A study assessed the reliability and some aspects related to the validity of an open sentence approach to assessing comprehension during silent reading. Subjects, 37 randomly selected seventh-grade students, read an "Opin" passage, prepared by deleting the second half of every other sentence. The blanks filled in by students were evaluated using a scale based on the response's resemblance to the author's intent. Responses were also divided into fragmented, projective, and interactive categories. Oral readings of a second, intact passage were used to determine the kinds of reading miscues made by students. The results confirm that the Opin sentence procedure has sufficient interrater reliability to be used for assessing a reader's ability to interact with text during silent reading. Both the percent score method and the category scoring method proved reliable for reading assessment purposes. The pattern of correlations with the miscue scores also indicates that readers who attempt to interact with the text while reading are more likely to make miscues that are consistent with their search for meaning. (Tables of data and a sample Opin passage are appended.) (HTH)