WHEN PEOPLE READ ORALLY, they sometimes appear to omit words that are in the text. This report presents evidence from oral reading miscue research to support a psycholinguistic view of why omissions take place and how they reflect the reading process. Word level omissions are classified as deliberate and non-deliberate. Deliberate omissions involve the readers choosing, after consideration, to omit a word rather than to substitute a real word, or a non-word, or to seek help. On the other hand non-deliberate omissions result from the operations of the reading process and are often indicators of readers' strengths in the use of that process. Readers are often unaware they have made non-deliberate omissions. Non-deliberate omissions are sub-divided as follows: 1) omissions incidental to a complex miscue on a phrase or clause level; 2) omissions of optional surface structure elements; 3) omissions of unpredictable elements; 4) omissions involving dialogue; 5) omissions of words in compounds; and 6) omissions involved in prediction of other structures. The evidence cited does not support old views that omissions result from inability to respond to an unknown word, lack of word attack skills, or carelessness. The report concludes that omissions are surface representations of an underlying interaction between the reader and the text.