Difference between revisions of "REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions (Summer 2007)"

From LearnLab
Jump to: navigation, search
 
(One intermediate revision by the same user not shown)
Line 1: Line 1:
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+
== REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions (Summer 2007) ==
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en" dir="ltr">
+
   
<head>
+
 
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+
=== Logistical Information ===
<meta name="keywords" content=" REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions (Summer 2007) " />
+
 
<link rel="shortcut icon" href="/favicon.ico" />
+
{| border="1"
<title> REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions (Summer 2007) - Pslc</title>
+
|+
<style type="text/css" media="screen,projection">/*<![CDATA[*/ @import "/research/wiki/skins/monobook/main.css?7"; /*]]>*/</style>
+
|-
<link rel="stylesheet" type="text/css" media="print" href="/research/wiki/skins/common/commonPrint.css" />
+
| '''Contributors''' || Christine M. Feeney and Michael Heilman 
<!--[if lt IE 5.5000]><style type="text/css">@import "/research/wiki/skins/monobook/IE50Fixes.css";</style><![endif]-->
+
|-
<!--[if IE 5.5000]><style type="text/css">@import "/research/wiki/skins/monobook/IE55Fixes.css";</style><![endif]-->
+
| '''Study Start Date''' || June, 2007 
<!--[if IE 6]><style type="text/css">@import "/research/wiki/skins/monobook/IE60Fixes.css";</style><![endif]-->
+
|-
<!--[if IE 7]><style type="text/css">@import "/research/wiki/skins/monobook/IE70Fixes.css?1";</style><![endif]-->
+
| '''Study End Date''' || July, 2007 
<!--[if lt IE 7]><script type="text/javascript" src="/research/wiki/skins/common/IEFixes.js"></script>
+
|-
<meta http-equiv="imagetoolbar" content="no" /><![endif]-->
+
| '''Learnlab Courses''' || no
<script type="text/javascript">var skin = 'monobook';var stylepath = '/research/wiki/skins';</script>
+
|-
<script type="text/javascript" src="/research/wiki/skins/common/wikibits.js"><!-- wikibits js --></script>
+
| '''Number of Students''' || 30
<script type="text/javascript" src="/research/wiki/index.php?title=-&amp;action=raw&amp;smaxage=0&amp;gen=js"><!-- site js --></script>
+
|-
<style type="text/css">/*<![CDATA[*/
+
| '''Total Participant Hours (est.)''' || 300
@import "/research/wiki/index.php?title=MediaWiki:Common.css&action=raw&ctype=text/css&smaxage=18000";
+
|-
@import "/research/wiki/index.php?title=MediaWiki:Monobook.css&action=raw&ctype=text/css&smaxage=18000";
+
| '''Data in Datashop''' || no  
@import "/research/wiki/index.php?title=-&action=raw&gen=css&maxage=18000";
+
|}
/*]]>*/</style>
+
 
<!-- Head Scripts -->
+
=== Abstract ===
</head>
+
 
<body  class="ns-0 ltr">
+
Previous psychological research has identified two types of comprehension: shallow, in which people can reproduce information, and deep, in which people can comprehend the meaning of information.  In addition, researchers have found that these two types of comprehension are separate and disparate.  The PSLC currently employs an English as a Second Language vocabulary tutor, REAP, that uses shallow, automatically generated, comprehension questions to check that students are actively reading, rather than just skimming, practice reading passages.  The purpose of the current study was to examine performance on the REAP-type questions with manually authored, deeper, reading comprehension questions.  Participants were thirty undergraduate students (male = 11) participating in summer research programs at Carnegie Mellon University.  The researcher predicted a positive correlation between the two testing types, which was supported by the data (r = .366, p < 0.0005, one-tailed t-test).   
<div id="globalWrapper">
+
 
<div id="column-content">
+
=== Glossary ===
<div id="content">
+
 
<a name="top" id="top"></a>
+
''Shallow Comprehension:'' Processing directed towards reproducing the learning material without necessarily understanding it.
<h1 class="firstHeading"> REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions (Summer 2007) </h1>
+
 
