A "Paradox" in Confidence Interval Construction Using Sufficient Statistics

Statistical inference about parameters should depend on raw data only through sufficient statistics-the well known sufficiency principle. In particular, inference should depend on minimal sufficient statistics if these are simpler than the raw data. In this article, we construct one-sided confidence...

Full description

Saved in:
Bibliographic Details
Published in:The American statistician 2018-10, Vol.72 (4), p.315-320
Main Author: Wang, Weizhen
Format: Article
Language:English
Subjects:
Admissible confidence interval
Binary data
Binomial distribution
Confidence intervals
Nonrandomized inference
One-sided confidence interval
Order
Random variables
Regression analysis
Statistical analysis
Statistical inference
Statistical methods
Statistics
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by cdi_FETCH-LOGICAL-c360t-1d831d7a1dcc01990f7b1a56bbbfee0b1d1e4feb21213671a90c0b0e17bba1dc3
cites cdi_FETCH-LOGICAL-c360t-1d831d7a1dcc01990f7b1a56bbbfee0b1d1e4feb21213671a90c0b0e17bba1dc3
container_end_page 320
container_issue 4
container_start_page 315
container_title The American statistician
container_volume 72
creator Wang, Weizhen
description Statistical inference about parameters should depend on raw data only through sufficient statistics-the well known sufficiency principle. In particular, inference should depend on minimal sufficient statistics if these are simpler than the raw data. In this article, we construct one-sided confidence intervals for a proportion which: (i) depend on the raw binary data, and (ii) are uniformly shorter than the smallest intervals based on the binomial random variable-a minimal sufficient statistic. In practice, randomized confidence intervals are seldom used. The proposed intervals violate the aforementioned principle if the search of optimal intervals is restricted within the class of nonrandomized confidence intervals. Similar results occur for other discrete distributions.
doi_str_mv 10.1080/00031305.2017.1305292
format article
fullrecord <record><control><sourceid>jstor_proqu</sourceid><recordid>TN_cdi_jstor_primary_45116979</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><jstor_id>45116979</jstor_id><sourcerecordid>45116979</sourcerecordid><originalsourceid>FETCH-LOGICAL-c360t-1d831d7a1dcc01990f7b1a56bbbfee0b1d1e4feb21213671a90c0b0e17bba1dc3</originalsourceid><addsrcrecordid>eNp9kFtLAzEQhYMoWC8_obDU560zm91N981SvGFBofY5JNlEUtqkJqnaf-8uVR99mss53wwcQoYIY4QJXAMARQrVuABk474rmuKIDLCiLC8YxWMy6D15L52SsxhX3QisLgbkaZqNXkQQrf8aZdZlM--MbbVTOnt0SYcPse53MYWdSta7bBmte8sWO2OsstqlbJFEsjFZFS_IiRHrqC9_6jlZ3t2-zh7y-fP942w6zxWtIeXYTii2TGCrFGDTgGESRVVLKY3WILFFXRotCyyQ1gxFAwokaGRS9hA9J1eHu9vg33c6Jr7yu-C6l7yoypKWdUWrzlUdXCr4GIM2fBvsRoQ9R-B9bvw3N97nxn9y67jhgVvF5MMfVFaIdcOaTr856NYZHzbi04d1y5PYr30wQThlI6f_v_gG5FF9fA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2544346535</pqid></control><display><type>article</type><title>A "Paradox" in Confidence Interval Construction Using Sufficient Statistics</title><source>JSTOR Archival Journals and Primary Sources Collection</source><source>Taylor and Francis:Jisc Collections:Taylor and Francis Read and Publish Agreement 2024-2025:Science and Technology Collection (Reading list)</source><creator>Wang, Weizhen</creator><creatorcontrib>Wang, Weizhen</creatorcontrib><description>Statistical inference about parameters should depend on raw data only through sufficient statistics-the well known sufficiency principle. In particular, inference should depend on minimal sufficient statistics if these are simpler than the raw data. In this article, we construct one-sided confidence intervals for a proportion which: (i) depend on the raw binary data, and (ii) are uniformly shorter than the smallest intervals based on the binomial random variable-a minimal sufficient statistic. In practice, randomized confidence intervals are seldom used. The proposed intervals violate the aforementioned principle if the search of optimal intervals is restricted within the class of nonrandomized confidence intervals. Similar results occur for other discrete distributions.