<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
    <title></title>
  </head>
  <body text="#000000" bgcolor="#ffffff">
    On 5/8/2011 3:28 PM, Emad Mohamed wrote:
    <blockquote
      cite="mid:BANLkTin0X-yiMz_+j6CdZJyS+ShvfoMY3A@mail.gmail.com"
      type="cite">
      <div dir="ltr"><font class="Apple-style-span" size="4"></font>
        <div>
          <div><font class="Apple-style-span" size="4">My question is:
              Do I need to use any significance test to show  that
              adding feature x really helps? if no, why not? and if yes,
              what are the appropriate tests in such a case?</font></div>
        </div>
      </div>
    </blockquote>
    <br>
        Significance tests are only relevant if you're testing a
    representative sample of the data.  If your sample is not
    representative, no amount of significance testing will make it so.<br>
    <pre class="moz-signature" cols="72">-- 
                                -Angus B. Grieve-Smith
                                Saint John's University
                                <a class="moz-txt-link-abbreviated" href="mailto:grvsmth@panix.com">grvsmth@panix.com</a>
</pre>
  </body>
</html>