Testing Environment and Sample
and Serving Design
The respondents tested the products in Portland, Oregon,
at an Oregon State University-identified canopied
booth during the outdoor “Bite of Oregon” food festival.
Other booths at the event featured regionally produced
foods, fine wines, and microbrewed beer. Music, chef
demonstrations, and competitions also took place. Although
this is a busy environment, Hersleth et al. (2005)
noted that location is not a factor in consumer acceptance
of cheese. Testing times ran from 11 a.m. to 10 p.m. over
2 d. Consumers received one pair of samples consisting
of a pasteurized and an unpasteurized milk cheese from
1 of the 3 producers. The samples were labeled with
product descriptors (“pasteurized” or “unpasteurized”)
or with 3-digit random codes in a multiple-sample presentation
served in balanced pairs across respondents.
Half of the consumers saw the product descriptor labels
and half saw the 3-digit random codes. The first 50
consumers received plates with 3-digit randomized code
identifiers (without any product information), whereas
the next 50 received plates with the labels “pasteurized”
and “unpasteurized” cheese, with switching continuing
after every 50 consumers throughout the test (Table 1).
Consumers were asked to taste both samples from left
to right and select the sample they preferred. They were
also given the options of “no preference” and “I can’t tell
a difference” (Lawless and Heymann, 1998). Bench testing
of the 8 cheeses was done ahead of time and it was
determined that each of the 2 cheeses within a cheese
pair was sufficiently different to conduct a preference
study. Because this was an off-site study conducted at
a food festival not specifically related to cheese, it was
unclear if the consumers attending, who would be the
participants of this study, would be frequent consumers
of specialty cheese. For this reason, the “I can’t tell a
difference” option was added to the paired preference
question to allow flexibility for those consumers who felt
they could not tell the difference between the 2 samples.