Abstract
One of the distinguishing characteristics of corpus-based dictionaries is that most entries contain example sentences or phrases that have been copied or adapted from corpora. Although examples are generally regarded as positive and have high face validity among learners, the body of evidence about their actual benefits is limited and inconclusive. My aim in the present study is to revisit the idea of testing the usefulness of corpus examples. However, unlike previous research, in the present study different words are used to test language comprehension and the ability to correct typical language production errors, and a distinction is also drawn with regard to examples intended to facilitate decoding and examples meant to benefit encoding. In addition to this, because a single example might not be enough to help people understand what a word means or how it is used, in this study I also test the value of presenting learners with multiple corpus examples.