What proportion is "a huge percentage"?

« previous post | next post »

Consider this passage from Kate Lahey, "DJs case will be a watershed", The Age, 8/4/2010, sent in by an alert reader:

Margaret Thornton, a law professor at the Australian National University specialises in discrimination law and policy.

She says the high profile of this {sexual harassment] case would undoubtedly encourage other women to speak out. […]

The $37 million message is also likely to resonate with employers, Thornton says.

"A huge percentage of women, if not most, have actually been subjected to some form of sexual harassment in the workplace … this is going to have an enormous ramification."

It's clear that Prof. Thornton means for "most" to be higher on the scale of proportions than "a huge percentage" is. But I'm also pretty sure that she intends "most" to mean "more than half", and thus she considers "a huge percentage" to be something like 40%-ish, at least in this case. As a result the whole phrase "a huge percentage of women, if not most" seems to mean something like "around 40% and maybe even a majority".

We expect that in "a huge percentage", huge will be (as usual) as interpreted relative to expectations — a huge cockroach is a lot smaller than a huge dog, which in turn is smaller than a huge bear, which is smaller than a huge mountain, which is… So there ought to be examples where "a huge percentage" is cashed out as pretty much any actual proportion, from 1% to 99%, depending on the proportions that we expect to find in a particular case.

In first few examples I was able to turn up, "a huge percentage" again turns out to mean "somewhat less than half":

HUGE percentage of Prehealth kids at Emory.
I found out today, that the amount of kids taking the intro science classes at Emory is around 560 kids (Intro to Bio has 560 kids, intro to chem has 500 kids). That means that the amount of Pre-Health ( pre-med, pre-dental, pre-pharmacy) is around 43 percent (our freshman class size is around 1299 kids). The vast majority of those kids, 90 percent of the 43 percent, are pre-med. That means the amount of pre-meds at Emory is around 39 percent. I think that percentage is far greater than the amount of pre-health kids at other top schools who are known for "Pre-Health schools" (ie JHU).

First Time Buyers Making Up Huge Percentage of Purchases
Nearly half the homes sold in March – 48.0 percent – were purchased by first-time buyers, according to a survey I read yesterday.

But then we get these, where a "huge percentage" is 80%-90%:

Huge Percentage of Students Fail Math Test
LAS VEGAS, NV. – An overwhelming number of students failed their end of semester exams in the Clark County School District. Across all schools, 90.5% failed Algebra 1, 86.6% failed Algebra 2 and 87.8% failed Geometry.

The toys are back in town: a huge percentage of toys sold in the UK are imported from China […]
… between 80 and 90 per cent of our toys [are] brought over from China …

And here's one where the "huge percentage" is 24%:

Once home, AJ began to do really well.  He had lost about 9 lbs in the hospital (he went from 38 lbs down to 29 lbs – so it was a huge percentage of his body weight).

For a child to lose a quarter of his body weight is indeed "huge". And here's a case where 10% — the alleged rate of bad reactions to a drug — is viewed as "a HUGE percentage":

I'm an advance-practice nurse who just got a colonoscopy..I made sure that the endo understood "no midazolam"…I specified this on the consent and they pushed it anyway…I had a very painful experience (even with the usual fentanyl), got very agitated, aggressive (I'm a very calm person), screamed my head off, had to be restrained, have the bruises to prove it and bit one of the nurses. Not really funny, as I have nightmares now about this terrible drug… […] I'm no wuss-I spent 2 tours in Iraq, but midazolam was the worst experience of my life. Why is this awful drug still on the market? 90% love it, but 10% is a HUGE percentage of patients who are traumatized by it.

I have no idea what the facts are, but if 10% of patients have that sort of reaction to a commonly-used anesthetic, that's indeed a "huge" proportion.

How low can "huge" go?  I imagine that if a commonly-used over-the-counter headache drug turned out to kill 1% of those who use it, that would be considered a "huge" proportion relative to our expectations for such things.



19 Comments

  1. Michele said,

    August 4, 2010 @ 8:45 am

    "A huge percentage" of political and marketing press releases try to imply significant scale by citing "a huge percentage." To me it's a red flag that announces, hello, you are about to be manipulated.

  2. Ellen said,

    August 4, 2010 @ 9:27 am

    Interesting. My first thought was, no such thing as a huge percentage, because that would have to be more than 100%. From there I thought, well maybe it's something like, 98%, that range. Thus, even with being someone who sees "most" as meaning well over 50%, I was surprised to find it used as less than most.

