SAT verbal scores, as used in this study, may not effectively identify poor and good readers. A comprehension test, such as the Nelson-Denny Reading Test, with subjects scoring in the lower 25th percentile considered poor readers and those scoring in the highest 25th percentile considered good readers, may be a more reliable measure of reading ability. Given this limitation, however, since previous studies show that more skilled readers are better able to use text structure to distinguish between important and unimportant information, we expected good readers to agree with expert readers about main idea paragraphs. Yet findings of this study indicate no clear pattern in ratings of text interest and importance by good and poor readers. This study also investigated whether ability would affect agreement with expert readers for paragraphs with "seductive details." In fact, there was no consistency in agreement of ratings for all paragraphs. This was true for percentage of agreement of ratings and for comparison of mean ratings by expert readers to poor and good readers.

When subjects are not grouped by ability and mean ratings by all readers are considered there is a clear pattern of agreement with expert readers for paragraphs 2, 3, and 4. College readers rate these main idea paragraphs as moderate in interest and high in importance, which is consistent with ratings by expert readers. This indicates students' ability to recognize main ideas. As with previous studies (Hidi & Baird, 1986; 1988), ratings by college readers diverged on main idea paragraphs showing little relationship between importance and interest. In other words, college readers in this study found important information to be of little interest.

Likewise, there is a clear pattern in ratings by college readers on paragraphs 5, 6, and 7. However, contrary to previous studies (Hidi et al, 1982; Jetton & Alexander, 1997), these ratings converged on paragraphs with "seductive details" showing a strong relationship between importance and interest. Typically readers recognize this type of material in a mixed text as high or moderate in interest but low in importance. College readers in this study consistently rated seductive detail paragraphs as moderate in interest and moderate in importance. The convergence of ratings may indicate that when seductive details are included in text they interfere with the readers' ability to distinguish important information from less important details.