<div id="bodyContent">
+
''Deep Comprehension:'' Processing directed towards comprehending the intended meaning of the learning material.
<h3 id="siteSub">From Pslc</h3>
+
 
<div id="contentSub"></div>
+
=== Research question ===
<div id="jump-to-nav">Jump to: <a href="#column-one">navigation</a>, <a href="#searchInput">search</a></div> <!-- start content -->
+
 
<table id="toc" class="toc" summary="Contents"><tr><td><div id="toctitle"><h2>Contents</h2></div>
+
Does performance on the shallow, automatically-generated questions correlate with performance on more sophisticated, deep reading comprehension questions?
<ul>
+
 
<li class='toclevel-1'><a href="# Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions"><span class="tocnumber">1</span> <span class="toctext"> REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions </span></a>
+
=== Dependent variables ===
<ul>
+
 
<li class='toclevel-2'><a href="#Logistical_Information"><span class="tocnumber">1.1</span> <span class="toctext">Logistical Information</span></a></li>
+
Percentage of Shallow Comprehension Questions Answered Correctly  
<li class='toclevel-2'><a href="#Abstract"><span class="tocnumber">1.2</span> <span class="toctext">Abstract</span></a></li>
+
 
<li class='toclevel-2'><a href="#Glossary"><span class="tocnumber">1.3</span> <span class="toctext">Glossary</span></a></li>
+
Percentage of Deep Comprehension Questions Answered Correctly
<li class='toclevel-2'><a href="#Research_question"><span class="tocnumber">1.4</span> <span class="toctext">Research question</span></a></li>
+
 
<li class='toclevel-2'><a href="#Dependent_variables"><span class="tocnumber">1.5</span> <span class="toctext">Dependent variables</span></a></li>
+
=== Independent variables ===
<li class='toclevel-2'><a href="#Independent_variables"><span class="tocnumber">1.6</span> <span class="toctext">Independent variables</span></a></li>
+
 
<li class='toclevel-2'><a href="#Hypotheses"><span class="tocnumber">1.7</span> <span class="toctext">Hypotheses</span></a></li>
+
Five reading passages, each of which was followed by four shallow reading comprehension questions and three to five deep reading comprehension questions.
<li class='toclevel-2'><a href="#Explanation"><span class="tocnumber">1.8</span> <span class="toctext">Explanation</span></a></li>
+
 
<li class='toclevel-2'><a href="#Descendents"><span class="tocnumber">1.9</span> <span class="toctext">Descendents</span></a></li>
+
=== Hypothesis ===
<li class='toclevel-2'><a href="#Notes"><span class="tocnumber">1.10</span> <span class="toctext">Notes</span></a></li>
+
 
<li class='toclevel-2'><a href="#Annotated_bibliography"><span class="tocnumber">1.11</span> <span class="toctext">Annotated bibliography</span></a></li>
+
A significant positive correlation between shallow and deep reading comprehension questions will occur in this study.
</ul>
+
 
</li>
+
=== Explanation ===
</ul>
+
 
</td></tr></table><script type="text/javascript"> if (window.showTocToggle) { var tocShowText = "show"; var tocHideText = "hide"; showTocToggle(); } </script>
+
There was a relationship between performance on the two question types (automatically-generated shallow reading comprehension questions and manually-generated deep comprehension questions), which could imply that using the automatically-generated questions as a check for reading comprehension for the REAP program is a sufficient measure (r = .366, p < 0.0005, one-tailed test). Further studies could examine different types of deep reading comprehension questions, removing the “giveaway words” from the automatically-generated distracters, and examining whether ESL students approach these questions in a different manner than native English speakers.
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions%28Summer_2007%29&amp;action=edit&amp;section=1" title="Edit section: REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions ">edit</a>]</div><a name=" REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions "></a><h2> REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions </h2>
+
 