    I'm guessing it means a huge number, expressed as a percent.

    [(myl) People seem to use "a huge percentage" to mean something like "a proportion that is large relative to the expected range of proportions for the sort of thing under discussion". This is exactly what huge means in uses like "huge cockroach" or "huge bear", so I think it's what we should expect.

    But I agree with you that there's a natural tendency for readers to react to "huge percentage" as relative to the possible range of percentages, i.e. 0% to 100%, and to take it as denoting the very top end of that range. Which is rarely what it's meant to mean.]

  3. Theodore said,

    August 4, 2010 @ 10:00 am

    So it appears that in the context of percentages in news items, when we say "huge", we mean "much more than the minimum threshold of notability".
    Half of toys sold in the UK being made in China is not even notable; we all know lots of toys are made there, and unless there's lead or cadmium in them who cares? You need 80% to be "huge".
    One out of 20 colonoscopy patients having serious side effects is notable (just about everyone will get a colonoscopy eventually, right?), so it only takes one of 10 to be "huge".

    [(myl) Yes, exactly.]

  4. Tom D said,

    August 4, 2010 @ 10:14 am

    My feeling here is that "a huge percentage" is just an intensified version of "a large percentage" in this context. It isn't necessarily above 50%, but it's enough that it is, as Theodore said, "above the threshold of notability."

    That, of course, also leads to the question, "What's a large percentage, exactly?"

  5. Mr Fnortner said,

    August 4, 2010 @ 10:19 am

    I'm equally concerned, if not more so, by the phrase this is going to have an enormous ramification. Aside from the subliminal naughtiness of the phrase, is there going to be only one ramification? Only one consequence stemming from the case? And how big must this outcome be to be considered enormous?

  6. Matt B said,

    August 4, 2010 @ 10:41 am

    Chris Barker, 2002 ("The dynamics of vagueness") (http://www.springerlink.com/content/4btjtj4h4xcgn461/) points out that words like "huge" have both "descriptive" and "metalinguistic" uses. My understanding of his terminology is that descriptive uses draw upon our present expectations to convey something about the properties of what's being described. So if we already know the threshold of notability or significance, we learn something about whatever is being described as "huge". A metalinguistic use tells us what should count as significant, assuming we already know the relevant properties of what's described. My guess is that most (more than half?) of the cases fall somewhere in between the two uses.

    I think that what's great about the above examples is that they are so clearly instructing us about what our expectations should be. We're told the percentage and then informed that we should think of it as "huge".

  7. richard said,

    August 4, 2010 @ 10:56 am

    I had a boss in a computer company who would occasionally describe something as a "huge problem," which invariably meant that it was his problem, not mine. So in that case "huge problem" meant "a problem I have to deal with," while "minor problem" meant "a problem someone else has to deal with." I have since heard this same practice (sometimes with different vocabulary) right here in academia, so I don't think my old boss was an isolated case–a case of some kind, certainly, but not an isolated one. I guess this speaks to the relative nature of these words, but in terms of who may be affected–if the journalist (in the initial example) wants to imply that you, the reader, might be affected, then the problem is huge; otherwise, it's minor. So, in a sense, in addition to trading on innumeracy among the readership, the practice also draws on the difficulties of risk assessment.

    It's obviously a huge problem–huge, I tell you!

  8. stephen said,

    August 4, 2010 @ 11:35 am

    It's possible for some quantities to be more than 100%. But I don't recall anybody using "a huge percentage" to mean greater than 100%. Has anybody else seen that?

    People might talk about a stock price going up 200%. They call that a huge gain or a large increase, but not "a huge percentage."

  9. D.O. said,

    August 4, 2010 @ 12:23 pm

    …a huge cockroach is a lot smaller than a huge dog, which in turn is smaller than a huge bear, which is smaller than a huge mountain…

    I would even say a huge cockroach is a lot smaller than a tiny dog, a huge dog in turn is smaller than a small bear, and a huge bear is smaller than an indiscernible mountain.

    [(myl) "A fortiori", as we say in the trade.]

  10. D said,

    August 4, 2010 @ 12:37 pm

    "How low can "huge" go?"

    In the minds of some people there's no tax increase that isn't huge.

