ABSTRACT
I had been in Cambridge for four years, the postcode project had finished and my PhD was completed. So what next? At that time, the same government department was responsible for both postal and telephone services and had a very active technical development programme. They were keen on improving the quality of telephone lines and for that purpose needed to measure their efficiency. This was, and I suspect still is, typically done through listening tests whereby potentially confusable items and phrases are spoken, and the quality of the line is measured in terms of the listener’s error rate. Might it be possible to obtain an even more sensitive measure, if, in addition to hearing the material, people had to use it? One can think of two possible reasons why this might be the case. It is possible that the effort of decoding a noisy signal might take attention away from other types of processing such as understanding the message. Another possibility is that decoding a noisy auditory signal itself might lead to a noisy memory representation, potentially leading to further error when retrieved.
