The Cocktail Party Effect

What it is

Colin Cherry's 1953 paper in the Journal of the Acoustical Society of America was about listening. He played two messages at once, one to each ear, and asked people to repeat one aloud. Listeners nearly always noticed when the voice in the ignored ear changed from a man's to a woman's, but could not report a word of what it said, and when the ignored message switched to German one listener assumed it had been English throughout. Cherry used the phrase "cocktail party problem".

The name came from Neville Moray in 1959. Almost nothing from the ignored ear got through, and his abstract ends: "The only stimulus so far found that will break through this barrier is the subject's own name." About a third of his participants heard theirs. Noelle Wood and Nelson Cowan repeated it in 1995 with 34 undergraduates and got 34.6 per cent, with attention disrupted for the two words after the name. A preregistered replication by Jan Philipp Röer and Cowan in 2021 found 29 per cent, and unexpected words were rarely noticed. In 2001 Andrew Conway, Cowan and Michael Bunting found that 65 per cent of people with low working memory capacity heard their name, against 20 per cent of those with high capacity.

Richard Shotton's The Choice Factory (Bias 24, pages 159 to 165) builds on this, telling the story as Cherry hearing his own name across a party; in the published record the own-name finding is Moray's. The chapter's advice is to localise a message rather than print a customer's name on it.

In effect

Your name buys a brief moment of attention from a minority of people. In a 2018 field experiment by Navdeep Sahni and colleagues, adding the recipient's name to an email subject line raised opens from 9.05 to 10.80 per cent and sales leads from 0.39 to 0.51 per cent, while unsubscribes fell from 1.2 to 1.0 per cent. Seeing your name is a signal that someone wants your attention, not that the message is relevant to you.

What it does not say

It does not say that everyone hears their name: in every shadowing study most people missed it. The capture is brief. Much of the commercial evidence rests on one paper covering three companies, and a name used everywhere may lose its pull, a worry no cited study has measured. Shotton's own survey of 304 people, in which 36 per cent called personalised banner ads completely unacceptable, and his localisation test with 500 people are unpublished.

---