How can you prefer one of two identical items?
- dsi657
- 1 day ago
- 2 min read

๐ฆ๐ฒ๐ป๐๐ผ๐ฟ๐ ๐๐ฐ๐ถ๐ฒ๐ป๐ฐ๐ฒ would go nowhere ๐๐ถ๐๐ต๐ผ๐๐ ๐ฝ๐๐๐ฐ๐ต๐ผ๐น๐ผ๐ด๐. I am often reminded that some basic principles involving psychological components can appear ๐ฐ๐ผ๐๐ป๐๐ฒ๐ฟ๐ถ๐ป๐๐๐ถ๐๐ถ๐๐ฒ to many, especially for those new to our field.
๐๐ฎ๐๐ฒ ๐ถ๐ป ๐ต๐ฎ๐ป๐ฑ: How to interpret the preference choices in a ๐ฏ๐น๐ถ๐ป๐ฑ preference test with a โNo preferenceโ option? What if the two items are ๐ฝ๐๐๐ฎ๐๐ถ๐๐ฒ๐น๐ ๐ถ๐ฑ๐ฒ๐ป๐๐ถ๐ฐ๐ฎ๐น?
๐ฐ ๐๐ป ๐ผ๐ฏ๐๐ถ๐ผ๐๐ ๐ฎ๐ป๐๐๐ฒ๐ฟ
When I ask students what results they would expect from this question:
โข Prefer one of the two: 0%
โข ๐ก๐ผ ๐ฝ๐ฟ๐ฒ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ฐ๐ฒ: ๐ญ๐ฌ๐ฌ%
โ๏ธ ๐๐ป ๐๐ป๐ฒ๐ ๐ฝ๐ฒ๐ฐ๐๐ฒ๐ฑ ๐ผ๐๐๐ฐ๐ผ๐บ๐ฒ
As many of you are aware, the ๐ฎ๐ฐ๐๐๐ฎ๐น outcome will be much closer to:
โข Prefer one of the two: 80%
โข ๐ก๐ผ ๐ฝ๐ฟ๐ฒ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ฐ๐ฒ: ๐ฎ๐ฌ%
This clearly challenges the meaning of โprefer one of the twoโ and how such results should be interpreted.
๐ง๐๐ผ ๐บ๐ฎ๐ถ๐ป ๐ฑ๐ฟ๐ถ๐๐ฒ๐ฟ๐ ๐ณ๐ผ๐ฟ ๐๐ต๐ถ๐ ๐๐ป๐ฒ๐ ๐ฝ๐ฒ๐ฐ๐๐ฒ๐ฑ ๐ผ๐๐๐ฐ๐ผ๐บ๐ฒ
- ๐ฅ๐ฒ๐๐ฝ๐ผ๐ป๐๐ฒ ๐ฏ๐ถ๐ฎ๐: The subject feels a subconscious drive to โpick a winner.โ
- ๐ ๐ผ๐บ๐ฒ๐ป๐๐ฎ๐ฟ๐ ๐ฝ๐ฒ๐ฟ๐ฐ๐ฒ๐ฝ๐๐๐ฎ๐น ๐๐ฎ๐ฟ๐ถ๐ฎ๐ฏ๐ถ๐น๐ถ๐๐, cornerstone of the Thurstonian/Signal Detection Theory frameworks. Our senses are noisy and no two momentary perceptions ๐๐ถ๐น๐น ๐ฒ๐๐ฒ๐ฟ ๐ฏ๐ฒ ๐๐ฟ๐๐น๐ ๐ถ๐ฑ๐ฒ๐ป๐๐ถ๐ฐ๐ฎ๐น.
๐ง๐ต๐ฒ ๐ป๐ฒ๐ฒ๐ฑ ๐ณ๐ผ๐ฟ ๐๐ฝ๐ฒ๐ฐ๐ถ๐ฎ๐น๐ถ๐๐ฒ๐ฑ ๐ฎ๐ป๐ฎ๐น๐๐๐ถ๐ฐ๐ฎ๐น ๐ฎ๐ฝ๐ฝ๐ฟ๐ผ๐ฎ๐ฐ๐ต๐ฒ๐ ๐ถ๐ป๐ฐ๐ผ๐ฟ๐ฝ๐ผ๐ฟ๐ฎ๐๐ถ๐ป๐ด ๐ฟ๐ฒ๐๐ฝ๐ผ๐ป๐๐ฒ ๐ฏ๐ถ๐ฎ๐
- ๐Use ๐ถ๐ฑ๐ฒ๐ป๐๐ถ๐ฐ๐ฎ๐น๐ถ๐๐ ๐ป๐ผ๐ฟ๐บ๐ (expected score distribution for identical samples)
- ๐ง Apply ๐ฝ๐๐๐ฐ๐ต๐ผ๐น๐ผ๐ด๐ถ๐ฐ๐ฎ๐น ๐บ๐ผ๐ฑ๐ฒ๐น๐ such as the Thurstonian 2-Alternative Choice
Statistics alone ๐๐ผ๐๐น๐ฑ๐ปโ๐ ๐ฏ๐ฒ ๐ฎ๐ฏ๐น๐ฒ ๐๐ผ ๐ต๐ฎ๐ป๐ฑ๐น๐ฒ this scenario, just as they canโt handle the ๐๐ป๐ฒ๐พ๐๐ฎ๐น ๐บ๐ฒ๐ป๐๐ฎ๐น ๐๐ฝ๐ฎ๐ฐ๐ถ๐ป๐ด of categories on rating scales or ๐ฒ๐ ๐ฝ๐น๐ฎ๐ถ๐ป ๐๐ต๐ฒ ๐ต๐ถ๐ด๐ต๐ฒ๐ฟ ๐ฝ๐ผ๐๐ฒ๐ฟ of the 3-AFC and tetrad methods over the triangle test.
๐ For a recent publication listing many relevant references (๐ฟ๐ฒ๐พ๐๐ฒ๐๐ ๐ผ๐ป ๐ผ๐๐ฟ ๐๐ฒ๐ฏ๐๐ถ๐๐ฒ): Biases in paired preference tests: Cross-cultural comparisons of Japanese and American consumers. 2019. Yuka Hasegawa ๐ฆ๐ต ๐ข๐ญ. Journal of Sensory Studies.
โข ๐ฏDo you have go-to approaches to handling ๐ฏ๐ถ๐ฎ๐๐ฒ๐ in ๐ฝ๐ฟ๐ฒ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ฐ๐ฒ ๐๐ฒ๐๐๐?
โข ๐กAlso, what ๐ผ๐๐ต๐ฒ๐ฟ ๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ of sensory science do you know can be ๐ฝ๐ฒ๐ฟ๐ฝ๐น๐ฒ๐ ๐ถ๐ป๐ด to students/professionals not -yet- versed in the intricacies of our field?