Monday, January 29, 2018

What do wine glasses and hamburgers have in common?

Short answer: their sizes have increased a lot during the past 30 years.

This is of importance because obesity is now considered to be an epidemic in many countries, particularly the USA. Indeed, for more than 20 years obesity has been seen as "the most challenging public health problem that we have faced" (Ref. 1). One of the things that is obviously associated with obesity is increasing consumption of energy-rich food; and one of the things implicated in increased consumption is the size of the portions being eaten — increased portions lead to increased energy intake, and if the energy is not immediately used then the person will put on weight.

Obesity in the USA, from the Obesity Rates & Trends Overview web site

The relationship between obesity and the size of portions is a complex issue, which I will not go into here. However, for those of you who are interested, there are a number of recent literature reviews (Refs 2, 3, 4, 5, 6, 7). The experimental consensus is that people consistently consume more food and drink when offered larger-sized portions, packages or tableware than when offered smaller-sized versions.

What is of particular interest for this blog post is that health professionals have been collecting data about increasing portion sizes over the years, and I thought that we might look at some of the data, and their consequences. Portion size is usually reported to be largest in the USA, and steadily increasing [Refs 8, 9], although data from Europe show similar trends — eg. Denmark (Ref. 10), the Netherlands (Ref. 11).

Increased portion sizes

People have now realized that there are two ways that increased portions have been happening, the first of which is related to hamburgers and the second of which is related to wine glasses. That is, food is often served as a discrete unit (Ref. 12), either as a single object (eg. a hamburger or a packet of crisps) or as a single serving (eg. a wine glass or a plate / bowl). Both units have been increasing in recent decades.

For hamburgers, Ref. 13 found that a single McDonald's hamburger weighed 1.6 oz when first introduced in 1955 but in 2002 varied in weight from 1.6-8.0 oz, Burger King's weighed 3.9 oz in 1954 but 4.4-12.6 oz in 2002, and Howard Johnson's weighed 3.5 oz in the 1970s and 5.0-8.0 oz in 2002. These are hardly small increases.

This is part of a general trend, in which the sizes of food products sold as a single package have increased through time. The next graph, from Ref. 14, shows the time when a larger unit size was introduced (ie. larger than when the product was first marketed) for a total of 181 different food products in the USA. The increase in portions has been rapid since 1980.

Year in which increase in food portion size was first recorded

Moving on to the other category, products sold "loose" (rather than in a fixed package size), the next graph, from Ref. 15, shows the capacity of wine glasses available in England over the past 3 centuries. Once again, the increase has been rapid since 1980.

Increase in English wine-glass sizes over three centuries

This is also part of a general trend, in which the sizes of other dinnerware have also increased through time. For example, Ref. 16 shows a similar graph of increasing size of dinner plates in the USA over the past century.

Studies of drink consumption

This raises the obvious question of whether increased sizes of glasses has any affect on the amount of alcohol consumed. The general consensus seems to be "yes". Here, I will briefly discuss a few recent experimental studies about the wine glasses, tumblers, etc into which drinks are poured in licensed premises (restaurants and bars), rather than at home.

Several years ago, the authors of Ref. 17 visited a number of bars in California, to check whether the wines, mixed drinks and straight drinks varied in alcohol level based on the size or shape of the glass it was served in (due to variation in the bartender pouring). They concluded that: “Straight shots and mixed drinks served in the relatively large pint glass ... were found to contain more alcohol than drinks served in a short wide glass. No other significant differences were found between glass types.” That is, in this situation wine glass size and shape made no difference.

The same amount of wine in glasses of two different sizes

Later, the authors of Refs 18 and 19 did similar experiments for wine glasses at three bar / restaurants in England. However, in this case the portion of wine served was identical in all cases — the only thing that varied was the size of the wine glass (see the picture above). They found that: “Daily wine volume purchased was [10%] higher when sold in larger compared to standard-sized glasses. This effect seemed principally driven by sales in the bar area ... Findings were inconclusive as to whether sales were different with smaller vs. standard-sized glasses.”

