The nagging question “Which is better, Coca-Cola or Pepsi-Cola?” sprang from an earlier, more basic question: “Can anyone tell the difference?”
Professor Nicholas H. Pronko and colleagues at the University of Wichita, Kansas, conducted a series of experiments in the 1940s and 1950s. They wrote five studies that brought rigor, sophistication, and cachet to the testing of Coke/Pepsi taste- discrimination.
Pronko’s final Coke/Pepsi paper appeared in 1958. In the ensuing half- century, other investigators have digested and challenged his methods and findings.
Advances in technology led, many years later, to investigation of the brain activity of
Coke and Pepsi drinkers. Here is a look back at the early work, the foundation upon which rests so much later planning, effort, and thoughtfully sipped cola.
This history is in some ways emblematic of experimental psychology as a whole—of its maturation and growth as an academic discipline.
Vess Cola, an obscure competitor to both Coca-Cola and Pepsi-Cola, was included in Pronko’s first “taste this and tell us what kind of cola it is” experiment. The test subjects could not reliably distinguish its taste from those of Coke, Pepsi or Royal Crown Cola.
Pronko’s first study asked not just one, but a series of complex, interrelated questions. Volunteer drinkers, some of them habitual cola drink drinkers, some of them not, tasted samples of four different kinds of cola: Coca-Cola, Pepsi-Cola, Royal Crown Cola (a brand that at the time was, like Coke and Pepsi, widely popular), and Vess Cola. Vess Cola was then (and remained) little known.
“Identification of Cola Beverages. I. First Study,” N.H. Pronko and J.W. Bowles, Jr.,
Journal of Applied Psychology, vol. 32, no. 3, June 1948, 304–12.
108 college students tasted and named four different brands of Cola beverages.... From one third to over two thirds of the responses were incorrect on the basis of the subjects’ likes and dislikes. It is concluded that the identification response of Cola beverages is not a function of the physico-chemical properties of the stimulus objects, but a matter of using an available tag or label for it based largely on familiarity.
Pronko and collaborator J.W. Bowles drew several conclusions. One—the most enduring—was that people cannot reliably identify the taste of Coke from that of Pepsi. A billboard advertising Pepsi-Cola.
The second study came out a mere six months after the first. Spurred and stimulated by criticism—mostly about the way they had labeled the cola glasses (the labels were W, X, Y and Z)—Pronko and Bowles simplified the experiment. This time there were three, rather than four different cola drinks, and the glasses were labeled X, Y and Z:
“Identification of Cola Beverages: II. A Further Study,” N.H. Pronko and J.W. Bowles, Jr., Journal of Applied Psychology, vol. 32, no. 5, October 1948, 559–64.
156 students in elementary psychology served as subjects in a taste experiment with three colas. When presented in random order and also when only one cola was given for an entire series, the results are comparable to a chance distribution and support the hypothesis that the pattern of naming responses was a function of the subjects’ familiarity with cola brand names.
Pronko, Without Coke or Pepsi
Pronko and Bowles then embarked on their third set of experiments. They began by reminding themselves that “when subjects were asked to identify the three leading brands of Cola, they might just as well have drawn their names out of a hat.” So this time, Pronko and Bowles used only obscure brands of cola.
Sipping three kinds of cola that they had probably seldom or never tasted before, almost everyone nonetheless said they were drinking Coke, Pepsi or Royal Crown, the USA’s three most domestically popular cola brands:
“Identification of Cola Beverages. III. A Final Study,” N.H. Pronko and J.W. Bowles, Jr., Journal of Applied Psychology, vol. 33, no. 6, Dec 1949, pp. 605–8.
60 subjects tasted samples of Hyde Park Cola, Kroger Cola, and Spur Cola. No correct identifications were made. Almost all responses identified the beverages in order as Coca Cola, Pepsi Cola, or Royal Crown Cola.
Pronko and Bowles intended this to be the end of things, to settle once and for all the
basic questions that scholars may have had about whether human beings can
reliably distinguish between, and identify, the tastes of Coke and Pepsi. But, as happens so often in modern research, nagging, niggling doubts remained.
