Correction for attenuation: WAIS-IV and test X

[UnnieE]

Hi,

I have been asked to do a correction for attenuation of the context between a specific test (Lets call it test X) and a specific part-test of WAIS-IV (word understanding, lets call it test Y).

The reliability for WAIS-IV word understanding (test Y) is r = 0,85 and we assume that the observed correlation between scores on these tests is r = 0,73.

Can anybody help me solve this problem? google gave me very little information on this topic

Thanks beforehand

 

[hlsmith]

So you are trying to see how correlated two tests are? Is there a gold standard as well? I did not really following what your question was!

 

[UnnieE]

So you are trying to see how correlated two tests are? Is there a gold standard as well? I did not really following what your question was!

Hi! And yes i think that is what they want me to do. However i understand that it might be hard to understand my question because i tried to translate it. English is not my first language so i'm sorry for the translation mistakes. And no sorry we have not been given a golden standard or any examples.

 

[spunky]

What is the reliability of test X? You need the reliabilities of both the WAIS-IV and test X to correct for attenuation.

 

[hlsmith]

And I am assuming reliability is established via a gold standard. