where are the examples of those below the impact region who obeyed authority and died? Was that the 7 or 11 people who died?

The population who could shed light on the conclusion are those above the impact zone in WTC 2. Those survivors who disobeyed authorities and left in the period between the first attack and the second attack support the conclusion; as do those who died because they obeyed authorities and remained. Sadly, there is no evidence from the second group, so we don’t even know if there were any members of this group. Maybe everyone who died was also disobeying authority.

Quotes like this one from the Google cache of an earlier presentation do not lend themselves to any single conclusion. This is from a survivor from a floor between 90 and 99:

We were instructed to go back to our offices and that the building was secure. So, I started back to my office. As I exited the fire escape and met my supervisor, I was told that he wanted me to leave the building and not follow the instruction to stay.

Yes, it was fortunate that this survivor disobeyed the authority who told them to stay, but also fortunate that they obeyed the other authority, their supervisor, who told them to evacuate.

Then there are dueling accounts,

Heard the building was secure by announcement… Kept moving down the stairs to exit the building. Came to a uniformed person holding the door open in the 44th floor sky lobby. We were asked to exit the stairwell area and to use the elevators to go back up, [but we] kept on going to exit the building.


‘I was raised in Los Angeles; I was taught that in the case of an emergency situations and disasters to avoid elevators.’ (100s) – Decision was made to use stairs until the PA announcement, when respondent used an elevator to evacuate.

The first respondent disobeyed authority twice, once walking down in spite of being told the building was secure, once by continuing to exit when instructed to return to office. The second respondent obeyed authority twice, once by not using the elevator because of childhood training, once by using the elevator in response to the PA announcement. What conclusion does that lead one to?

It is easy to agree with Wolf. The notion of disobeying authority confirms so many people’s opinions. Libertarians agree with him. Liberals of some stripes agree with him. Conservatives of other stripes agree with him. People rebelling from the conservatism of their family and community agree with him. People rebelling from big government and big media agree with him.

Even I agree with him. But, I agree with a hypothesis that is open to study, not a proven fact. It is pretty easy to imagine a scenario where the converse rule applied. His argument depends on the asymmetry of information present at the World Trade Center on the morning of September 11th. This asymmetry is not always present, and certainly cannot be covered by a blanket conclusion to “Disobey authority.” For instance, I predict that a similar analysis of responses to the mailings of Anthrax to various offices and government sites, would not support the conclusion.

This story has generated 37 technorati links as of this time. Looking through them, and following the links, I find completely uncritical acceptance of Wolf’s conclusion. As a curious case of obedience to authority, many bloggers have adopted Wolf’s conclusion without any investigation of the report.

The alert among you will realize that I have taken the opposite tack. I am disputing the authority of Wolf’s conclusions without ever examining the document he is basing his conclusions on. In my defense, the acutal documents are not widely available. I have examined some of drafts and oral presentations of the data. My initial response, beyond wanting to disobey Wolf’s authority of course, was that it smelled fishy. “Disobey authority,” doesn’t sound like the conclusion of a group of civil engineers preparing a report for the National Institute of Science and Technology. I am eager to be proven wrong.

In any case, I would be interested in any evidence or methodology that could cast a light on the general idea that obeying or disobeying authority has a positive correlation with life expectancy.

2 Comment(s)

  1. Gary Wolf | Jun 7, 2005 | Reply

    Thanks, Mr. Baxter, for the intelligent comments. I think that you are absolutely right to question “disobey authority” as a simple rule. My small story about the NIST report was part of a much larger project of looking at public warning and emergency response, and when I have a chance to write about it at length I will try and correct my oversimplification.

    But let me defend the substance of my story. I think you have probably read the study by now, and seen that the numbers I cite are from the report itself. The two types of “disobedience” I was referring to were the orders to “shelter in place” and the long-standing rule to avoid the elevators and take the stairs in case of disaster. The NIST report attempts to estimate the number of people who would have died if:

    1. Nobody from the second building to be hit had begun to evacuate in the time between the first attack and the second attack.

    2. Nobody had taken the elevator.

    This is just an estimate of course, but I found the argument in the report plausible. The point I meant to make was not that you should always disobey authority, but that, in an emergency, authoritative commands have to be weighed against other evidence.

    There is a flip side to this that applies to emergency planning. Planners who intend to warn the public of some danger and give them notice of what to do ought to expect people to seek additional information before complying. This has all kinds of practical implications. For instance, it means that emergency messages should not sacrifice too much complexity for the sake of brevity. When there is a disaster, the public enters an “information seeking” mode. Good disaster response means taking this seriously, giving them sources to confirm or extend the warnings, and expecting them to distrust authority. In fact, the cutting-edge of disaster response attempts to integrate “disobedience” into its model – even taking it as a good thing.

  2. R Cozakos | Jan 9, 2006 | Reply

    Uhhhh, do you have a large heat sink on top of your head? Do you always think this much? I hope you occassionally drop back a few Ghz and just enjoy knowing your children. I know reading your stuff has cranked my brain up just trying to imagine how all this analysis makes your life better (like I know the working out does.)

    Please understand, no offense intended. I just stumbled into your blog from the one Aaron Margosis has on Windows XP (The easiest way to run as non-admin). Maybe I’m just intimidated by your intensity and openness? And never having thought much myself about the names of Jelly(fish) or melting bicycle tires …

    BTW, thanks for the scoop on Automated Backups on Windows XP Pro (which is what got me here). Might save me a few stubbed toes wandering through the same paths.

    I had started to wonder about the tradeoffs of NOT having passwords on a stand-alone XP system at home when I kept running into things I could not do that would also be good for an intruder to not be able to do when they eventually get past my firewall through some dumb/ignorant move on my part.

    I had never intended to become this involved in the details of XP administration. I’m just trying to educate myself enough about security to keep the vandals from my Quicken data without having to keep it on a separate system with an air gap, now that I’m finally hooked to the web at home.

    I would love to read your thoughts on system restore points — keeping them on to be able to restore an earlier version of the registry after getting cooties, vs cooties using the restore points to hide out? (Hope that name is close enough for your sensitivities.)

    Hmmm. Microsoft and Dell security apps warn me about having them turned off, and Trend Micro’s PC-cillin (or was it Spy Sweeper?) tells me to turn them off when getting rid of bad stuff I have come to believe was a false positive. Can I borrow your heat sink? I thought I was having fun back when I learned Lotus 123 1.0 running on DOS 1.0, but THIS is way beyond that, huh?

    It seems like the more capable my computers become, the less time I have to take advantage of it because of the ever-increasing overhead of maintenance. And I once thought it was a lot of work backing up a 20 MB hard disk to 1.44 MB floppies once a week.

1 Trackback(s)

  1. Jun 10, 2005: AETHER

Sorry, comments for this entry are closed at this time.