So, serving people in over-sized glasses led them to purchase more glasses of wine. It is worth looking at how this effect is assumed to work. The Delboeuf Illusion is usually considered to be the prime force, as shown in the next picture (from Ref. 16) — the plate on the left looks like it has less food, even though the two amounts are identical. The same is true of the two wine glasses shown above — if it looks like you are drinking less per glass, then you buy more glasses.

The Delboeuf Illusion as applied to dinner plates

This idea was experimentally tested by the authors of Ref. 20, who asked their participants to fill a second glass with the same amount of wine as was present in the first (standard) glass — the second glass was either wider or taller or both, compared to the reference. The authors concluded that: “Participants under-filled the wider glass relative to the reference glass and and over-filled the larger glass relative to the reference glass. These results are broadly consistent with people using the relative fullness of glasses to judge volume, and suggest both the shape and capacity of wine glasses may influence perceived volume.” A similar conclusion was reached by the authors of a similar but much earlier study (Ref. 21).

However, these same authors (Ref. 22) also tested a different idea in the laboratory, that the effect might be produced by differences in drinking behavior — drinking rate, sip number and sip duration might be increased for larger glasses. Sadly, they found that: “Wine drunk from the larger, compared with the smaller glass, was consumed more slowly and with shorter sip duration, counter to the hypothesised direction of effect.” This might be due to the participants drinking alone in a laboratory rather than socially in a bar / restaurant.

However, for lager consumed from straight-sided versus curved beer glasses, the authors of Ref. 23 found that: “Participants were 60% slower to consume the alcoholic beverage from a straight glass compared to a curved glass. This effect ... was not observed for a non-alcoholic beverage. Participants also misjudged the half-way point of a curved glass to a greater degree than that of a straight glass.” So, glass shape mattered.


This inevitably leads us to consider the affect that of all of this research might have on marketing. Because the actual food and drinks are usually only a small percentage of the cost of eating and drinking, giving customers a larger portion is a strategy that might increase total sales. This leads to the widespread practice of “super-sizing” — giving customers a lot more food (and calories) for only a small additional cost. This seems to be the main motivation for sellers increasing their food portions, as described above.

People seem to purchase larger portion sizes than they actually need (in terms of energy), for two main reasons (Ref. 2):
  • "value for money" — larger portions are attractive because they offer more value for money, ie. a lower price per unit;
  • "portion distortion" — larger portions can be perceived as an appropriate amount to consume on a single occasion, due to continuous exposure to larger food portion sizes.
These two ideas combine, because people tend to "clean their plate" when given a particular portion of food — they consume what they are served, no matter how much (or little) it is.

For alcoholic beverages, marketing has been discussed by Ref. 24, particularly with regard to “strategies aimed at recruiting new young customers, increasing consumption and nudging drinkers towards stronger products ... Particular attention is paid to the size of the glass and how this can best be manipulated to increase sales.”

The authors quote a number of UK trade publications encouraging the use of larger glasses:
Since 1995, the legally permitted measure sizes for wine by the glass in the UK have been 125 ml, 175 ml and multiples thereof ... In 1996, a spokesman from Carlsberg-Tetley brands urged licensees to call 175 ml a “standard” glass and 125 ml “small” ... Similarly, in 2001, licensees were advised that ‘The basic step is to move from 125 ml glasses to 175 ml ... With 175 ml as the standard and 250 ml as the large size, pubs increase both volumes and profits’ ... A 2004 article declares that small glasses ‘give the impression that the pub doesn’t take wine seriously,’ and that ‘getting customers to “go large” in glass size not only improves the experience, it’s also a tremendously efficient way of cranking up profit.’
So, it does seem that the drinks business is actively involved in the current obesity epidemic.


There are "standard portions" described for many foods, usually based on daily diet requirements or other health considerations. For example, in the USA these standards are provided by the Food and Drug Administration (FDA), the Department of Agriculture (USDA), and the Centers for Disease Control and Prevention (CDC). These portions are often listed as "standard servings" on packaged food.