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%28Summer_2007%29&amp;action=edit&amp;section=2" title="Edit section: Logistical Information">edit</a>]</div><a name="Logistical_Information"></a><h3> Logistical Information </h3>
+
=== Descendents ===
<table border="1">
+
 
<caption>
+
=== Notes ===
</caption>
+
 
<tr>
+
The researchers chose five passages of between five hundred to one thousand words because this length is comparable to reading comprehension passages that appear on standardized tests, such as the GRE, and to REAP textsThe researchers used a variety of sources in order to have passages of varying difficulty.
<td> <b>Contributors</b> </td><td> Christine M. Feeney and Michael Heilman
+
 
</td></tr>
+
=== Annotated bibliography ===
<tr>
+
 
<td> <b>Study Start Date</b> </td><td> June, 2007
+
Curran, James R. Curran & Moens, Marc. (2002a). Improvements in automatic thesaurus extraction. Proceedings of the Workshop of the ACL Special Interest Group on the Lexicon (SIGLEX), 59-66.
</td></tr>
+
 
<tr>
+
Davoudi, Mohammad. (2005). Inference Generation Skill and Text Comprehension. The Reading Matrix, 5(1), 106-123.  
<td> <b>Study End Date</b> </td><td> July, 2007
+
 
</td></tr>
+
Heilman, M., Collins-Thompson, K., Callan, J. & Eskenazi, M. (2006). Classroom success of an Intelligent Tutoring System for lexical practice and reading comprehension. Proceedings of the Ninth International Conference on Spoken Language Processing.
<tr>
+
 
