<?xml version="1.0" standalone="yes"?>
<Paper uid="W02-0405">
  <Title>Using Summaries in Document Retrieval</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
LEAD(BUSH AND GORE), or HEADLINE
</SectionTitle>
    <Paragraph position="0"> and LEAD combination, e.g., HLEAD(BUSH AND GORE). Through three separate experiments, the value of leading text as a general purpose summary for news documents has been verified. This paper describes a fourth experiment that investigates whether and how searches limited to this type of summary benefit the targeted customers In many information retrieval experiments, a single user perspective, i.e., a single answer key, is used to evaluate the results. If that perspective matches that of the targeted customer set, the evaluation is meaningful.</Paragraph>
    <Paragraph position="1"> However, different customer segments perform information seeking tasks with different goals and perspectives in mind, even when they are interested in the same topic. Just as potential search tool enhancements are not one-size-fitsall, a one-size-fits-all answer key should not be used to determine the value of a search aid for two sets of customers with fundamentally different goals. In this experiment, the results of each query were evaluated using two different user perspectives, highly relevant references only and all references. Through this approach we were able to determine whether Searchable LEAD satisfied the goal that motivated its creation.</Paragraph>
  </Section>
class="xml-element"></Paper>