A study investigated the effect on reading comprehension of using semantic impressions as a previewing activity to provide an overall impression of a story's structure. Subjects, 31 eighth grade remedial reading students from a suburban junior high school, were divided into two groups: a treatment group that received semantic impressions as a prereading activity, and a control group. The semantic impressions--words extracted directly from a story--provided clues to the setting, characters, and major elements of the plot. A maximum of three words was used for each impression. The 19 clues were arranged vertically and marked with arrows to indicate clue order. After fifteen minutes of demonstration and instruction, students were given the impression and asked to link the clues together in the same manner and to generate a hypothesis story of their own. The next day, the students read their stories before being given a copy of the short story to read. After reading it, students from both groups were tested on it. Results indicated that students receiving the semantic impressions as a prereading treatment correctly answered a higher percentage (72%) of the comprehension questions than students who simply read the story (60%). A table of findings, examples of semantic impressions for two stories, and a student hypothesis story are appended. (EL)