A corpus is a collection of texts stored on a computer. This dictionary was created by a group of editors using sophisticated software tools to search the American English parts of the Cambridge International Corpus (CIC) for words and phrases and to analyze the results.

Cambridge University Press has assembled a corpus that includes 700 million words of written and spoken English texts. The texts in the CIC come from newspapers, novels, nonfiction books on a wide range of topics, websites, magazines, junk mail, TV and radio programs, recordings of people's everyday conversations, and many other sources. The American English parts of the CIC include 175 million words of written texts, 22 million words of spoken English, 7 million words of academic texts, and 25 million words taken from business texts.

Although idioms in general are used frequently in both speech and writing, individual idioms may not be used frequently. For example, the idiom give up the ghost occurs approximately once in every 7 million words of corpus texts. (By comparison, the word give occurs approximately 3,300 times in every 7 million words.) Because there are many more idioms in American English than we could include in this book, we decided to include only idioms that occurred at least once in every 10 million words of corpus texts.

The idioms included in the Cambridge Dictionary of American Idioms have been carefully checked in the American English parts of the CIC. Our editors have checked to make sure that each idiom is actually used in current American English, that it is used often enough to include in this book, and that the information we give about each idiom describes what people mean when they say or write it.