</description><identifier>ISSN: 0003-1305</identifier><identifier>EISSN: 1537-2731</identifier><identifier>DOI: 10.1080/00031305.2017.1305292</identifier><language>eng</language><publisher>Alexandria: Taylor &amp; Francis</publisher><subject>Admissible confidence interval ; Binary data ; Binomial distribution ; Confidence intervals ; Nonrandomized inference ; One-sided confidence interval ; Order ; Random variables ; Regression analysis ; Statistical analysis ; Statistical inference ; Statistical methods ; Statistics</subject><ispartof>The American statistician, 2018-10, Vol.72 (4), p.315-320</ispartof><rights>2018 American Statistical Association 2018</rights><rights>Copyright 2018 American Statistical Association</rights><rights>2018 American Statistical Association</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c360t-1d831d7a1dcc01990f7b1a56bbbfee0b1d1e4feb21213671a90c0b0e17bba1dc3</citedby><cites>FETCH-LOGICAL-c360t-1d831d7a1dcc01990f7b1a56bbbfee0b1d1e4feb21213671a90c0b0e17bba1dc3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://www.jstor.org/stable/pdf/45116979$$EPDF$$P50$$Gjstor$$H</linktopdf><linktohtml>$$Uhttps://www.jstor.org/stable/45116979$$EHTML$$P50$$Gjstor$$H</linktohtml><link.rule.ids>314,776,780,27901,27902,58213,58446</link.rule.ids></links><search><creatorcontrib>Wang, Weizhen</creatorcontrib><title>A "Paradox" in Confidence Interval Construction Using Sufficient Statistics</title><title>The American statistician</title><description>Statistical inference about parameters should depend on raw data only through sufficient statistics-the well known sufficiency principle. In particular, inference should depend on minimal sufficient statistics if these are simpler than the raw data. In this article, we construct one-sided confidence intervals for a proportion which: (i) depend on the raw binary data, and (ii) are uniformly shorter than the smallest intervals based on the binomial random variable-a minimal sufficient statistic. In practice, randomized confidence intervals are seldom used. The proposed intervals violate the aforementioned principle if the search of optimal intervals is restricted within the class of nonrandomized confidence intervals. Similar results occur for other discrete distributions.</description><subject>Admissible confidence interval</subject><subject>Binary data</subject><subject>Binomial distribution</subject><subject>Confidence intervals</subject><subject>Nonrandomized inference</subject><subject>One-sided confidence interval</subject><subject>Order</subject><subject>Random variables</subject><subject>Regression analysis</subject><subject>Statistical analysis</subject><subject>Statistical inference</subject><subject>Statistical methods</subject><subject>Statistics</subject><issn>0003-1305</issn><issn>1537-2731</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2018</creationdate><recordtype>article</recordtype><recordid>eNp9kFtLAzEQhYMoWC8_obDU560zm91N981SvGFBofY5JNlEUtqkJqnaf-8uVR99mss53wwcQoYIY4QJXAMARQrVuABk474rmuKIDLCiLC8YxWMy6D15L52SsxhX3QisLgbkaZqNXkQQrf8aZdZlM--MbbVTOnt0SYcPse53MYWdSta7bBmte8sWO2OsstqlbJFEsjFZFS_IiRHrqC9_6jlZ3t2-zh7y-fP942w6zxWtIeXYTii2TGCrFGDTgGESRVVLKY3WILFFXRotCyyQ1gxFAwokaGRS9hA9J1eHu9vg33c6Jr7yu-C6l7yoypKWdUWrzlUdXCr4GIM2fBvsRoQ9R-B9bvw3N97nxn9y67jhgVvF5MMfVFaIdcOaTr856NYZHzbi04d1y5PYr30wQThlI6f_v_gG5FF9fA</recordid><startdate>20181002</startdate><enddate>20181002</enddate><creator>Wang, Weizhen</creator><general>Taylor &amp; Francis</general><general>American Statistical