Similarity (different/similar)

Factsheet Similarity (different/similar)
O P P
product research
Similarity
(different/similar)
Similarity (different/similar)
O P P
product research
Can two variants be distinguished from each other or are they equal? Does the consumer notice that an ingredient is changed with
another ingredient? Is the product from another supplier equal to the current product? In all cases there is a very small difference and
we want to know if it is noticed by the consumers.
Very often we don’t want to know if the two variants differ from each other but if they are the same.
To determine if two variants differ significantly from each other is rather easy, to conclude if they are the same is much more difficult.
If two variants don’t differ significantly from each other this does not mean that they are the same. When only 9 of 24 assessors detect
a difference this is not significant, but this does not implies that the two variants are the same. If instead of 24 assessors there had been
120 assessors and 45 of them detect a difference (same proportion) than the difference was highly significant.
We can only conclude that there is no difference if we have searched well enough, with other words if the sample size was sufficiently
large. To get a reliable answer for the question if there is no difference the power of the test must be sufficiently high (we should have
search well enough).
The most frequently used difference test is sensory research is the triangle test. The respondent has to select the odd one from a set
of three samples (two A and one B). For a triangle test with sufficient power to conclude that there is no difference a sample size of at
least 240 people/judgments is needed. The tetrad test has significantly more power and therefore approximately 80 judgments are
enough.
With the tetrad test the assessors get four instead of three samples. With a triangle test they have to indicate the odd one, with a tetrad
they have to indicate which two pairs are the same. In a session of 45 minutes at most3 tetrads can be evaluated per person (this can
be three different products or three replicates).
Besides the triangle and tetrad test there is a variety of other difference test, the choice depends partly on the kind of product that has
to be tested.
Tetrad test
Set up
In one 45 minute session, each assessor can judge four tetrads
(this can be four different pairs or four presentation of the same pair)
Session duration
45-60 minutes
# samples
Every presentation consists of four samples (two A and two B)
Time between presentations 5 minutes
Preparation
Depending on the product
Sample size
Groups of maximal 16 consumers (minimum 80)
Incentives
On hourly basis
Respondents
Target group or on request
Report
Power point report