Inaugural Cosmic Variance Whisky Tasting

By Sean Carroll | May 12, 2010 10:15 am

One of the consequences of having a blog is that people occasionally offer to send you free stuff. Not out of the goodness of their hearts, for the most part; rather, because they’d like some free publicity in return. Usually it’s a book of some sort, and usually I just decline; I can’t possibly get through all the books I hope to read on my own, much less other books that people want me to read.

So when I received an email from Kimberly Moniz at SHIFT Communications asking if I’d be interested in receiving a free sample to possibly mention on the blog, I almost replied automatically in the negative. But then the nature of the product sunk in, and I paused — this wasn’t a book, this was booze! Specifically, Canadian Club whisky.

I’ll admit that I’m an occasional Scotch drinker, but not much of a Canadian whisky fan. To be honest, the mention of Canadian Club conjures images of something my grandmother would have been drinking (while smoking her Pall Malls), although that seems to be changing. I suspect the marketing people recognize that, and thought it would be good to freshen their image among a younger, hipper crowd. And what better way to do that than by reaching out to science blogs? (Especially ones that occasionally rhapsodize about the perfect martini.) This is some new-media marketing savvy I can get behind. Also, free booze.

But our honor is not sold so cheaply — we’re not going to provide free advertising just because someone sends us some loot. We have our scruples, and everything we post must adhere to the guiding principles of our Mission Statement. But then I remembered that our Mission Statement says we post about whatever we feel like posting about. Still, we like to convey at least the illusion of integrity.

So I hit upon the perfect solution: talk about the whisky, but do so through the lens of Science! That is, we would accept the free booze, but only under the understanding that we would subject it to a rigorous taste-test in comparison with other comparable whiskies, apply the time-honored techniques of the experimental method to the results, and publish whatever they might turn out to be. Kim was up for this adventure, so we set the wheels in motion. Results below the fold.


We were sent a bottle of Canadian Club’s super-special 30-year-old whisky, as well as a bottle of their slightly-special 12-year-old offering. I judged that it would make the most sense to put them up against a variety of other kinds of whisky, so we assembled a worthy array of adversaries from around the globe. (Note that the Scots and the Canadians spell it “whisky,” while the Americans and Irish use “whiskey,” so there’s no consensus choice in this case.) From left to right, we have the two Canadian Club entries; a bottle of Jim Beam Distiller’s Series, a Kentucky bourbon; Jameson, a classic Irish whiskey; and Glenlivet, a well-known single-malt scotch. I wasn’t smart enough to keep track of the prices for the different bottles, but you should imagine that the 12-year CC is about $18, the Jim Beam is about $18, the Jameson is about $23, and the Glenlivet is about $35. The 30-year CC is apparently a collector’s item that will set you back about $200.

Then, to the experimental procedure. I rounded up a few friends (8 people in all), none of whom were really whisky experts, but none of whom were really averse to the idea, either. One of them was Risa, who was passing through town for other reasons, so it was a true Cosmic Variance event. We disguised the bottles by putting them all in plain brown bags and assigning numbers to each. All of the testers were given samples of each whisky, as well as a score card where they were asked to give each entry an overall score between 1 and 5, as well as to write down more specific impressions.

And here are the results! Complete with error bars (at least, standard deviations computed from the raw data), so that makes it science, right?

Whisky Data

And the results are — inconclusive. Well, that’s not exactly true. We learned a lot, actually. It’s true that all of the scores ended up being within the error bars of each other. So we didn’t technically learn a lot about whisky. But we learned a lot about the proper experimental protocol for conducting a whisky tasting! Mostly, I don’t think we gave people enough of it. That is, more than one person remarked that it took more than a couple of sips to really start to appreciate the merits of any individual liquor. If we were to do it again, we’d take more time, offer lots of food, and let people really enjoy each of the offerings, so that their individual ratings were more reliable. In the meantime, if you’d like useful insight into the nature of these different drinks, you’d be better off checking out Drinkhacker or similar sites. They actually know what they’re talking about.

The other thing we learned is that, despite the obvious scatter in the numerical results, people did have somewhat consistent reactions to the different whiskies. Both kinds of Canadian Club were judged as sweet, mellow, vanilla; the Jameson was lighter and thinner; the Jim Beam was a bit more harsh; and the Glenlivet was thought to be more complex but softer. It’s just that different people liked different characteristics. So perhaps there is no absolute scale of whisky achievement, but rather a subjective relationship between the liquor and the imbiber? Of course, down that road lies moral relativism and nihilism, and ultimately the Taliban will win, so we should probably avoid that path.

Most of all, we learned that we like getting sent free stuff that doesn’t require hours of our time to read and think about. Other good possibilities for this category might include: iPads, trips to Vegas, and Tesla roadsters. You know where to find us.

Finally, here’s the scientific explanation behind the production of Canadian Club whisky. Thanks to Kim Moniz for sending this, as well as the whisky, and for playing along. And apologies for my usual tardiness — it took a while to put the tasting event together, and then I went and lost the score sheets for a matter of months. But science eventually triumphs!

