<div style="line-height:1.7;color:#000000;font-size:14px;font-family:Arial"><font face="宋体" size="3">

</font><font face="宋体" size="3">


<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线">Dear NCL experts,</font></span></p>

<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线"> </font></span></p>

<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线">There are several bootstrap approaches to significance
testing:</font></span></p>

<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线">1.Permutation tests (without replacement)</font></span></p>

<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线">2.Bootstrap confidence intervals (with
replacement, including bootstrap t confidence interval, bootstrap percentile
confidence interval, bootstrap bias-corrected accelerated (BCa) interval and
bootstrap tilting interval etc).</font></span></p>

<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线">According to the description of the
bootstrap function “</font><a href="https://en.wikipedia.org/wiki/Bootstrapping_%28statistics%29"><span style="color: windowtext; text-decoration: none; text-underline: none;"><font face="等线">Bootstrapping</font></span></a><font face="等线">
is a statistical method that uses data resampling with replacement”, it seems
that NCL chooses the bootstrap confidence intervals method. Could anyone tell
me which one of the several bootstrap confidence intervals is used to get the confidence
intervals, BCa or tilting, or other methods?</font></span></p>

<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线"> </font></span></p>

<p style="margin: 0cm 0cm 0pt;"><span lang="EN-US"><font face="等线">Sincerely,<br>
Shouli Xuan<br>
Email:</font><a href="mailto:15295538792@163.com" data-auto-link="1"><font color="#0000ff" face="等线">15295538792@163.com</font></a></span></p>

</font></div><br><br><span title="neteasefooter"><p><br/> </p></span>