Should You Always Use The Tetrad Test?

🌍 𝗧𝗵𝗼𝘂𝗴𝗵𝘁 𝘁𝗵𝗲 𝘁𝗲𝘁𝗿𝗮𝗱 𝘄𝗼𝘂𝗹𝗱 𝘀𝗮𝘃𝗲 𝘁𝗵𝗲 𝘄𝗼𝗿𝗹𝗱? 𝗧𝗵𝗶𝗻𝗸 𝗮𝗴𝗮𝗶𝗻.

Back in my grad school days with Prof. Michael O’Mahony, my early research focused on ever more powerful discrimination methods. The motto was simple: 𝘁𝗵𝗲 𝗺𝗼𝗿𝗲 𝗽𝗼𝘄𝗲𝗿𝗳𝘂𝗹, 𝘁𝗵𝗲 𝗯𝗲𝘁𝘁𝗲𝗿.

• The duo–trio and triangle methods lacked power

• The 2-AFC improved sensitivity, but required specifying the sensory dimension

• Then came the 𝘁𝗲𝘁𝗿𝗮𝗱, championed by my former colleagues Dr. John Ennis and Dr. Daniel Ennis at The Institute for Perception.

It seemed like the perfect solution — still lacking the power of the 2-AFC, but with a clear power advantage over the triangle and duo-trio tests, with no need to define an attribute.

Problem solved… or so many users thought.

In a recent webinar, Benoît Rousseau shared his results where:

• The tetrad 𝗱𝗶𝗱 𝗻𝗼𝘁 𝗱𝗲𝘁𝗲𝗰𝘁 a difference between two beverages (84 trials correct/228, 𝙥 > 𝟬.𝟬𝟱)

• Yet consumers 𝗽𝗿𝗲𝗳𝗲𝗿𝗿𝗲𝗱 one over the other (74/104, 𝙥 < 𝟬.𝟬𝟬𝟭) – You can find the research summary here: https://lnkd.in/gvuEBgmw

One attendee exclaimed, “𝘽𝙪𝙩 𝙄 𝙬𝙖𝙨 𝙩𝙤𝙡𝙙 𝙩𝙝𝙚 𝙩𝙚𝙩𝙧𝙖𝙙 𝙬𝙤𝙪𝙡𝙙𝙣’𝙩 𝙢𝙞𝙨𝙨 𝙖𝙣𝙮𝙩𝙝𝙞𝙣𝙜!”

The truth? Every pair of products differs in 𝘴𝘰𝘮𝘦 𝘸𝘢𝘺 — and as many of you know, a non-significant result simply means the test wasn’t powerful enough to detect that difference. As is often the case in sensory science, psychology (thank you, 𝗧𝗵𝘂𝗿𝘀𝘁𝗼𝗻𝗲 🧠) helps explain this apparent 𝘱𝘢𝘳𝘢𝘥𝘰𝘹.

The tetrad isn’t a silver bullet. To make it predictive, it needs to fit within a broader risk framework— one that clearly defines the target consumer difference threshold, alpha and beta levels, and sample size.

👉 And when product equivalence is the goal, a similarity test designed on the same principles is the better tool.