Alicia has taught students of all ages and has a master's degree in Education
A Test Without Standards
Your teacher hands out a test on simple addition. You're ready to ace this thing. After all, you have been adding since you were a baby. Mrs. Krause sets the paper in front of you, and you see that the entire test consists of word problems written in a foreign language you have never seen. Immediately, you foresee your final grade in math plummeting into an abyss of unfair testing.
Well, after you fail the Slavonic math test, you can go to the library and open up Standards for Educational and Psychological Testing to build a case against this test destroying your math grade.
Standards exists to provide guidelines for fair and accurate testing. Associations of psychologists (American Psychological Association) and educators (American Educational Research Association and National Counsel on Measurement in Education) worked together to develop this book, which is used and recommended by professionals in those fields as well as policy-makers whose work includes education. And perhaps even well-informed students who want to prove that they shouldn't need to know a foreign language to pass a math test.
Mostly, Standards is used for standardized tests. However, since Standards does everything possible to create and encourage fairness, you should be able to build quite a case for yourself. Let's take a look at each section of the book.
Part I: Test Construction, Evaluation, and Documentation
Our first stop on the fairness patrol is a visit to the test maker. Standards directs test makers to ensure that their tests only measure the one thing they say they test. In our case, that thing is addition. Our test maker, however, accidentally let a little bias sneak in by writing the test in a foreign language. Bias exists when certain groups will do better or worse based on something irrelevant to the test. For instance, a group of students who speak Slavonic will excel on our math test whether they know how to add better than us or not.
Fair and accurate testing starts with a test maker carefully hunting for bias by trying out this test to ensure that only addition is being tested (or whatever the subject may be). This might mean translating the test into any language represented by the test group.
This also might mean explaining cultural references. For instance, a bad question on an addition test might read, 'Five new reindeer join Santa's usual team. How many reindeer are on the team now?' This question assumes that the test taker has cultural knowledge about the number of reindeer on the team. The question should say, 'Five new reindeer join Santa's usual team of 13. How many reindeer are on the team, now?' This way, the test taker who isn't familiar with Santa, but knows how to add, will have all the needed information.
Since tests are never completely perfect, Standards also tells test makers that they are responsible to make clear exactly what the test has been proven to show and its margin of error. The margin of error indicates just how accurate the test is predicted to be.
For instance, let's say you study your addition facts for the retest you have argued yourself into, but you only study the addition facts for numbers 0-8, leaving off the 9s. In reality, you now know 90% of addition facts.
If the test happened to leave off all the 9s, the test would show that you knew 100% of addition facts, and if it consisted only of 9s, the test would show that you knew 0% of addition facts.
Neither of these would be accurate tests. As the test maker tests her tests, she tries to reduce the margin of error as much as possible and then reports whatever margin of error remains.
Part II: Fairness in Testing
Once the test maker produces the test, it falls into the hands of the test user. The test user is the person who administers the test and reports or uses its scores. Perhaps Mrs. Krause found this math test on a website advertising an excellent addition test. She used it without taking the time to look at the test maker's documentation, which would have told her that this is a very accurate test for measuring addition knowledge among Slavonic speakers.
Standards instructs the test user to read all documentation provided with a test and make informed choices about the test's application to her situation. Additionally, the test user must report the margin of error with the scores. This section goes into detail about how to be fair when selecting tests for diverse groups.
Part III: Testing Applications
Part III goes on to tell us that if a test user changes the way a test gets administered, perhaps by translating it or reading it aloud, the test user must give evidence that these changes are supported by research. She must record and report any effects these changes have on the margin of error. So, when Mrs. Krause translates your test to English, she needs to test the translated test before using it to test you.
She also might think to herself that the first test she gave you could work to test the 'Doing Math in Slavonic' class she teaches on the weekends. However, if she wants to use it to measure how well students can add in Slavonic, she will need to provide some evidence that the test is a good indicator of that. The test only labeled itself as an addition test, after all, and a class adding in a foreign language is slightly different than a native speaker adding in her own language.
All 194 pages of Standards of Educational and Psychological Testing boil down to two words: fairness and accuracy. Whether you are the teacher or the student or the policy maker selecting tests for state-wide use, the heart of testing is finding out what the test seeks to find and eliminating anything that hinders fairness and accuracy.
To unlock this lesson you must be a Study.com Member.
Create your account
Register to view this lesson
Unlock Your Education
See for yourself why 30 million people use Study.com
Become a Study.com member and start learning now.Become a Member
Already a member? Log InBack