04 Jun 2014

Check Your Hurricane Privilege

All Posts 18 Comments

A lot of people were having fun with this study:

People don’t take hurricanes as seriously if they have a feminine name and the consequences are deadly, finds a new groundbreaking study.
Female-named storms have historically killed more because people neither consider them as risky nor take the same precautions, the study published in the Proceedings of the National Academy of Sciences concludes.

Researchers at the University of Illinois and Arizona State University examined six decades of hurricane death rates according to gender, spanning 1950 and 2012. Of the 47 most damaging hurricanes, the female-named hurricanes produced an average of 45 deaths compared to 23 deaths in male-named storms, or almost double the number of fatalities. (The study excluded Katrina and Audrey, outlier storms that would skew the model).

The difference in death rates between genders was even more pronounced when comparing strongly masculine names versus strongly feminine ones.

“[Our] model suggests that changing a severe hurricane’s name from Charley … to Eloise … could nearly triple its death toll,” the study says.

I didn’t read the study itself, so I’m hoping they accounted for this, but in any event: It occurs to me that the “sexism” could go the other way: When male researchers see a force of nature that is going to destroy everything in its path for no good reason…they name it after a woman.

See? That’s plenty sexist and it goes the other way!

18 Responses to “Check Your Hurricane Privilege”

  1. guest says:

    Ba-dum bum. 😀

  2. Harold says:

    I guess this is a joke, but unfounded in reality. Each tropical storm is named with alternate male and female names form a set of 6 lists in sequence whether it develops into a hurricane or not. Every year the first name alternates between male and female. Dangerous storms are effectively randomly assigned male or female names.

    This is an interesting study, and if correct reveals a certain lack of rationality. There is no rational reason why someone would risk death so much more because of a random factor like the gender of the name of the storm.

    The wisdom of using human names has been questioned: “this practice also taps into well-developed and widely held gender stereotypes, with unanticipated and potentially deadly consequences”

    Despite the humorous tone of Bob’s comment, this study apparently reveals by peoples actions – not by surveys and questionnaires – that these deeply held gender stereotypes are real.

    • Tel says:

      Going by Wikipedia (Atlantic):

      Names retired in the 1950’s: Carol, Edna, Hazel, Connie, Diane, Ione, Janet, Audrey, Gracie — 100% feminine.

      Names retired in the 1960’s: Donna, Carla, Hattie, Flora, Cleo, Dora, Hilda, Betsy, Inez, Beulah, Camille — 100% feminine.

      Names retired in the 1970’s: Celia, Agnes, Carmen, Fifi, Eloise, Anita, Greta, David, Frederic — 78% feminine.

      If that’s random, then I was the gunman behind the grassy knoll.

      • Tel says:

        It wasn’t until 1979 that hurricane names started to alternate between female names and male names. Considering there are so many more hurricanes with girl names than boy names, wouldn’t this skew the data set?

        The researchers say it doesn’t. Their study looks not just at whether a hurricane has a female name or a male name, but at how feminine or masculine the name is. For example, the name Bertha sounds less feminine than Laura.

        Ha ha ha… how stupid is that?

        I mean, it’s not like technology has done anything whatsoever since 1979. Just as an example, our budget allocation to nutbag research is far bigger now than it should be.

    • Bob Murphy says:


      I didn’t read the study, and I didn’t know (assuming you’re right) that the names of hurricanes are pre-ordained.

      I guess a more serious formulation would be: They didn’t merely look at the damage by sex, right? They did a full-blown regression holding other things constant? I’m assuming they did, but in the journalist summary it didn’t jump out that they did (which of course means nothing).

      • Harold says:

        Coupled with Tel’s post above, this needs more looking into. The naming I described is currently correct, but as Tel pointed out it was not always thus.

        USA started giving hurricanes women’s names in 1953, and only switched to male and female names in 1978, so that is a lot of female named storms. The authors say “We use more than six decades of death rates from US hurricanes to show that feminine-named hurricanes cause significantly more deaths than do masculine-named hurricanes.” The first 25 of those 60 years only had female names. Whatever analysis they did would have to compensate for this. Apparently “The pattern held true even if they concentrated only on storms that hit after the naming convention changed. The authors were careful to factor in how risk of storm damage increases as population grows over time.” So that should take care of that concern.