Although a "standard drink" is defined in terms of the amount of alcohol it contains, most people have little idea about what constitutes a standard drink in terms of volume, irrespective of whether it refers to wine, beer or spirits. In the experiment conducted by the authors of Ref. 25, almost all of the participants poured a drink that was too large.

Unfortunately, the so-called “standard drink” can vary considerably among countries. For example, in the USA a standard drink contains 14 g (0.6 oz) of ethanol, which would be 150 mL (5 fl. oz) of 12% ABV wine. However, in Australia, the standard drink contains only 10 g of ethanol, which is equivalent to 105 mL of wine. So, a bottle of 12% wine in the USA is considered to contain 5 standard drinks but in Australia it contains 7 such drinks. Obviously, with increased alcohol strengths of wine in recent years there are even more standard drinks per bottle these days!

Literature references

[1] Barbara J Rolls (2003) The supersizing of america: portion size and the obesity epidemic. Nutrition Today 38: 42-53.

[2] Ingrid HM Steenhuis, Willemijn M Vermeer (2009) Portion size: review and framework for interventions. International Journal of Behavioral Nutrition & Physical Activity 6: 58.

[3] M Barbara E Livingstone, L Kirsty Pourshahidi (2014) Portion size and obesity. Advances in Nutrition 5: 829-834.

[4] E Robinson, S Nolan, C Tudur-Smith, EJ Boyland, JA Harrold, CA Hardman, JCG Halford (2014) Will smaller plates lead to smaller waists? A systematic review and meta-analysis of the effect that experimental manipulation of dishware size has on energy consumption. Obesity Reviews 15: 812-821.

[5] David Benton (2015) Portion size: what we know and what we need to know. Critical Reviews in Food Science and Nutrition 55: 988-1004.

[6] Gareth J Hollands, Ian Shemilt, Theresa M Marteau, Susan A Jebb, Hannah B Lewis, Yinghui Wei, Julian Higgins, David Ogilvie (2015) Portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco. Cochrane Database of Systematic Reviews 9: CD011045.

[7] C Peter Herman, Janet Polivy, Lenny R Vartanian, Patricia Pliner (2016) Are large portions responsible for the obesity epidemic? Physiology & Behavior 156 177-181.

[8] Samara Joy Nielsen, Barry M. Popkin (2003) Patterns and trends in food portion sizes, 1977-1998. Journal of the American Medical Association 289: 450-453.

[9] Helen Smiciklas-Wright, Diane C Mitchell, Sharon J Mickle, Joseph D Goldman, Annetta Cook (2003) Food commonly eaten in the United States, 1989-1991 and 1994-1996: are the portion sizes changing? Journal of the American Dietetic Association 103: 41-47.

[10] Jeppe Matthiessen, Sisse Fagt, Anja Biltoft-Jensen, Anne Marie Beck, Lars Ovesen (2003) Size makes a difference. Public Health Nutrition 6: 65-72.

[11] Ingrid HM Steenhuis , Franca H Leeuwis, Willemijn M Vermeer (2010) Small, medium, large or supersize: trends in food portion sizes in The Netherlands. Public Health Nutrition 13: 852-857.

[12] Andrew B Geier, Paul Rozin, Gheorghe Doros (2006) Unit bias: a new heuristic that helps explain the effect of portion size on food intake. Psychological Science 17: 521-525.

[13] Lisa R Young, Marion Nestle (2003) Expanding portion sizes in the US marketplace: implications for nutrition counseling. Journal of the American Dietetic Association 103: 231-234.

[14] Lisa R Young, Marion Nestle (2002) The contribution of expanding portion sizes to the US obesity epidemic. American Journal of Public Health 92: 246-249.

[15] Zorana Zupan, Alexandra Evans, Dominique-Laurent Couturier, Theresa M Marteau (2017) Wine glass size in England from 1700 to 2017: a measure of our time. British Medical Journal 359: j5623.

