Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

[Phys-L] requiring /valid/ tests



As the proverb says:
Be careful what you test for; you might get it.

I'm all in favor of testing ... /if it is done right/.

I'm all in favor of accountability ... /if it is done right/.

========

In connection with the federally-mandated high-stakes
testing program, many states are currently shopping
for a new generation of tests. Some of this is driven
by Common Core and/or NGSS.

Here is a deceptively simple idea that you might
recommend to your state legislators and/or Supt.
of Public Instruction:

Anybody who wants to become a test supplier must
a) State in detail what the test is supposed to
measure, and
b) provide clear, objective evidence that the test
actually measures what it is supposed to.

In educationese, part (b) is called "validity". A test
is "valid" if it measures what it is supposed to. Of
course this is meaningless without part (a).

Conversely, part (a) is meaningless without part (b).

===

Validity seems like a reasonable requirement, indeed
a very low bar. Who could object to requiring a test
to be valid?

On the other hand ... this requirement is sorely needed.
Vendors have a long history of foisting grossly invalid
tests onto the states. On the other edge of the same
sword, bureaucrats have a long history of pretending
that tests measure things that they were never intended
to measure ... and then making life-changing decisions
about students, teachers, and entire schools, based
on invalid measurements.

If anybody at the Board of Education -- or at the
vendors -- dares object to the validity requirement,
it creates a tremendous campaign issue. Why should
we spend money on tests with no demonstrated validity?

============================

Suggested action item: Send a letter along these lines
to your state legislators and/or Supt. of Public Instruction.
Please let me know [off list] what response you get.