<td> <b>Learnlab Courses</b> </td><td>none
+
Heilman, M. & Eskenazi, M. In Press. Application of Automatic Thesaurus Extraction for Computer Generation of Vocabulary Questions. Proceedings of the SLaTE Workshop on Speech and Language Technology in Education.
</td></tr>
+
<tr>
+
<td> <b>Number of Students</b> </td><td> 30
+
</td></tr>
+
<tr>
+
<td> <b>Total Participant Hours (est.)</b> </td><td> 300
+
</td></tr>
+
<tr>
+
<td> <b>Data in Datashop</b> </td><td> no
+
</td></tr></table>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%28Summer_2007%29&amp;action=edit&amp;section=3" title="Edit section: Abstract">edit</a>]</div><a name="Abstract"></a><h3> Abstract </h3>
+
<p> Previous psychological research has identified two types of comprehension: shallow, in which people can reproduce information, and deep, in which people can comprehend the meaning of information.  In addition, researchers have found that these two types of comprehension are separate and disparate.  The PSLC currently employs an English as a Second Language vocabulary tutor, REAP, that uses shallow, automatically generated, comprehension questions to check that students are actively reading, rather than just skimming, practice reading passages.  The purpose of the current study was to examine performance on the REAP-type questions with manually authored, deeper, reading comprehension questions.  Participants were thirty undergraduate students (male = 11) participating in summer research programs at Carnegie Mellon University.  The researcher predicted a positive correlation between the two testing types, which was supported by the data (r = .366, p < 0.0005, one-tailed t-test).   
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%28Summer_2007%29&amp;action=edit&amp;section=4" title="Edit section: Glossary">edit</a>]</div><a name="Glossary"></a><h3> Glossary </h3>
+
<p><i>Shallow Comprehension:</i> Processing is directed towards reproducing the learning material without necessarily understanding it.
+
</p>
+
<p>
+
<i> Deep Comprehension:</i> Processing is directed towards comprehending the intended meaning of the learning material.
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%%28Summer_2007%29&amp;action=edit&amp;section=5" title="Edit section: Research question">edit</a>]</div><a name="Research_question"></a><h3> Research question </h3>
+
<p> Does performance on the shallow, automatically-generated questions correlate with performance on more sophisticated, deep reading comprehension questions?
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%28Summer_2007%29&amp;action=edit&amp;section=6" title="Edit section: Dependent variables">edit</a>]</div><a name="Dependent_variables"></a><h3> Dependent variables </h3>
+
<p> Percentage of Shallow Comprehension Questions Answered Correctly.
+
</p><p> Percentage of Deep Comprehension Questions Answered Correctly.
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%28Summer_2007%29&amp;action=edit&amp;section=7" title="Edit section: Independent variables">edit</a>]</div><a name="Independent_variables"></a><h3> Independent variable </h3>
+
<p>Five reading passages, each of which was followed by four shallow reading comprehension questions and three to five deep reading comprehension questions.
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%28Summer_2007%29&amp;action=edit&amp;section=8" title="Edit section: Hypotheses">edit</a>]</div><a name="Hypotheses"></a><h3> Hypotheses </h3>
+
<p> A significant positive correlation between shallow and deep reading comprehension questions will occur in this study.
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title= REAP_Study_on_the_Correlation_Between_Automatically_and_Manually_Generated_Reading_Comprehension_Questions_%28Summer_2007%29&amp;action=edit&amp;section=9" title="Edit section: Explanation">edit</a>]</div><a name="Explanation"></a><h3> Explanation </h3>
+
<p>The study found no significant differences between the two conditions (2 or 4 words/text).  This null result may be due to statistical power and mistakes in the study design.  The tutor had difficulty providing high quality readings with four vocabulary words.  Thus, the tutor often had to give texts with 3 or even 2 words to students in the 4-word condition.  Additionally, the optimal number of words per text (the density of practice) may be higher than 4 words.  There may be a threshold at which the density of practice becomes too high.  This threshold may actually be at 5 or even 10 words given the average length of the texts used by the REAP tutor.  Some research suggests that up to 5% of the words in a text can be unknown before comprehension is impeded (Laufer, 1992).  Although the task for that study was different (reading comprehension rather than vocabulary acquisition), it suggests that the optimal level of stretch may be such that as many as 20 words may be unknown in a text with 400 word types (which is approximately the mean number of word types in REAP texts). Other research, however, suggests higher thresholds up to 98% (Nation, 2001).