[16] Koert Van Ittersum, Brian Wansink (2012) Plate size and color suggestibility: the Delboeuf Illusion’s bias on serving and eating behavior. Journal of Consumer Research 39: 215-228.

[17] William C Kerr, Deidre Patterson, Mary A Koenen, Thomas K Greenfield (2009) Large drinks are no mistake: glass size, not shape, affects alcoholic beverage drink pours. Drug and Alcohol Review 28: 360-365.

[18] Rachel Pechey, Dominique-Laurent Couturier, Gareth J Hollands, Eleni Mantzari, Marcus R. Munafò, Theresa M Marteau (2016) Does wine glass size influence sales for on-site consumption? A multiple treatment reversal design. BMC Public Health 16: 390.

[19] Rachel Pechey, Dominique‐Laurent Couturier, Gareth J Hollands, Eleni Mantzari, Zorana Zupan, Theresa M Marteau (2017) Wine glass size and wine sales: a replication study in two bars. BMC Research Notes 10: 287.

[20] Rachel Pechey, Angela S Attwood, Dominique‐Laurent Couturier, Marcus R Munafò, Nicholas E Scott-Samuel, Andy Woods, Theresa M Marteau (2015) Does glass size and shape influence judgements of the volume of wine? PLoS One 10: e0144536.

[21] Brian Wansink, John S Dyson, Koert van Ittersum (2005) Shape of glass and amount of alcohol poured: comparative study of effect of practice and concentration. British Medical Journal 331: 1512.

[22] Zorana Zupan, Rachel Pechey, Dominique‐Laurent Couturier, Gareth J Hollands, Theresa M Marteau (2017) Micro-drinking behaviours and consumption of wine in different wine glass sizes: a laboratory study. BMC Psychology 5: 17.

[23] Angela S Attwood, Nicholas E Scott-Samuel, George Stothart, Marcus R Munafò (2012) Glass shape influences consumption rate for alcoholic beverages. PLoS One 7: e43007.

[24] Martine Stead, Kathryn Angus, Laura Macdonald, Linda Bauld (2014) Looking into the glass: glassware as an alcohol marketing tool, and the implications for policy. Alcohol & Alcoholism 49: 317-320.

[25] Aaron M White, Courtney L Kraus, Lindsey A McCracken, H Scott Swartzwelder (2003) Do college students drink more than they think? Use of a free-pour paradigm to determine how college students define standard drinks. Alcoholism: Clinical & Experimental Research 27: 1750-1756.


  1. "Unfortunately, the so-called 'standard drink' can vary considerably among countries. For example, in the USA a standard drink contains 14 g (0.6 oz) of ethanol, which would be 150 mL (5 fl. oz) of 12% ABV wine. However, in Australia, the standard drink contains only 10 g of ethanol, which is equivalent to 105 mL of wine. So, a bottle of 12% wine in the USA is considered to contain 5 standard drinks but in Australia it contains 7 such drinks. Obviously, with increased alcohol strengths of wine in recent years there are even more standard drinks per bottle these days!"

    Elaborating on this USA example.

    In decades past, California vineyard managers and winemakers studiously followed their UC Davis and Fresno State enology professor instructions and chose their harvest date based on the brix reading using a refractometer. By today's standards, those comparatively less ripe grapes resulted in 12% to 13% Alcohol By Volume (ABV) wines.

    Today in California, vineyard managers choose their harvest date based on what is called the "physiological ripeness" of the grape. These contemporary riper grapes result in 14% to 16% Alcohol By Volume (ABV) wines.

    By way of example, a 2014 California Zinfandel clocking in at 15% ABV is +20% higher in alcohol per fluid ounce than a 1984 California Zinfandel clocking in at 12.5% ABV.

    (Aside: The highest naturally attained -- not fortified like Port -- ABV in a California wine was the 1978 Mayacamas "Late Harvest" Zinfandel. See this label photo with its 17% ABV declaration: )

    Today, higher ABV levels combined with more generous (than 5-ounce) pours in larger bowl wine glasses result in more alcohol intake per "glass" of wine consumed.