Association</general><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>20181002</creationdate><title>A "Paradox" in Confidence Interval Construction Using Sufficient Statistics</title><author>Wang, Weizhen</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c360t-1d831d7a1dcc01990f7b1a56bbbfee0b1d1e4feb21213671a90c0b0e17bba1dc3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2018</creationdate><topic>Admissible confidence interval</topic><topic>Binary data</topic><topic>Binomial distribution</topic><topic>Confidence intervals</topic><topic>Nonrandomized inference</topic><topic>One-sided confidence interval</topic><topic>Order</topic><topic>Random variables</topic><topic>Regression analysis</topic><topic>Statistical analysis</topic><topic>Statistical inference</topic><topic>Statistical methods</topic><topic>Statistics</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Wang, Weizhen</creatorcontrib><collection>CrossRef</collection><jtitle>The American statistician</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Wang, Weizhen</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>A "Paradox" in Confidence Interval Construction Using Sufficient Statistics</atitle><jtitle>The American statistician</jtitle><date>2018-10-02</date><risdate>2018</risdate><volume>72</volume><issue>4</issue><spage>315</spage><epage>320</epage><pages>315-320</pages><issn>0003-1305</issn><eissn>1537-2731</eissn><abstract>Statistical inference about parameters should depend on raw data only through sufficient statistics-the well known sufficiency principle. In particular, inference should depend on minimal sufficient statistics if these are simpler than the raw data. In this article, we construct one-sided confidence intervals for a proportion which: (i) depend on the raw binary data, and (ii) are uniformly shorter than the smallest intervals based on the binomial random variable-a minimal sufficient statistic. In practice, randomized confidence intervals are seldom used. The proposed intervals violate the aforementioned principle if the search of optimal intervals is restricted within the class of nonrandomized confidence intervals. Similar results occur for other discrete distributions.</abstract><cop>Alexandria</cop><pub>Taylor &amp; Francis</pub><doi>10.1080/00031305.2017.1305292</doi><tpages>6</tpages></addata></record>
fulltext fulltext
identifier ISSN: 0003-1305
ispartof The American statistician, 2018-10, Vol.72 (4), p.315-320
issn 0003-1305
1537-2731
language eng
recordid cdi_jstor_primary_45116979
source JSTOR Archival Journals and Primary Sources Collection; Taylor and Francis:Jisc Collections:Taylor and Francis Read and Publish Agreement 2024-2025:Science and Technology Collection (Reading list)
subjects Admissible confidence interval
Binary data
Binomial distribution
Confidence intervals
Nonrandomized inference
One-sided confidence interval
Order
Random variables
Regression analysis
Statistical analysis
Statistical inference
Statistical methods
Statistics
title A "Paradox" in Confidence Interval Construction Using Sufficient Statistics
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-12T21%3A01%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-jstor_proqu&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=A%20%22Paradox%22%20in%20Confidence%20Interval%20Construction%20Using%20Sufficient%20Statistics&rft.jtitle=The%20American%20statistician&rft.au=Wang,%20Weizhen&rft.date=2018-10-02&rft.volume=72&rft.issue=4&rft.spage=315&rft.epage=320&rft.pages=315-320&rft.issn=0003-1305&rft.eissn=1537-2731&rft_id=info:doi/10.1080/00031305.2017.1305292&rft_dat=%3Cjstor_proqu%3E45116979%3C/jstor_proqu%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c360t-1d831d7a1dcc01990f7b1a56bbbfee0b1d1e4feb21213671a90c0b0e17bba1dc3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2544346535&rft_id=info:pmid/&rft_jstor_id=45116979&rfr_iscdi=true