+
</p><p>Although the results for stretch from this study were inconclusive, a great deal of interesting data were gathered. Some interesting results have been found, and further analysis is possible.
+
</p><p>Many potential variables exist that could lead to successful learning of new vocabulary: general ESL ability as measured by the MTELP, time on task, number of words looked up, pre-test scores on the academic word list. A step-wise linear regression showed that the most significant predictor of acquisition of new vocabulary was the size of the student's vocabulary scores on the academic word list prior to instruction (r= 0.77, p &lt; .0001). However, the effect of this prior knowledge was diminished by the number of texts the students read r= -0.17). Together, in the stepwise regression, these two variable accounted for 67% of the variance (r2= .67).
+
</p><p>This finding corroborates findings of Stanovich (1986) and James (1996) who refer to the phenomenon as the "Matthew effect" (after a Bible passage), whereby the rich get richer and the poor get poorer. In this case, it is not 'general richness' expressed in terms of higher general ESL proficiency, but specifically knowledge of words on the academic word list. In other words, the more knowledge components in the domain of acquisition a learner brings to the task, the greater the potential for future learning.  
+
</p><p>It was also found that students frequently access dictionary definitions for non-target words that are well below their expected reading level (e.g., "hitting").  This may lead to revisions of curricula for ESL vocabulary learning.
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;action=edit&amp;section=10" title="Edit section: Descendents">edit</a>]</div><a name="Descendents"></a><h3> Descendents </h3>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;action=edit&amp;section=11" title="Edit section: Notes">edit</a>]</div><a name="Notes"></a><h3> Notes </h3>
+
<p>Target vocabulary lists were determined by a lengthy pre-test using multiple choice cloze questions.  Each question takes 30-60 seconds on average, depending on the student.  In later studies, self-assessment pre-tests were preferred due to time constraints and initial impressions of the studentsLengthy pre-tests for determining target word lists leave a poor initial impression of the tutor.
+
</p>
+
<div class="editsection" style="float:right;margin-left:5px;">[<a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;action=edit&amp;section=12" title="Edit section: Annotated bibliography">edit</a>]</div><a name="Annotated_bibliography"></a><h3> Annotated bibliography </h3>
+
<p>Laufer, B. (1992). How much lexis is necessary for reading comprehension? In Vocabulary and Applied Linguistics, Pierre J. L. Arnaud and Henri Béjoint (eds.), 126–132. London: Macmillan.
+
</p>
+
<!-- Saved in parser cache with key theory_wiki:pcache:idhash:1510-0!1!0!0!!en!2 and timestamp 20070725163805 -->
+
<div class="printfooter">
+
Retrieved from "<a href="http://www.learnlab.org/research/wiki/index.php/REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29">http://www.learnlab.org/research/wiki/index.php/REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29</a>"</div>
+
<!-- end content -->
+
<div class="visualClear"></div>
+
</div>
+
</div>
+
</div>
+
<div id="column-one">
+
<div id="p-cactions" class="portlet">
+
<h5>Views</h5>
+
<ul>
+
<li id="ca-nstab-main" class="selected"><a href="/research/wiki/index.php/REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29">Article</a></li>
+
<li id="ca-talk" class="new"><a href="/research/wiki/index.php?title=Talk:REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;action=edit">Discussion</a></li>
+
<li id="ca-edit"><a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;action=edit">Edit</a></li>
+
<li id="ca-history"><a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;action=history">History</a></li>
+
<li id="ca-move"><a href="/research/wiki/index.php/Special:Movepage/REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29">Move</a></li>
+
<li id="ca-watch"><a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;action=watch">Watch</a></li>
+
</ul>
+
</div>
+
<div class="portlet" id="p-personal">
+
<h5>Personal tools</h5>
+
<div class="pBody">
+
<ul>
+
<li id="pt-userpage"><a href="/research/wiki/index.php/User:Feeney" class="new">Feeney</a></li>
+
<li id="pt-mytalk"><a href="/research/wiki/index.php/User_talk:Feeney" class="new">My talk</a></li>
+
<li id="pt-preferences"><a href="/research/wiki/index.php/Special:Preferences">Preferences</a></li>
+
<li id="pt-watchlist"><a href="/research/wiki/index.php/Special:Watchlist">My watchlist</a></li>
+
<li id="pt-mycontris"><a href="/research/wiki/index.php/Special:Contributions/Feeney">My contributions</a></li>