  11. Jonathan said,

    August 4, 2010 @ 12:40 pm

    If 40% of women are harassed would it still be accurate to say "Most women are never harassed in the workplace" (with the meaning "most" > 50%) but at the same time "but a huge percentage of them are." Both those things could be true, but they imply opposite conclusions if put as stand alone sentences.

    "Most women are never harassed in the workplace."

    "A huge percentage of women are harassed in the workplace."

    or

    "Most people suffer no major ill effects from this drug." (90%)

    "A huge percentage of patients get serious side effects from this drug." (10%).

  12. Margaret L said,

    August 4, 2010 @ 1:34 pm

    When I taught at Flyover State University, slightly less than 1% of the population of the town passed through my Intro class each year. That's a huge percentage.

  13. davep said,

    August 4, 2010 @ 2:36 pm

    Maybe, "huge percentage" means "much larger percentage than one would expect". Which is what you said (in fewer words)!

  14. Peter Taylor said,

    August 4, 2010 @ 3:07 pm

    If "huge percentage" means "far more than 50%", which would be my take lacking contextual clues otherwise, then "a huge percentage…, if not most" is certainly infelicitous, and I think that's why that sentence made me mentally double-take.

    But even if it's interpreted as "much more than you would expect", that relies on speaker and hearer having similar expectations, which I don't think reasonable in this context. It may be reasonable to expect your hearer to agree on the expected size of a cockroach (although even then, only within limits – the largest cockroaches you see in the street where I live are much smaller than the ones I've seen in the Americas); it's less reasonable in the case of a mountain; but in the case of the number of (presumably, from context, working) women who "have actually been subjected to some form of sexual harassment in the workplace" it's reasonable to expect variation in expectations from about 1% to upwards of 80%.

    So am I supposed to peg it against her expectation (or my expectation of what her expectation is)? That clearly doesn't work either, because she's not presenting new figures which might contradict her expectation and so clearly her expectation is pretty close to the value she's presenting as "huge".

    So the only sensible conclusion I can come to is to agree with Michele: it's marketing speak aiming to manipulate.

  15. Rick S said,

    August 4, 2010 @ 9:37 pm

    I think "huge percentage", in many contexts, notionally means a proportion yielding consequences that result in an elevated (positive or negative) emotional state. The lower limit for "huge" then depends inversely on how high the emotional state could go. For example, a new cancer-fighting drug that reduces deaths from cancer by 1% would save 56,000 lives per year in America alone; that's a huge advance. Likewise, a 1% increase in the risk of containment failure during a nuclear excursion at a generating station upwind from you is a huge cause for alarm. If the consequences are huge, the percentage inherits from them.

  16. Robert V said,

    August 4, 2010 @ 10:04 pm

    I always felt like "a huge percentage" was a hedge phrase used instead of "most" when the speaker believed the number was more than half (possibly much more?) and would have been inclined to use "most" but didn't have the proof to back that claim up if called on it. That reading seems to lead easily to phrases like "a huge percentage, if not most" although it doesn't support a 40%-ish interpretation as you propose. I wouldn't be surprised, however, to find a difference between casual speech and careful writing (which could be a valid distinction given your principal example appears to be a spoken quote while the rest are first hand writings).

  17. Katherine said,

    August 4, 2010 @ 11:22 pm

    It seems it would make much more sense to say "a huge *number* of X" rather than "a huge percentage", as it's the overall effect that is worrying rather than the percentage. At least in a huge percentage of the quoted stories. 40% of women in the workforce in a fair number of countries is a huge number while not being a particularly high percentage.

  18. J. Goard said,

    August 5, 2010 @ 2:37 am

    "Most", like "a huge percentage", also seems dependent upon world knowledge and default expectations, except that the former always requires at least slightly more than 50%.

    Most means >50% if the predicate is normally expected to be rare, but something substantially larger if the predicate is expected. For example, as a straight man, if I were to say

    Most of my friends are gay.

    the figure might very well be 51%, but if I said

    Most of my friends are straight.

    it would sound as if I had only a few token gay friends, i.e. a very high percentage for most. Therefore, I would not assert either of these sentences.

  19. Army1987 said,

    August 5, 2010 @ 7:42 pm

    I too at first am inclined that "huge percentage" means "much more than 50%", then when seeing that that cannot be the intended meaning decide it must mean "much more than one would expect". But on the other hand, I remember reading that such statistics have a veeeery broad definition of "harassment" (they give a long list of very vaguely phrased questions, and any woman answering "yes" to at least one of them is included in the percentage).

RSS feed for comments on this post