Running head: TASTE TEST WITH TWO COLAS

This was an experiment that investigated peoples ability to discriminate among 2 types of cola, Pepsi and Coke. Our hypothesis was that they would prefer Pepsi to Coke. Thirty opportunistically selected people participated in two randomly assigned 15 people groups for counterbalancing. A Likert scale was used to collect the data of subject’s ratings for each cola. The results were that people could distinguish between these two colas. They significantly preferred Pepsi to Coke, which agreed with our hypothesis.

Many experiments have investigated the sense of taste. Among some of the more famous experiments, the ‘Pepsi Challenge’ taste test has become part of American pop culture. Some of the more interesting experiments in this area have studied how people distinguish among similar commercial foods. One example of a taste test similar to the one we performed was the Thumin (1962) experiment, the "Identification of Cola Beverages.’

The purpose of our experiment was to investigate whether people could discriminate among colas solely on the basis of taste. Our null hypothesis was that people would be unable to distinguish between the colas-Pepsi and Coke-that we presented to them. Our hypothesis was that people would prefer Pepsi to Coke.

Participants

Thirty student and adults were opportunistically selected from outside the library at the Lewis and Clark campus and randomly assigned to one of two 15-people groups.

Materials

Two colas were tasted, with the same size cups of water in between each cup of cola. One was canned Pepsi, and the other canned Coke. They were given a general explanatory consent form and each participant signed them to show they had read and understood the experiment. A seven point Likert scale was used to rate the subjects scoring of each cola.

Procedure

Subjects were given a cup, asked to drink and rate the taste according to the Likert scale. The instructions we gave subjects and recording their responses were all done orally; we used no written instructions. The subjects were given cups of water in between tasting the two colas, and the colas were served in opposite order for each group, for counterbalancing. The procedure was double blind, that is, the experimenters who served and recorded the data, along with the subjects, did not know which cola was which.

Data was collected by listing each subjects’ rating of the particular cola on the Likert scale. We were testing whether our subjects could discriminate among colas purely on the basis of taste. Our results indicated that the subjects preferred Pepsi over Coke. The mean for Coke-preferring subjects was M=4.25 (1.65); which was lower than the mean for subjects preferring Pepsi, M=5.12 (1.28). We ran a within-subject t-test which turned out to be significant, (29)=2.832, p=.0083.

The results were that our subjects preferred Pepsi to Coke, which agreed with our hypothesis. People can and do discriminate amongst these two colas by taste.

We ran into multiple confounding variables and difficulties that could be eliminated in a similar future experiment. One of the possible confounds we encountered was that since outside the library where we conducted our experiment is a popular smoking spot on campus, often people would smoke while tasting the different colas. This could conceivably affect their perception of the flavor of each cola. In the future, perhaps when we have more willing subjects and can afford to be stricter with the guidelines of the experiment, we can ask the subjects not to ingest anything other than what we’re giving them, for the duration of the taste-test experiment. The lack of willing subjects was another difficulty, since cajoling them into participating could have biased our results.

Another possible confound was our lack of consistent instructions. We failed to make a set of written instructions for administering the test. Next time, we should make sure to standardize what we tell each subject, so that we don’t introduce an extraneous variable into the experiment by telling different things to each subject.

Another possible confound had to do with the amount of liquid the subjects drank. Some subjects drank all of the cola in the cup, and others drank only a sip. The varying levels ingested could have been a confound, especially when that might have influenced their preference. It could be that when drinking only a small amount of cola, people might prefer a sweeter taste, but when drinking a whole can, they prefer a less sweet taste. Since Pepsi has a sweeter taste than Coke, the different amounts that our subjects ingested might have been a serious confound. Also many subjects only drank a tiny sip of water, and others all of the water in the cup they were given, so that some subjects might have had cleaner palates than others. In the future, to account for these possible confounds, one could make sure to standardize the amount of all liquids the subject drinks. Also, a future experiment could include different levels of cola for different groups. This might help control for the possible interactive effects the different levels of sugar and the amount they drink of each cola have on people’s preferences.

Publication Manual of the American Psychological Association. (1998). American Psychological Association. Washington, DC: APA.

Thumin, F. J. (1962). Identification of Cola Beverages. Journal of Applied Psychology, 46, 358-360.