Production Process of Canadian Club

Canadian Club is made from a blend of corn, rye, rye malt and barley malt. Once the grain arrives at the distillery it is split up: the corn is milled and put in a premix tank where water is added to prepare it for cooking. The other grains are milled together, mixed with water and the product, called mash, is then transferred to the batch cooker. The cooked mash is moved from the batch cookers to conversion tanks where malts are added to change the starch into fermentable sugars. Once the malts are added it produces high sucrose mash, which is then transferred from the conversion tanks to the fermenters. There are 39 fermenters in the distillery, each stands over three stories high and is able to hold up to 218,000 liters. Once the fermenter is filled with the high sucrose mash, two types of yeast are added to the mix –commercial yeast is added to the corn mash, while a specially developed yeast patented in 1858 by Hirman Walker is added to the blender grains mash. After 72 hours of fermentation, the product is 11% alcohol and is referred to as “distiller’s beer”. The fermented mash is pumped into a whisky still where it flows through horizontal filtration plates. Steam is injected at the base of the still and as it rises, it causes the alcohol to evaporate out of the mash. The alcohol is then carried to the top of the still and condensed into liquid form called “new whisky”(this is the process of distillation). The “new whisky” is then run through other stills in order to remove unwanted elements, such as fusel oils. The removal of these items is what makes Canadian Club have a light, smooth flavor. The corn distillate is now 95% alcohol per volume, colorless and odorless. The blended grains are distilled through the column still, but at a lower alcohol strength in order to produce a higher grain character in the taste of the whisky. The grain mash is further refined in a copper still, enhancing the flavor of the whisky. After quality control checks, each whisky is then pumped into large blending tanks prior to aging. This process is referred to as pre-barrel blending (blending the corn and grain mash), a unique process to Canadian Club. The colorless new whisky is then pumped to drain and fill where it is tapped into the once used American oak charred barrels for its aging process.

  • Evan Harper

    If you intend to repeat this trial, I have to suggestions. First, consider tasting both straight whiskies, and significantly watered-down whiskey (1:1 or even more.) The alcohol “bite” can disrupt one’s enjoyment of subtle flavors, and some whiskys just seem to taste better when they are mellowed out. I would assume that’s even more true of the single-malt scotch than the blended Canadian whisky, but never having conducted a blinded trial myself, I can’t say that for sure.

    Second, since you’re already tasting Canadian whiskys, consider including Forty Creek‘s whisky in your batch. I like it, and it’s local owner-operated, rather than part of a giant booze conglomerate.

  • Richard Scalzo

    Hi Sean,

    Have you considered cross-correlating scores from the same participant among different varieties, or with (if they put them up) numerical scores for the different taste characteristics? This might control somewhat for the vagaries of individual taste, assuming that the varieties in the palette sample a reasonable chunk of possible booze experiences.


  • Sean

    It might. But it sounds like work. Maybe someday we’ll do this more carefully. Probably when someone pays me to do it on TV.

  • Eugene

    Not to discourage any future freebies for you and CV, but isn’t the fact that the CC 30 year being an order of magnitude more expensive than the rest, and yet couldn’t muster enough oomph to go rise above the noise of sqrt{8}, tells you that maybe it’s just not worth its price?

  • Non-Believer

    I think the biggest issue here is you didn’t have a large enough group of people tastibg to get a TRUE result. Now if you were to increase the number of participants, to say, well at least 9, which would of course include me, you would get a better more accurate result. I offer out of the kindness of my heart. No other motivation. Just a thought.

  • Martin

    Why did you use a line graph for data in discrete categories? You should have used a bar graph.

    Also, I agree with Eugene. The marginal improvement in taste of top-shelf liquor doesn’t justify the cost in most cases. Usually there’s enough of a difference between bottom-shelf and mid-range liquor to justify the costs, but not between mid-range and top-shelf. I remember watching a TV show where they blindly gave people a variety of vodkas and the test subjects liked the mid-range stuff just as much as the top-shelf stuff.

    People drink expensive liquor for reasons other than taste.

  • Christopher Todd

    Hi Sean,

    Dig this post, but as others have mentioned the methodology could use a tweak or two. Putting the 30 year old up to other similarly aged and priced, and then the 12 y.o. along with its’ kindred would yield more interesting results.

    Another thing: The spelling of the booze isn’t geographically based. Case in point: Maker’s Mark (Kentucky bourbon) says “WHISKY” right on the front, and Bushmills (Irish Whiskey) uses the spelling “WHISKEY” on its’ label.

  • Chris Palmer

    Palmer’s Law of Alcoholic Drinks:

    The third glass of anything (from single malt scotch to MD 20/20) is delicious.

  • Possible pseudonym

    @Christopher Todd: I assure you, all Irish whiskeys are whiskeys, and all Scottish whiskys are whiskys. What you get up to your side of the pond is your own business. (“,)

  • Andy Fleming

    Hi Sean,
    If you’re after European participation in scientific alcoholic beverage analysis, please let me know…. myself and a group of workmates are fully trained in testing and have had plenty of experience!

  • steeleweed

    However you spell it, there are better brands of whisk(e)y to choose from.
    I can’t comment on the Scotch or Canadian, as I don’t care for either type.
    Suggest Murphy’s Irish if you can get it (don’t think they export it to US any more) and Old Charter bourbon ( available West of the Mississippi).

