ASA Connect

Hi all, 

     First, I'd like to briefly introduce myself... I'm the ASA's Science Policy Fellow, and my job is to promote the practice of statistics in public policy making decisions. I've just started at the ASA in the last month, but I've been enjoying following some of the discussions so far,  especially those that can be related back to policy. If anyone ever has any questions or topics that could be related, please feel free to let me know.

    Now on to my main question... I got a call today asking about standardized tests and how often they are or should be re-normed. I've found some evidence online, but the answers vary from every seven years to every twenty years, and all are from older or not official sources. Can anyone point me in the right direction? Answers from members who work regularly with standardized testing are much appreciated!

Hi Howard, 

     I don't actually know a lot of the details... This is a question coming from a quick phone call that I just happened to answer. Here's what I do know... This particular school board is advocating the Iowa Test of Basic Skills as THE assessment to use. The ITBS was last used renormed in 2005. In 2012, the test was renamed and rebranded into a different test (looks like something called the Iowa Assessments), but for whatever reason, the school board wants to continue using the ITBS. It sounds like the school superintendent was looking for expert opinion regarding using the latest or renormed version so that she could convince them otherwise.

     Hopefully this helps!

Hi Amy,

Irregardless of norming, I think the most important comment here is that it is illogical to advocate any one particular test as THE test.  To me, "the test was renamed and rebranded into a different test" speaks volumes.  Companies like Seton, ACT, etc. are in it primarily to make money.  I haven't looked, but I'm guessing that any information on just how the Iowa Tests are normed is likely proprietary.  And while there may be some underlying statistics involved in their methods, when you start in the middle of a big patch of quicksand (non interval-ratio data for sure), it's nearly impossible to find your footing on solid ground.  

Another point that might be made to this school board is that perhaps they are barking up the wrong tree when it comes to what is best for their students.  Rather than relying on the "classics" (old-school tests designed in some sense to compare one school to another, but in many senses built out of sand), perhaps they should consider newer individualized testing systems that are designed to be adaptive to the needs of an individual and not only test but provide students with interactive materials based on that testing that will help them to learn those things that they didn't already know (ALEKS is one example that comes to mind).  

Cheers,
Joe

Hi all, 

     First, I'd like to briefly introduce myself... I'm the ASA's Science Policy Fellow, and my job is to promote the practice of statistics in public policy making decisions. I've just started at the ASA in the last month, but I've been enjoying following some of the discussions so far,  especially those that can be related back to policy. If anyone ever has any questions or topics that could be related, please feel free to let me know.

    Now on to my main question... I got a call today asking about standardized tests and how often they are or should be re-normed. I've found some evidence online, but the answers vary from every seven years to every twenty years, and all are from older or not official sources. Can anyone point me in the right direction? Answers from members who work regularly with standardized testing are much appreciated!