Re: More intersting revelations from EdTrust

[William Cala <Wcala@SERVTECH.COM>]

Allen,

Just can't figure out what's wrong with this.  I have Acrobat.  I open pdf all the time.  I got an "access denied" on this link as well.
Bill
  ----- Original Message ----- 
  From: Allen Flanigan. 
  To: ARN-L@LISTS.CUA.EDU 
  Sent: Thursday, January 03, 2002 11:43 AM
  Subject: Re: More intersting revelations from EdTrust


  Bill,

  The link I gave is to a PDF file, so if you don't have acrobat reader "hooked up" to your browser (so that it automatically opens pdf file addresses in acrobat reader) you might not be able to go directly to that address.

  Try going to this link:

  http://64.224.125.0/dtm/


  and clicking on "instructions", then on "important cautions".  You will still need acrobat reader to open the file.

  You can also see a similar disclaimer under the "read me" link.

  [Flanigan, Allen]  -----Original Message-----
  From: Wcala@SERVTECH.COM [mailto:Wcala@SERVTECH.COM]
  Sent: Thursday, January 03, 2002 7:04 AM
  To: ARN-L@listsrva.CUA.EDU
  Subject: Re: More intersting revelations from EdTrust


    Allen,

    The URL you listed doesn't work.  The message I got was "Unresolved IP address".  I'd like to find the statement you cited as it came through somewhat messed up.

    thanks

    Bill
      ----- Original Message ----- 
      From: Allen Flanigan. 
      To: ARN-L@LISTS.CUA.EDU 
      Sent: Wednesday, January 02, 2002 12:30 PM
      Subject: More intersting revelations from EdTrust


      Sorry for multiple postings, but the more I weed through this "dispelling the myth" report, the more interesting stuff I find.

      We have had a few discussions about the alleged differences between norm-referenced and criterion or standards-referenced tests (most recently George C. and I traded messages about CRTs, NRTs, Standards-referenced Tests, etc.)

      My suspicion that State tests commonly referred to as "criterion-referenced", or Standards-referenced, are little different from off-the shelf NRTs in their construction and in the items used therein is confirmed by this statement from EdTrust's website:

      http://64.224.125.0/dtm/popups/tutorial.dtm.pdf

      Criterion-Referenced and Norm-Referenced Tests

      Much confusion surrounds the use of these terms. In general, a norm-referenced

      test compares a studentTs performance on the exam with the performance of

      other students on the same test. Criterion-referenced tests and standards-based

      tests measure a studentTs performance against external criteria for evaluating the

      extent to which a student has mastered the material covered on the test. There

      are two main ways these differences come into play.

      First, many experts believe that test questions are chosen differently depending

      on which kind of test is being created. Since the goal of norm-referenced tests is

      essentially competitive, questions that all or most students answer correctly are

      weeded out in favor of questions that only some students answer correctly. In

      contrast, experts constructing criterion-referenced tests are supposed to focus

      primarily on whether the test question is aligned with the knowledge and skills the

      test is intended to measure. If all students answer a question correctly, it should

      mean that teachers are doing a good job teaching that particular part of the

      curriculum.

      Second, student performance on different kinds of tests is often reported using

      different kinds of scores. For example, while the results of norm-referenced tests

      often are reported in the form of ifpercentile rank,le the results of criterion-

      referenced tests often are reported using some variation on inpercent passinglt (or

      iaproficientle or irmeeting standardle). The percentile score tells you where a

      schoolTs average student performs compared with a nationally representative

      sample of students, whereas percent proficient tells you what percent of a

      schoolTs students met the stateTs benchmark for mastering the material covered

      on the test.

      However, in practice these distinctions blur. Many companies that traditionally

      have sold generic norm-referenced tests now help states build iucustomizedle tests

      that are meant to measure whether students are meeting the stateTs own unique

      academic standards. Often, however, the companies create such tests using the

      same set of questions (or iobank of test itemslo) that they also use to construct

      their more generic norm-referenced exams, with the result that such customized

      tests can end up looking much like the companyTs iooff the shelfly norm-referenced

      tests! Also, more and more, a single test can be used to yield widely different

      kinds of scores, from simple scale scores to norm-referenced percentile ranks to

      criterion-referenced percentages of students meeting state standards.