+
<li id="pt-logout"><a href="/research/wiki/index.php?title=Special:Userlogout&amp;returnto=REAP_Study_on_Vocabulary_Stretch_(Spring_2006)">Log out</a></li>
+
</ul>
+
</div>
+
</div>
+
<div class="portlet" id="p-logo">
+
<a style="background-image: url(/research/wiki/skins/pslc_logo.png);" href="/research/wiki/index.php/Main_Page" title="Main Page"></a>
+
</div>
+
<script type="text/javascript"> if (window.isMSIE55) fixalpha(); </script>
+
<div class='portlet' id='p-navigation'>
+
<h5>Navigation</h5>
+
<div class='pBody'>
+
<ul>
+
<li id="n-mainpage"><a href="/research/wiki/index.php/Main_Page">Main Page</a></li>
+
<li id="n-Root-Node"><a href="/research/wiki/index.php/Root_node">Root Node</a></li>
+
<li id="n-Glossary"><a href="/research/wiki/index.php/Category:Glossary">Glossary</a></li>
+
<li id="n-recentchanges"><a href="/research/wiki/index.php/Special:Recentchanges">Recent changes</a></li>
+
<li id="n-randompage"><a href="/research/wiki/index.php/Special:Random">Random page</a></li>
+
<li id="n-help"><a href="/research/wiki/index.php/Help:Contents">Help</a></li>
+
</ul>
+
</div>
+
</div>
+
<div class='portlet' id='p-clusters'>
+
<h5>clusters</h5>
+
<div class='pBody'>
+
<ul>
+
<li id="n-Coordinative-Learning"><a href="/research/wiki/index.php/Coordinative_Learning">Coordinative Learning</a></li>
+
<li id="n-Interactive-Communication"><a href="/research/wiki/index.php/Interactive_Communication">Interactive Communication</a></li>
+
<li id="n-Refinement-and-Fluency"><a href="/research/wiki/index.php/Refinement_and_Fluency">Refinement and Fluency</a></li>
+
</ul>
+
</div>
+
</div>
+
<div class='portlet' id='p-courses'>
+
<h5>courses</h5>
+
<div class='pBody'>
+
<ul>
+
<li id="n-Algebra"><a href="/research/wiki/index.php/Algebra">Algebra</a></li>
+
<li id="n-Geometry"><a href="/research/wiki/index.php/Geometry">Geometry</a></li>
+
<li id="n-Chinese"><a href="/research/wiki/index.php/Chinese">Chinese</a></li>
+
<li id="n-French"><a href="/research/wiki/index.php/French">French</a></li>
+
<li id="n-English"><a href="/research/wiki/index.php/English">English</a></li>
+
<li id="n-Chemistry"><a href="/research/wiki/index.php/Chemistry">Chemistry</a></li>
+
<li id="n-Physics"><a href="/research/wiki/index.php/Physics">Physics</a></li>
+
</ul>
+
</div>
+
</div>
+
<div class='portlet' id='p-datashop'>
+
<h5>datashop</h5>
+
<div class='pBody'>
+
<ul>
+
<li id="n-DataShop-Pipeline"><a href="/research/wiki/index.php/DataShop_Pipeline">DataShop Pipeline</a></li>
+
</ul>
+
</div>
+
</div>
+
<div id="p-search" class="portlet">
+
<h5><label for="searchInput">Search</label></h5>
+
<div id="searchBody" class="pBody">
+
<form action="/research/wiki/index.php/Special:Search" id="searchform"><div>
+
<input id="searchInput" name="search" type="text" accesskey="f" value="" />
+
<input type='submit' name="go" class="searchButton" id="searchGoButton" value="Go" />&nbsp;
+
<input type='submit' name="fulltext" class="searchButton" value="Search" />
+
</div></form>
+
</div>
+
</div>
+
<div class="portlet" id="p-tb">
+
<h5>Toolbox</h5>
+
<div class="pBody">
+
<ul>
+
<li id="t-whatlinkshere"><a href="/research/wiki/index.php/Special:Whatlinkshere/REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29">What links here</a></li>
+
<li id="t-recentchangeslinked"><a href="/research/wiki/index.php/Special:Recentchangeslinked/REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29">Related changes</a></li>
+
<li id="t-upload"><a href="/research/wiki/index.php/Special:Upload">Upload file</a></li>
+
<li id="t-specialpages"><a href="/research/wiki/index.php/Special:Specialpages">Special pages</a></li>
+
<li id="t-print"><a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;printable=yes">Printable version</a></li> <li id="t-permalink"><a href="/research/wiki/index.php?title=REAP_Study_on_Vocabulary_Stretch_%28Spring_2006%29&amp;oldid=4799">Permanent link</a></li> </ul>
+
</div>
+
</div>
+
</div><!-- end of the left (by default at least) column -->
+
<div class="visualClear"></div>
+
<div id="footer">
+
<div id="f-poweredbyico"><a href="http://www.mediawiki.org/"><img src="/research/wiki/skins/common/images/poweredby_mediawiki_88x31.png" alt="MediaWiki" /></a></div>
+
<ul id="f-list">
+
<li id="lastmod"> This page was last modified 15:26, 14 April 2007.</li>
+
<li id="viewcount">This page has been accessed 17 times.</li>
+
<li id="privacy"><a href="/research/wiki/index.php/Pslc:Privacy_policy" title="Pslc:Privacy policy">Privacy policy</a></li>
+
<li id="about"><a href="/research/wiki/index.php/Pslc:About" title="Pslc:About">About Pslc</a></li>
+
<li id="disclaimer"><a href="/research/wiki/index.php/Pslc:General_disclaimer" title="Pslc:General disclaimer">Disclaimers</a></li>
+
</ul>
+
</div>
+
<script type="text/javascript"> if (window.runOnloadHook) runOnloadHook();</script>
+
</div>
+
<!-- Served by einstein.pslc.cs.cmu.edu in 0.16 secs. -->
+
</body></html>
+