  • onymous

    I remember watching a TV show where they blindly gave people a variety of vodkas and the test subjects liked the mid-range stuff just as much as the top-shelf stuff.

    I’ve seen claims that blind taste tests show that people prefer dirt-cheap vodka run through a Brita filter to expensive vodka, but other claims saying differently. Someone should apply for funding to do a really large-scale study.

  • spyder

    An interesting but way too brief, an experiment. Comparing Canadian whiskeys with Scottish single malts is nearly apples to oranges. Bourbon whiskeys, and rye ones too, are also very different from sour mash to malted grains and so forth. Aging is also a variable that cannot be accounted for in the same way that price fluctuates. Start with Johnnie Walker labels: Red, Black, Gold, Green, and Blue. They represent various blends and aging processes from relatively inexpensive to staggeringly luxurious. That would make a great test batch. Then do the same with single-malts from isles and highlands; albeit the starting prices are usually much higher than the others. This needs to be a long-term ongoing study. hehehe

  • Patrick Kelso

    I thnk a longitudinal study is needed to ensure that the same level of quality is evident in each brand.

    If you want a decent sample I can recommend a good whisky bar in Roppongi, Tokyo; over 800 different Whiskys available. I recall that the Blackadder Whisky was particularly nice.

  • Ted Dunning

    I hate to be too serious, but there is a nice statistical (Bayesian even!) way to analyze this that let’s you untangle the offsets due to individual cussedness and due to inherent whiskiness.

    Send me your data and I will bump out the analysis for you.

  • Sili

    My late Computer Science lecturer used whisky for the final big programming exercise on his course from time to time – though not the year I finally bothered to do the homework. I think it dealt with having to divide whiskies into groups based on a list of yes/no characteristics.

  • marcel

    Don’t the quantities tasted have to be much greater for this to be a blind test?

  • layman

    How can you pretend this is Science! C’mon read Popper, Kuhn… THIS is Science, and they all agree: it takes at least 9 subjects. Well if I can help…

  • Fairchild

    Wait, let me get this straight. This company sends you a $200 bottle of whisk(e)y just because you have a blog? When I set up my blog, how long does it take to get these free gifts? You can respond at my new blog,… :)

  • JoAnne

    Yes, please do send us Tesla roadsters! I’d be happy to perform rigorous scientific tests on their capabilities.

  • Malc

    European physics whisky tasting clubs have more refined science goals and higher luminosity.

  • Carl Brannen

    Palmer’s law must work the opposite on me. I never have a second shot glass cause the first kills my taste buds. And I also vote for Glenlivet. I just finished off a bottle that was sold to me as a 12-year-old, but finished off as a 21-year-old. Another favorite is Balvenie.

    For years I’ve had a hankering for a vodka (Polish buffalo grass) I haven’t tasted since 1984. I finally looked it up and discovered that soon after then it was declared illegal in the US. But whenever I really feel like a glass of it, it’s too late to drive up to Canada and illegally import one.

  • JMW

    “…the Jim Beam was a bit more harsh…”

    I’ve heard that the only reason people drink Bourbon is to demonstrate that they can.

  • George

    So the title should “A bunch of people tasting whiskey”?

  • Dionigi

    @Carl Brannen
    I’m pretty sure that I have read that whisky does not age when in the bottle, it has to be in a cask so that the angels can get their share to alter the taste.So a 12 year old would remain a 12 year old. BTW how do you keep a bottle of scotch 9 years? mine dont last the year.

  • Sean Peters

    I’m not really sure it makes much sense to compare completely different products in this way. It seems most likely that what you’re actually measuring is the relative popularity of the different kinds of spirit, rather than anything about the spirit itself. I’m a bourbon drinker, and I’m naturally going to prefer the Jim Beam to the other spirits. That doesn’t necessarily mean that Beam is higher quality than, say, CC.

  • Sili

    European physics whisky tasting clubs have more refined science goals and higher luminosity.

    They also get more intimate by obeying Bose-Einstein statistics.

  • Niall

    I must be a pedant with this and emphasise you drank whisky (Scotch) and whiskey (Irish), or whisk(e)y (mixture).
    Plurals thereof are whiskies (Scotch) and whiskeys (Irish). I’m not sure what the plural of whisk(e)y is… I’d go for whisk(e)y(s)ies.

    Also. I’d recommend doing different whisk(e)y(s)ies over different days, or at least with a large gap between drinks. After a few you couldn’t detail the finer points of a good one.

  • Mike

    Your plotted error bars should be the standard deviation of the mean, not the standard deviation. If you do so, I suspect you will find a marginally significant result.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

About Sean Carroll

Sean Carroll is a Senior Research Associate in the Department of Physics at the California Institute of Technology. His research interests include theoretical aspects of cosmology, field theory, and gravitation. His most recent book is The Particle at the End of the Universe, about the Large Hadron Collider and the search for the Higgs boson. Here are some of his favorite blog posts, home page, and email: carroll [at] .


See More

Collapse bottom bar