Paired Preference Tests: “50:50” and “Alternating” No Preferences

Abstract

Because the concept of “No Preference” is ambiguous, two types of “No Preference” are defined: a “50:50” No Preference, defined operationally as consumers presented with both products having a 50% chance of choosing either, and an “Alternating” No Preference, where the probability is not 50%. The consumer would be more likely to choose one of the products. However, later, he or she might choose the other product. To investigate this, paired preference tests were performed with and without the “No Preference” option for various products. No discernible trend was noted for consumers who had chosen the “No Preference” option and distributed their votes among the preference options. These consumers and those who changed their preferences were classified according to the “50:50” No Preference and “Alternating” No Preference dichotomy accordingly. PRACTICAL APPLICATIONS Besides the tendency of consumers to report false preferences, paired preference tests have another problem: the concept of “No Preference” is ambiguous. It could mean that consumers liked both products equally and so would always have a 50% chance of choosing either. It could also mean that the probability is not 50% on any one occasion but varies so that one product is chosen on one occasion and the other product on another occasion. To avoid confusing the consumer so as to obtain more valid data, it is important to specify these two meanings of “No Preference.” This study is a first step in that direction and should help consumers understand better the instructions for preference tests.

Assessment of Preference With Controls for Response Bias Operating in the Test Situation: A Practical Example Using Omega-3 Enriched Wholegrain Breads With Ecuadorian Consumers

“…The second approach for utilizing chi‐squared is simpler and accordingly has been applied by various authors (Alfaro‐Rodriguez et al . 2007, 2008; Kim et al .…”

“…2003; Chapman and Lawless 2005; Chapman et al . 2006; Alfaro‐Rodriguez et al . 2007, 2008; Kim et al .…”

Applying Disruptive Preference Test Protocols to Increase the Number of “No Preference” Responses in the Placebo Pair, Using Chinese Consumers

“…The “No Preference” frequencies vary but are mostly in the range 20% to 30%. (Marchisano and others ; Chapman and Lawless ; Chapman and others ; Alfaro‐Rodriguez and others , , ; Kim and others ; Angulo and others ; Alvarez‐Coureaux and others ). Consequently, rather than using a “one‐size‐fits‐all” identicality norm, it was considered appropriate for a sample of consumers to generate their own identicality norms, according to the appropriate experimental conditions, the products being tested and the types of consumers making the assessments.…”

Paired Preference Tests: A signal detection based analysis with separate d′ values for segmentation

“…Ennis and Collins's () 40‐20‐40 frequency distribution for putatively identical stimuli does not seem to be general. For reasons as yet unresolved, other authors (Alfaro‐Rodriguez, Angulo, & O'Mahony, ; Alfaro‐Rodríguez, O'Mahony, & Angulo, ; Alvarez‐Coureaux, Aguilar, O'Mahony, & Angulo, ; Angulo, Okayama, Nakamura, Yuen, & O'Mahony, ; Calderón et al, ; Chapman, Grace‐Martin, & Lawless, ; Chapman & Lawless, ; Sung et al, ; Kim, Lee, O'Mahony, & Kim, ; Marchisano et al, ; Xia et al, ) found different frequencies, the numbers varying with the products being tested, the experimental conditions, the types of consumers tested and the types and numbers of response options available in the test. The frequencies of reported “No Preferences” vary a great deal but most are in the range 20–35%.…”

“…The safest approach at the present time, would appear to be to require consumers to assess the target (different) pair and also the putatively identical (placebo) pair to give a measure of the effect of the extraneous factors. This was adopted and involved giving consumers two pairs to assess in a given test (Alfaro‐Rodriguez et al, ; Alvarez‐Coureaux et al, ; Kim et al, ; Sung et al, ). Using Chi‐squared, the responses elicited by the target pair of (different) stimuli were not compared with the null hypothesis but were instead compared with the hypothesized identicality norm, derived from the putatively identical pair, to see whether they were significantly different.…”