|
SensoGram 20: The Spurious Paired-Comparison Test
The paired-comparison taste test is the simplest and most time-honoured of sensory methods: taste this, now taste the other; which do you prefer?
This seems a plausible method and has been used in countless product tests over the years. The fact is, however, that the paired-comparison test can be misleading - sometimes seriously so.
Perhaps the first objection to this test is that it does not simulate the real-world. How often does the consumer sit down to two competitive products, side by side, and conduct a taste comparison?
But there is a larger issue at stake. In the paired comparison test, respondents are forced to make a single preference decision between two products which, in most cases, differ on several sensory dimensions (eg appearance, aroma, flavour character, sweetness, texture etc). This decision is easy if one product is vastly superior to the other, but not so straightforward when the products are close. For example, one product may have a better flavour but the other may have a better mouth feel. In this event, what respondents tend to do - although they may not be consciously aware of it - is to select one particular sensory dimension and base their preference judgment on it alone.
Often the sensory dimension which is (unconsciously) selected turns out to be sweetness, with the sweeter product being preferred. Note that this can and does happen even with products that are not conspicuously sweet (eg savoury sauces).
Yet when the same products are evaluated in a one-at-a-time protocol, we find that the sweetness differential, while detected, does not necessarily confer a liking advantage on the sweeter product. In one-at-a-time evaluation, all sensory attributes can be given their due consideration, and the taster is not forced by context to weight artificially any one attribute.
The implication of all this is that the preference obtained from the paired-comparison test can sometimes be an artifact of the methodology.
This is not merely a point of academic interest; billion dollar decisions have rested upon it! You may remember the paired-comparison `Pepsi Challenge' taste test. Many say that Pepsi tastes sweeter than Coke; consequently, in a paired-comparison test, Pepsi can tend to be preferred. It was this, according to reports, that prompted Coca-Cola to reformulate ... and the rest is marketing history.
Does this rule out the paired-comparison test? Not necessarily. But we need to be aware of when it is liable to mislead.
|