And so N.H. Pronko, teaming up this time with colleague D.T. Herman, ran a fourth set of experiments. This time, there would be no cola anonymity. The volunteer cola drinkers were explicitly told they would be sampling Coca-Cola, Pepsi-Cola and Royal Crown Cola. The task before them was starkly simple: identify which glass contained which cola. This time, the results were almost the same as before, but with one new twist:
“Identification of Cola Beverages. IV. Postscript,” N.H. Pronko and D.T. Herman, Journal of Applied Psychology, vol. 34, no. 1, February 1950, pp. 68–9.
Where subjects were given 3 samples of the same beverage, their identifications were not significantly different than when the beverages were unknown Colas or actually these Colas unspecified. The same was true for Pepsi Cola and Royal Crown Cola identifications when 3 different samples were given. However, Coca Cola identifications yielded statistically significant results.
The “statistically significant” results for Coca-Cola were, technically, “significant.” They were also small. But, for the first time, there was evidence that maybe a few people, occasionally, might be able to reliably, if tentatively, identify the flavor of Coca-Cola.
Pronko’s Side Exploration
Here is a side note. Eight years later, Pronko tried a very different experiment involving colas. It did not involve the taste of the colas themselves. Rather, it examined people’s ability to recognize cola bottle caps, cola bottles and cola logos. Details are in the study:
“Identification of Cola Beverages: V. A Visual Check,” G.Y. Kenyon and N.H. Pronko, Journal of Applied Psychology, vol. 42, no. 6, December 1958, pp. 419–22.
“A group of 210 subjects was asked to identify 45 tachistoscopic slides presented
at 1/400-seconds exposure.” Slides of bottles, bottle caps, and type-written brand names of 3 leading colas were presented singly, in pairs, or as triples. Subjects showed a greater accuracy in responding to the brand names Coca Cola, Pepsi Cola, and Royal Crown Cola in that order.
A Cola Experiment in the Middle East
The Pronko experiments, wide-ranging as they were, examined only the ability of American cola-drinkers to recognize the flavors of American colas. The larger world, with its wider range of colas, was still virgin territory for cola-centric experimental psychology.
Edwin Terry Prothro of Brooklyn College changed that. His experiment, documented in a 1953 paper in the Journal of Applied Psychology, supplied data that still, these many years later, awaits its definitive interpreter:
“Identification of Cola Beverages Overseas,” E. Terry Prothro, Journal of Applied Psychology, vol. 37, no. 6, December 1953, pp. 494–5.
60 students in American University of Beirut were asked to identify Coca-Cola, Pepsi-Cola, and a local cola, in blindfold taste tests. Ten students each took the three beverages in one of the six possible orders. The American brands could not be distinguished from one another by the subjects, but the local beverage was correctly identified in 51 out of 60 identifications.
A recent portrait of Frederick J. Thumin, who in 1962 quibbled with N.H. Pronko’s corpus of cola work.
Enter the Quibblers
Until 1962, the four Pronko studies (though not the fifth) stood as a colossus, intimidating all who might consider criticizing them. Then along came Frederick J. Thumin of Washington University.
Thumin tried to point out what he saw as glaring defects in Prinko’s work. In October 1962, Thumin hurled a javelin, metaphorically speaking:
“Identification of Cola Beverages,” Frederick J. Thumin, Journal of Applied Psychology, vol. 46, no. 5, October 1962, pp. 358–60.
An attempt was made to overcome certain methodological inadequacies of earlier studies in determining whether cola beverages can be identified on the basis
of taste. Some 79 subjects completed questionnaires on their cola drinking habits and brand preferences, then were tested individually on samples of cola beverages presented under methods of paired comparisons. Significant chi square values
were obtained for Coca Cola and Pepsi Cola, due to the large number of correct
identifications for these brands. Correct identification of Royal Crown, however, did
not differ from chance expectancy.
Thumin’s report implied that if people drink lots of cola regularly, they can develop a slight, perhaps temporary, ability to identify one brand from another, at least sometimes.
The Quibbler Quibbled
A decade later, three researchers at Cleveland State University took aim at Thumin. Their study, published in the same journal in which most of the cola warfare had already transpired, implied (a) that Pronko was mostly wrong, and (b) that Thumin was mostly wrong:
“Cola and Diet Cola Identification and Level of Cola Consumption,” Sam H. Lane, James Zychowski and Kenneth Lelii, Journal of Applied Psychology, vol. 60, no. 2, April 1975, pp. 278–9.
