Has wine a magnetic personality?
By a panel of wine-loving skeptics doing their bit for Western Australia
Don't worry if you don't understand the more technical parts of this hitherto unpublished report. You will still get the message.
An article in The West Australian (12 August 2006, page 60) caught our attention as an example of American pseudoscience -- too good to be true and therefore potentially good at parting the unwary from their hard-earned money (well, money anyway).
According to the article, Patrick Farrell, a resident of California and described as a Master of Wine, has invented a magnetic device that sits around the neck of a wine bottle and transforms cheap plonk into a fuller, better tasting wine. For example he claims his Bev Wizard, which retails for around $US40, transformed a bottle of young and aggressive Bordeaux from the Medoc into a wine that was obviously softer, fruitier, and altogether more sophisticated.
Naturally, as regular wine drinkers (for medicinal purposes only of course), we were soon planning the conversion of various $8 cleanskin Pinot Noirs into wines which, when treated by miracle magnets, would be immediately mistaken by our oenophile friends for a $60 classic Central Otago. Why stop there, we thought, here is a golden opportunity to upgrade quantities of cask Shiraz to the equivalent of Grange Hermitage while promoting our budget Merlots to the status of quality Pomerols or St-Emilions. (For readers lacking the fine appreciative skills that we and our eastern states colleagues take for granted, this is equivalent to converting diet Coke into regular non-diet Coke.)
Farrell says the magnetic field generated by his Bev Wizard alters the shape of the tannins in the wine. Tannins give a wine much of its distinctive taste and come from the grape skin, grape seeds, or oak if the wine is aged in oak barrels. Any old tannin won't do, as anyone can discover for themselves by dunking a tea bag in their favourite plonk. Farrell claims that any bitter taste due to tannins can be softened by his Bev Wizard because the magnetic field makes the tannin molecules larger and thus more acceptable to demanding palates. Tannins are traditionally softened by bottle ageing and (to avoid sediment) by decanting the wine prior to drinking. But Farrell is adamant that his Bev Wizard fast-tracks the whole process. As far as we know, such a process is completely unknown to science.
We noticed that other wine experts had pitched in with their support for the Bev Wizard. Peter Marks, also a Master of Wine and Director of Wines at the Napa Valley Centre for Wines in California, reported a blind tasting in which many of the tasters said they found a difference between $US10 wines and their magnetized versions. They preferred the magnetized wines and claimed that the Bev Wizard worked particularly well with less-expensive wines. There was no information on how the test was designed, or whether the tasters knew in advance which wine they were tasting (which as we shall see is crucial), or whether at some stage the tasters became unable to distinguish reds from whites.
We can also turn to those who know a thing or two about molecules and magnetism. David Ball, for instance, is a professor of chemistry at Cleveland State University (you can find his credentials at http://academic.csuohio.edu/ball/cv.htm) and he is not convinced. He warns us: "All that the magnetic field is doing is separating you from your money. Magnetic fields aren't strong enough to change the shape of tannins." Professor Ball then speculates: "If there were any improvement in the taste of wine which had passed over the Bev Wizard, it was probably due to oxygenation or imagination."
The role of imagination is by no means a figment of itself. Many studies have shown how easily our perceptions (what we expect) prevail over reality (what we actually get). For example our perceptions of how Coke and Pepsi taste prevail over how they actually taste (no perceptible difference when we can't see the label), see the article "Pepsi versus Coke: labels, not tastes, prevail" in Psychological Reports, 52, 185-186, 1983. If we think magnets work, whether improving our sleep, easing foot pain, or improving wine, our perceptions will probably confirm it despite the evidence against the first two, see the article on magnet therapy at www.quackwatch.org. Which is why tests of new drugs must be performed under double blind conditions to make sure nobody knows whether they are getting the real drug or an imitation one.
Grasping the nettle
So we chose the middle road and selected four red wines similar in colour and price, namely a 2003 South Australian Vintage Merlot, a 2004 South Australian Shiraz-Cabernet, a 2003 Victorian Temperanillo, and a 2005 South Australian Cabernet-Merlot.
Our testing procedure
(1) The four bottles had their labels replaced by labels A to D by a person who, at great personal sacrifice, would not be involved in the tasting.
(2) Each of the four bottles A to D were carefully decanted into a glass jug by a second self-sacrificing person blind to which wine was in which bottle. The small quantity of wine that remained was swirled around and poured off to remove any sediment. The wine in the glass jug was then carefully decanted back into its original bottle.
(3) With collars holding 0, 1, 2, or 3 magnets around the neck of the bottle, the same person filled plastic cups the size of an egg cup with the various wines. Each magnet was a flexible sheet magnet attached to the neck of the bottle like the one apparently used in the Bev Wizard.
The cups were of white semi-opaque plastic that helped to mask any colour difference between wines. The samples were divided into sets (also known as flights by habitual wine tasters, presumably because among bystanders unfamiliar with such rituals they invite flights of stares), each set having six samples of wines that had been poured past 0, 1, 2, or 3 magnets under double-blind conditions.
Four sets of six samples were taken from each bottle so a total of 4 x 4 = 16 sets of six samples were available to our eager tasters. Tasters were told to select any set, and to taste every sample within the set before moving on to a different set. Each of the six samples in each set had to be rated on a scale of 1 to 6 (representing very poor, poor, fair, good, very good, excellent), according to the taster's assessment of appearance, aroma, and taste.
Left: Wine bottle (not yet relabelled) with its original label hidden. On the neck is the holder for 0, 1, 2 or 3 magnets. The bottle is carefully held with the holder side downwards so the wine flows over, not under, the magnets. Middle and right: wine samples laid out in plastic cups. Each taster selects one set of six samples at a time.
Initially 13 skeptics had volunteered their services as tasters, but the regimented array of samples was so intimidating to their unassuming and delicate sensitivities, as were the anti-drink-driving messages being aired on television, that only five unshaken stalwarts actually took to the cups. The rest sought solace in stares, leaving the tasters stared, not shaken. At the end of the evening all results were collected for subsequent analysis by our statistical experts.
Ratings of the six wine samples
The spread of ratings for each sample shows that the judges tended to disagree on its quality, which was not unexpected since none of them pretended to be experts in wine. To see what effect magnets may have, we combined ratings according to the number of magnets. The graph below shows the results.
Effect of magnets on the taste of wine. r = correlation between average rating and number of magnets. sd = standard deviation (a statistical measure of scatter in the individual ratings).
The results show that magnets had no effect. Indeed, the average rating tended to decrease (albeit very slightly) as the number of magnets increased, suggesting if anything that magnets made the taste worse. But thanks to the small sample sizes the scatter attached to each average is so large that we cannot rely on this eyeball conclusion. Instead we must apply what statisticians call a t-test, which tells us whether two averages are significantly different when their scatter is taken into account. In this case the t-test tells us that the difference between wine without magnets (mean rating 2.60 sd 0.95 based on 48 samples) and wine with 1-3 magnets (mean rating 2.56 sd 1.07 also based on 48 samples) is not even weakly significant, p being 0.85. (As every statistician knows, the t does not stand for taste nor p for plonk, but we are working on it.) So what conclusion can be drawn?
In hindsight it could always be argued (not by us of course) that, unlike the products from California, the plonk was too cheap to be redeemable even by magnets. But we leave that one to our eastern states colleagues, who will of course feel duty bound to completely mistrust their earlier findings and our present results, and to run their own new tests on a suitably huge scale. We look forward to their findings.