Commentary on Research Paper: Do More Expensive Wines Taste Better?
I’ve always thought that wine is a fine beverage, but I’ve never had the desire to learn more than a few basic facts. I was satisfied just knowing that the type of wine is determined by either the variety of grape or by the region where the grapes are grown. I was proud enough of the fact that I could generally tell the difference between a Pinot Noir and a Cabernet Sauvignon, but I was generally stymied by the differences between two Pinot Noirs. This is probably how most people feel, and therefore it seems ridiculous to spend $50 on a bottle of Pinot that tastes similar to a $10 bottle of Pinot.
It’s pretty obvious when you walk into a wine shop that the wine suppliers are trying desperately to establish a brand. Half of the names include outlandish words like “Bitch” and “Devil” and have pitchforks and flames all over the bottle. It’s an obvious attention grab that leads the consumer to believe that having a clever name is more important than having a high-quality wine. My thought is generally to pass on all of these products because it is a clear attempt to increase sales through marketing rather than through improving the quality of the product.
What criteria do I use to buy a bottle of wine, you ask? Without having any real knowledge about wine, I judge a wine based on rating (if available), price, and origin, in that order. You’ll argue that the label actually affects my decision, but that’s a different discussion for another article.
I’ll get to price and rating later; let’s take care of country of origin first. In general, I will drift toward a foreign wine if price and rating are equal. My thinking is that some of these countries have been making wine for centuries, so maybe they’ve got it figured out. This is a total stereotype, and I could very well be buying wine from a just-opened winery in a foreign country. Argument destroyed. My limited knowledge is that French wines are good. German white wines are good. Southeast Australia and New Zealand produce good Sauvignon Blancs. Argentina has good Malbecs. Northern California wines are good. Italy has Chianti, which is made legitimate by the presence of an ink rooster stamp on the paper loop by the cork. That’s about it.
Another problem is the mystery of the winery itself. As I already mentioned, I have no idea whether a winery has been in business for a long time, which tells me that they have a good product and have stayed around. Furthermore, marketers intentionally make this very confusing. One winery in California called Stag’s Leap Wine Cellars, per Wikipedia, “became famous when its 1973 Cabernet Sauvignon won first place in a blind taste test of reds at the 1976 Judgment of Paris.” There is a different brand called Stags’ Leap Winery, which is basically just riding the brand of Stag’s Leap Wine Cellars. This type of thing causes me to simply throw up my hands and refuse to remember specific brands. Which leads me to my second and third criteria, rating and price.
I basically buy wine based on price and a rating provided by some third-party service like Wine Spectator. I’m ideally looking for someone to tell me that a bottle is of high quality and is also a good price.
But do expert ratings and price matter? This question was answered in a 2008 research paper by Goldstein et al titled “Do More Expensive Wines Taste Better? Evidence From a Large Sample of Blind Tastings,” published in the Journal of Wine Economics. The authors analyzed data obtained by the first author, Robin Goldstein.
Goldstein separated blind tasters into two groups, which he called “experts” and “non-experts.” I am clearly a “non-expert,” so I’m going to focus on that. The authors note that they suspect the “non-experts” to be overrepresented by “highly educated individuals and individuals working in the food and wine industries.” Which actually means the paper is conservative in their prediction of just how little the “non-expert” actually knows.
The tasters were asked to rate wines on a scale from 1 to 4, corresponding to the descriptions “Bad,” “Okay,” “Good,” and “Great.” Goldstein took over 6,000 data samples, and the authors of the paper crunched the numbers to get four regressions. I wanted to grab the data to try to set up the regressions, but the data weren’t readily available1.
One conclusion of the study is that the average, “non-expert” wine drinker actually gives a higher rating to cheaper wines. Another interesting conclusion of the paper is that expert ratings are next to useless for the average “non-expert” wine drinker. The authors’ justification for this conclusion is that although “expert” wine tasters are able to distinguish between different vintages of the same wine, they aren’t able to actually tell which is the better of the two vintages. This is humorous because you hear in movies the outlandish statements like, “OOOOH. The 1973! What a great vintage. I’d rather vomit than drink the 1974, but the 1973 is FANTASTIC!” This study says that the expert tasters identified the “better” vintage only 50% of the time. That is literally a coin flip guess.
So here we have it: a smoking gun that tells me I should basically buy the cheapest wine available. The paper essentially proves that the average wine drinker should rely on neither a high price nor a high rating to select wine. That leaves the country of origin and the label (basically a shot in the dark). My recommendation to you is to simply pick a bunch of bottles of cheap wine ($6 – $12) and try to find something you like. Goldstein et al cut off cheap wines at $6 per bottle, so if you are following the paper you should spend about $6.
I will be writing an article in the future where I will buy a number of bottles of a single wine variety (probably Malbec, Pinot, or Chianti) and perform a blind taste test to select my preferred wine. I’ll stick with this one going forward.
1. I tried to recreate the equations outlined by the paper in Excel but the authors do not provide all of the information required, particularly the epsilon value. Taking the authors’ description as a linear function with a y-intercept of 2.297 and a slope of -0.038 and using price ranging from $6 to $150, one gets results ranging from a rating of 2.23 to 2.11. This is a very tight result that corresponds to a description of “Good” for the full extent of the range. I interpret this as all of the “non-experts” saying that all of the wines are “Good.” There is probably some scatter where people say “Great” and “Okay” occasionally, but pretty much everyone is saying everything is “Good.” This is bogus in my opinion and calls into question the paper. These qualifying terms could be a real problem. Maybe a better strategy would have been to rate the wine from 1-10. Then you could get a spread of values rather than simply “Good.” Furthermore, it credits the “experts” with the ability to make finer distinctions than simply saying everything is “Good.” This totally makes sense.