    As the writer of a Wall Street Journal article [*] on the subject observed:

    "A standard 'serving' for an alcoholic beverage is 5 fluid ounces of wine, 12 ounces of regular beer or 1.5 ounces of distilled spirits, according to the Centers for Disease Control and Prevention. All three portions contain 0.6 ounce of alcohol. But glasses today come in so many different shapes and sizes -- tall 'highballs,' wide tumblers, bowl-shaped wine goblets and now the new popular stemless wine glasses -- it's virtually impossible to estimate the right serving amount. Although a traditional wine glass holds about 7 ounces, many wine glasses today hold 16 ounces or more. Beer glasses often hold 20 ounces.

    "'Often my clients think they are just having one or two drinks, when really they're having more like three or four,' says Lisa R. Young, a New York University nutritionist.

    "Dr. Young says that a solution is for drinkers at restaurants to count each glass of wine, beer or spirits as two servings."

    ["The Accidental Binge Drinker: How Much We Really Pour," Tara Parker-Pope, The Wall Street Journal "Personal Journal" Section, May 1, 2007, Page D1.


    Related: see David's proffered literature references 13 and 14.]

  2. California-based wine writer W. Blake Gray has been reporting on the implications of higher alcohol levels in USA wines, given a change in out tax laws.

    First blog:

    "Public safety alert: Americans could get more drunk than they expect because of tax law change"



    "One important aspect is that it raises the amount of alcohol allowed in "table wine" from 14% to 16%. Previously, wines over 14% were taxed at a higher rate. Now, that higher-taxation line moves to 16%.

    "By itself, this is a good change that reflects the way wine is made in the United States today. But there's a catch that could be not just bad for wine lovers, but dangerous: label tolerance, or how much a winery is legally allowed to misstate a wine's true alcohol level on the label.

    "Currently, the label ABV must be accurate within 1.5% for wines under 14% alcohol, and within 1% for wines over 14%. However, because of the difference in taxation, 14% was a dividing line that a winery could not legally cross. A wine labeled at 13.5% alcohol might actually contain 12% or 14%, but it could not contain 14.1%."

    Second blog:

    "New wine tax law is more tolerant of alcohol error, apparently"



    "... For many years, wine was taxed at about 21 cents per 750 ml bottle if it contained less than 14% alcohol, and 31 cents a bottle if it contained more than 14%.

    "As part of the tax reform passed by Congress last month, that line of higher taxation moved up to 16%. All wines under 16% alcohol (except sparkling wine, sigh) will be taxed at the lower 21-cent rate. This is good. Winemakers shouldn't have to make decisions based on taxation rate.

    "However, the label tolerance -- the amount a winery can legally misstate the actual alcohol percentage -- was not addressed in the tax reform. And that was very important. Previously, wines under 14% ABV had a label tolerance of 1.5%, while wines over 14% had a tolerance of 1%. But, and this is key, the label had to be on the correct side of 14%, so that a wine labeled at 13.5% might have 12% alcohol or 14%, but not 14.1%.

    "Last Friday, the TTB sent out a mailer that answers some questions about the way the tax law will be enforced. ...

    "What is NOT addressed in this answer is whether or not 14% is a line that cannot be crossed. I'm sorry, but I don't know the answer to that, and the TTB might not yet either.

    "Previously, the tolerance was based as much on the label ABV as the actual ABV. Both had to be on the same side of 14%. But now, the tolerance is based purely on the actual ABV, apparently.

    "What this means is that a wine that says 14.5% on the label might actually contain as much as 15.5%, and it might actually contain as little as 13%, which would have been impossible before.

    "I believe this means almost all wines now have a 2.5% swing in label tolerance from least to most, which is actually larger than before.

    "In other words, a wine labeled at 13.9% might actually have anywhere from 12.4% to 14.9%."