Latest revision as of 18:36, 25 July 2007

REAP Study on the Correlation Between Automatically and Manually Generated Reading Comprehension Questions (Summer 2007)

Logistical Information

Contributors Christine M. Feeney and Michael Heilman
Study Start Date June, 2007
Study End Date July, 2007
Learnlab Courses no
Number of Students 30
Total Participant Hours (est.) 300
Data in Datashop no

Abstract

Previous psychological research has identified two types of comprehension: shallow, in which people can reproduce information, and deep, in which people can comprehend the meaning of information. In addition, researchers have found that these two types of comprehension are separate and disparate. The PSLC currently employs an English as a Second Language vocabulary tutor, REAP, that uses shallow, automatically generated, comprehension questions to check that students are actively reading, rather than just skimming, practice reading passages. The purpose of the current study was to examine performance on the REAP-type questions with manually authored, deeper, reading comprehension questions. Participants were thirty undergraduate students (male = 11) participating in summer research programs at Carnegie Mellon University. The researcher predicted a positive correlation between the two testing types, which was supported by the data (r = .366, p < 0.0005, one-tailed t-test).

Glossary

Shallow Comprehension: Processing directed towards reproducing the learning material without necessarily understanding it.

Deep Comprehension: Processing directed towards comprehending the intended meaning of the learning material.

Research question

Does performance on the shallow, automatically-generated questions correlate with performance on more sophisticated, deep reading comprehension questions?

Dependent variables

Percentage of Shallow Comprehension Questions Answered Correctly

Percentage of Deep Comprehension Questions Answered Correctly

Independent variables

Five reading passages, each of which was followed by four shallow reading comprehension questions and three to five deep reading comprehension questions.

Hypothesis

A significant positive correlation between shallow and deep reading comprehension questions will occur in this study.

Explanation

There was a relationship between performance on the two question types (automatically-generated shallow reading comprehension questions and manually-generated deep comprehension questions), which could imply that using the automatically-generated questions as a check for reading comprehension for the REAP program is a sufficient measure (r = .366, p < 0.0005, one-tailed test). Further studies could examine different types of deep reading comprehension questions, removing the “giveaway words” from the automatically-generated distracters, and examining whether ESL students approach these questions in a different manner than native English speakers.

Descendents

Notes

The researchers chose five passages of between five hundred to one thousand words because this length is comparable to reading comprehension passages that appear on standardized tests, such as the GRE, and to REAP texts. The researchers used a variety of sources in order to have passages of varying difficulty.

Annotated bibliography

Curran, James R. Curran & Moens, Marc. (2002a). Improvements in automatic thesaurus extraction. Proceedings of the Workshop of the ACL Special Interest Group on the Lexicon (SIGLEX), 59-66.

Davoudi, Mohammad. (2005). Inference Generation Skill and Text Comprehension. The Reading Matrix, 5(1), 106-123.

Heilman, M., Collins-Thompson, K., Callan, J. & Eskenazi, M. (2006). Classroom success of an Intelligent Tutoring System for lexical practice and reading comprehension. Proceedings of the Ninth International Conference on Spoken Language Processing.

Heilman, M. & Eskenazi, M. In Press. Application of Automatic Thesaurus Extraction for Computer Generation of Vocabulary Questions. Proceedings of the SLaTE Workshop on Speech and Language Technology in Education.