PDF pending restoration
ERIC Number: ED246407
Record Type: RIE
Publication Date: 1984-Mar
Reference Count: 0
On-Line Processing of Relative Importance Information during Reading.
Lorch, Elizabeth Pugzles; Lorch, Robert F., Jr.
One hundred thirty-nine college students participated in an experiment designed to examine the basis of their abilities to make relative importance distinctions among ideas in a text while reading. Subjects read two texts of 13 paragraphs each. The topics of the two texts were unrelated, but their topic structure was similar. For each of the critical paragraphs in each text, reading times were recorded for two target sentences: the initial topic sentence and a matched nontopic sentence expressing an idea subordinate to the topic sentence. To test the hypothesis that reading times may have been longer on topic sentences because they contained more new information, two versions of each text were constructed. In one version, the first paragraph of the text did not contain any information about the topics to be discussed. In the second version, the initial paragraph specified the topics of text and their organization. Two kinds of instructions were also given to readers. Half were given a sample outline of the first paragraph of the text before reading the body of the text; the other half read the text straight through, in preparation for a recognition test. The major results of the experiment were that subjects spent nearly half a second more time processing topic sentences than nontopic sentences, and that the magnitude of this "topic sentence effect" depended upon several other variables such as whether the topic sentence effect was greater when the topic sentence introduced a new attribute than when it did not. (HOD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: Text Structure; Topic Sentences
Note: Paper presented at the Annual Meeting of the Southeastern Psychological Association (New Orleans, LA, March 1984). Print is broken and may not reproduce well.