Assessed the ability of 72 undergraduates to identify correctly 2 beverages (Coca Cola and Pepsi Cola) and their diet complements (Tab and Diet Pepsi Cola). Subjects were administered the 4 beverages in a paired-comparisons design, and questionnaire data were collected regarding their beverage consumption habits. Results indicate that subjects could correctly identify each beverage in excess of chance expectancy, and that their ability to identify beverages was not related to their level of consumption.
And so it goes, quibble begetting quibble begetting quibble.
The basic question—”Is it Coke or is it Pepsi?”—is now addressed often, if not always rigorously, in introductory psychology courses around the world. It is a simple experiment that most students can, at least in theory, get excited about and perhaps even carry out.
Postscript 1: Brain Scans of Pepsi vs. Coke Drinkers
In 2005, nearly two decades after Sam Lane, James Zychowski and Kenneth Lelii had at the cola work of Frederick Thumin, and four decades after Thumin took on the imposing cola legacy of Nicholas Pronko, and more than fifty years after Pronko himself began the series of projects that would make him the name of names in early Coke/Pepsi taste-discrimination research, six scientists in Texas got the first half-good look at part of what happens in portions of the brain of a cola drinker while that cola drinker is drinking cola:
A recent portrait of P. Read Montague of Baylor College of Medicine. Dr. Montague has studied brain activity of cola drinkers.
“Neural Correlates of Behavioral Preference for Culturally Familiar Drinks,” Samuel M. McClure, Jian Li, Damon Tomlin, Kim S. Cypert, Latané M. Montague, and P. Read Montague, Neuron, vol. 44, October 14, 2004, pp. 379–87. The authors, who are at Baylor College of Medicine, say:
We report a consistent neural response in the ventromedial prefrontal cortex that correlated with subjects’ behavioral preferences for these beverages. In the brand-cued experiment, brand knowledge for one of the drinks had a dramatic influence on expressed behavioral preferences and on the measured brain responses.
Postscript 2: Controversy Begets Controversy
The basic question—Is it Coke or is it Pepsi? —remains unanswered.
In the 1960s, Coke and Pepsi engaged in a marketing battle, each company announcing the results of blind taste tests that it had conducted. Volunteer cola-drinkers sampled cola from several glasses. Each glass was marked only with an identifying letter such as M or Q.
The tests conducted by Pepsi found that people preferred the taste of Pepsi.
The tests conducted by Coke found that people preferred the taste of Coke. Coke also announced that when it offered a choice of two glasses, both containing Coke, people said they preferred the taste of one glass of Coke to the taste of another, differently marked, glass of the exact same beverage.
This battle of company-sponsored research itself led (as many things do) to a spate of
academic squabbling. The Coke versus Pepsi question led, or fed, an ongoing controversy about whether people are prejudiced for against particular letters (for example, whether something marked “M” is probably better than something marked “Q”).
The burgeoning mélange of contention is exemplified by this passage from a 1977 study by Kenneth A. Coney of Arizona State University:
“Order-Bias: The Special Case of Letter Preference,” Kenneth A. Coney,
Public Opinion Quarterly, vol. 41, no. 3m Autumn, 1977, pp. 385–8.
Pepsi conducted blind taste tests using letters [for] identification and determined that Coke drinkers prefer Pepsi over Coke. Coke has countered that the letters used to identify their brand have negative connotations causing the choice of Pepsi.
This study is neither the first nor the last word on Coke or Pepsi or letter preference. Perhaps a study published six years later by researchers at King’s College in Wilkes Barre, Pennsylvania, is the last word:
“Pepsi versus Coke: Labels, Not Tastes, Prevail,” Mary E. Woolfolk, William Castellan and Charles I. Brooks, Psychological Reports, vol. 52, no. 1, February 1983. pp. 185-6.
30 subjects were asked their preference for either Pepsi or Coke. Then they drank from a Pepsi bottle (which contained Coke) and from a Coke bottle (which contained Pepsi). Subjects were significantly influenced by the label of the product they preferred and not by taste differences between these products. It is concluded that a taste comparison of colas should avoid using any labels.
And perhaps it is not.
_____________________The article above is from the January-February 2007 issue of the Annals of Improbable Research. You can download or purchase back issues of the magazine, or subscribe to receive future issues. Or get a subscription for someone as a gift!
Visit their website for more research that makes people LAUGH and then THINK.