The Second Glass

Introduction

It is sometimes necessary for individuals to make personal sacrifices for the greater good, to put their minds and bodies in the way of danger for the sake of pure altruism. Think of firefighters, soldiers, proctologists, and now… friends and staff of The Second Glass. On the evening of March 9, 2007, nine able-bodied and wholesome 20-somethings willingly subjected themselves to unknown hazards, and they did it in the name of science.

“Try to get reasonably tipsy by the final tasting round,” we suggested. How every single one of our guinea pigs instead got totally hammered is something of a mystery, but what we do know for sure is that (editor in chief) Tyler is not going to get his security deposit back when he eventually moves out of his wine-soaked and glass-encrusted apartment. Though, on the upside, the carpet in his living room does look like a Jackson Pollack.

Destruction and chaos aside, we set out to answer the age-old question, “Do drinks taste better as you get progressively more drunk?” Everyone knows that the Beer-Goggles Effect is real; it was even verified in a recent and legitimate scientific study. But what about the enjoyment of wine, is the perception of flavor affected by the amount of alcohol in your blood? Our intuition and experience told us, “yes”, but science requires that pesky thing called data. Thus, our experiment was born.

Materials and Methods

We told our test subjects that they would taste a large variety of wines over many separate tasting rounds, but we lied to them to prevent bias and expectation. In fact, they tasted the same set of (randomly ordered) four wines over and over:

1. Taurus, Toro, Spain 2003 - $6.99
2. San Fereolo, Dolcetto, Italy - $11.99
3. Marques de Riscal, Rioja Reserva, Spain 2002 - $19.99
4. Rombauer, Zinfandel, California 2004 - $29.99

We also told them that they were responsible for “kicking it up a notch” after each round, and provided Black Box Wines Monterey County Chardonnay 2005 and Pasa Robles Cabernet Sauvignon 2004 for just such a purpose. Data was collected continuously throughout the experiment, with participants repeatedly recording answers to the following questions:

1. How drunk are you?
2. How much do you like this wine?
3. How much do you think this wine costs?

Results and Discussion

We ran some basic algorithms on the data, “crunching” it, if you will. See the sidebar entitled “Statistical Methods & Your Drunk Ass” for a detailed explanation, but for the mathematically averse we’ll just go ahead and summarize. The principle finding was this- no significant correlation between drunkenness and enjoyment of flavor whatsoever! That is to say, in general, as people got drunker they did NOT rate wines higher (or lower). Indeed, this shocking revelation throws into question nearly everything we’ve come to believe about alcohol.

Another major finding was a very significant positive correlation between enjoyment and estimated cost. Our test subjects apparently think that if a wine is good then it probably costs more. While somewhat expected, this result is itself a raison d’etre for The Second Glass magazine. Wines do not have to cost more to taste great! In fact, while our test subjects’ brains were busy with price prejudice, their palates were sorting out a different story altogether: the most expensive wine got the worst score, and the least expensive wine got the best score.

Actually, the most expensive wine, the Zin, was said by many to taste like Manischewitz. We tasted it and disagreed, but some of us have the taste of Manischewitz permanently etched onto our palates from ten childhood years of drinking nothing else. That, and we were probably about a bottle behind the others.

Maybe we were two bottles behind test subject # 5, who wouldn’t shut up about how he knew what we were doing, how the experiment was so transparent; he flagrantly violated our repeated instruction to refrain from all experiment-related comments until after it was over. So, in the next round we switched one of his samples to a mixture of half wine and half port, and with the glass in hand he declared, “This was sample C from last round. Or maybe it was sample B, but I definitely had this before!” Ahh, sweet revenge.

To complete our insidious prank, we next gave him a glass of port with more than a few splashes of balsamic vinegar… and he didn’t notice anything out of the ordinary! He even gave it a higher rating than other, “real” wines he tasted that evening. Don’t even consider for a moment the possibility that we miraculously happened upon something reasonable- we sampled the concoction, it was most disgusting. Though, it’s probably worth mentioning that the next day he had no memory of this incident whatsoever, despite our humiliating announcement of its details at the end of the night. Call it “Gonzo Science”, inspired by Hunter S. Thompson.

So there you have it- a barely scientific study that wouldn’t stand up for ten seconds under the scrutiny of peer-review. Perhaps the most interesting result of all was that our friends needed only the slightest encouragement from a “scientific authority” to start acting like freshmen at a frat party.
Statistical Methods & Your Drunk Ass

Ari Friedland

For you mathy types, I present here the calculated coefficient of correlation (R) for each pair of variables. This coefficient can have values anywhere between –1 and 1, with the extremes signifying a perfectly (negative for –1, positive for 1) linear correlation between the two variables. Lesser values signify weaker correlation, and R = 0 signifies no correlation whatsoever. As you can see, most of the values in the top two rows range between –0.15 and 0.15, indicating essentially no correlation between drunkenness and estimated cost or enjoyment rating. Estimated cost and rating, on the other hand, are apparently well correlated, with R values in the 0.45 to 0.65 range.