        People rated nameless hurricanes as slightly less dangerous than female named ones. The authors say that if action is required it should be to make people perceive storms as more dangerous rather than less dangerous. This would mean fewer lives lost as a direct result of the storms, but this may not be the optimum outcome. Maybe people actually over-react, and overall better outcomes would result from downplaying the dangerousness, even at the cost of a few more storm victims.

        Whichever is the case, people seem to be reacting differently given the name only, a reaction which is irrational.

  3. Josiah says:

    I did read the study. They regressed for normalized damages, death toll, and minimum pressure (and the “gender” of the hurricane). There did not appear to be any attempt to account for the fact that hurricanes all had female names prior to 1979.

    The experimental portion of the study was interesting. It makes me wonder what you would find if you did a similar study based on the perceived racial identity of storm names.

    • Bob Murphy says:

      I did read the study. They regressed for normalized damages, death toll, and minimum pressure (and the “gender” of the hurricane).

      I don’t mean to be stupid, but how does that make sense? Weren’t they trying to, say, explain death toll and/or damages as a function of the objective strength of the hurricane (top wind speed, population density of landfall, etc) and sex of the name?

    • Matt M -Dude Where's My Freedom) says:

      I’d run for my life from Hurricane Shaniqua.

  4. Andrew' says:

    But the male hurricanes are assertive, while the female ones…

  5. Keshav Srinivasan says:

    Here’s the paper if anyone’s interested:

    And here is a reply to critics from the authors:

  6. Ben B says:

    I for one will never take a female storm lightly again. I’ve seen the Discovery Channel’s “Deadly Women”.

  7. Yancey Ward says:

    Well, they need to start using male names exclusively as a matter of public policy, and make the names as masculine as possible. Would you ignore Hurricane Brutus, or would you move inland to Manitoba?

  8. Philemon says:

    William M. Briggs site had a fun take down of this study with good comments as well:


    Of course, I am prejudiced in the completely reverse direction from the Illinois college students surveyed because I lived in places at risk from hurricanes back when they all had female names and did more damage to life and property due to a variety of factors. Therefore, I tend to view the male-named storms as politically correct sops until I see evidence otherwise whereas female-named hurricanes evoke a more visceral fear reaction.

    • Harold says:

      Briggs misses the point. he says “Just as do the theories, “Hurricanes starting with the letter P kill more than the letter R” or “Hurricanes with three syllables kill more than with one or two.” These have low power, too. Just as will any of the infinite theories you can think up which would explain the data.” These hypotheses are without any evidence whereas the one in the paper is supported by evidence. The number of deaths DID correlate with femininity of the name to a statistical significance. There would probably be no significant correlation between storms beginning with letter P or those with 3 syllables and number of deaths.

      A correlation does not mean a causal connection of course. Causal connections are more likely if you have a hypothesis for the connection which you can test using other means. This is what the authors do. They construct the hypothesis that people are less scared of female named hurricanes, and therefore less likely to take evasive action. They then test this hypothesis using students and find it is true – they are less scared of more feminine named hurricanes. It is unlikely that this supporting evidence would be obtained if you tested the two-vs-three syllable name hypothesis.

      This provides a coherent, evidence based hypothesis to explain the correlation.

      It is not conclusive, of course. The data is sparse, the correlation not that strong (it is not statistically significant if you look at 35 years instead of 60 years) and there are other possible hypotheses to explain it. Nonetheless, it is a very far cry from “any of the infinite hypotheses you can come up with.”

  9. Philemon says:

    Well, Harold, you go to your church and I’ll go to mine. Illinois college students who were never near experiencing a hurricane ever in their short little lives are not the best judges of hurricane strength, no matter the name.

    “The number of deaths DID correlate with femininity of the name to a statistical significance.” Only if they threw out outliers.

    As I mentioned, the comments were good:

    Nir Friedman
    4 JUNE 2014 AT 8:37 AM
    I really think the most important point is completely being missed. You can’t just throw away outliers because they seem large relative to your completely empirical estimate of the distribution.
In fact, what Katrina and other large storms tell us, is that the distribution for deaths from hurricanes is reasonably fat tailed. So using negative binomial, or Poisson, as models, are both completely unreasonable as they are thin tailed.
In fat tailed distributions, fluctuations are much more significant, so the sample mean will vary much more. Hence the results are not statistically significant.

Leave a Reply