Abstract:

 

Sequence effects introduce bias in product testing.  If not properly balanced, this bias can cause product means to be inaccurate and product comparisons to be misleading.  This is a well-known issue, but there has been a lack of good solutions to address it.  In this webinar we provide a general, computer intensive, approach to address this problem and also review current knowledge on sequential effects.

​

Using examples from actual consumer investigations we first illustrate how ratings for the same product, such as a beverage or a shampoo/conditioner bundle, can vary up to a full rating scale unit depending on its position in the design or the nature of the sample that precedes it.

​

We continue by focusing on three main rotation generation approaches: Order randomization within a subject, William Square designs, and the column randomization and search method or ‘CR&S’.  We compare these methods and show how CR&S is best equipped to minimize position, sequences, and sequence position biases.

​

We first discuss complete block designs and then extend our overview to more complex situations including multiple-day research and incomplete block designs.  In multiple-day applications, sequences between days are less relevant than those found within a day.  Incomplete designs will also be illustrated with optimal and near-optimal rotation sequence optimization.

​

Note: A software implementation of the CR&S method can be found in the IFPrograms® software.  Webinar participants not currently holding a subscription can request a two-month trial license to become more familiar with this design balancing method and see how it can help improve the quality of their sensory and consumer collected data.

​

This webinar is intended for a general audience of sensory professionals and graduate students. No detailed technical knowledge is assumed.

​

A digital recording of this webinar is available for purchase ($289).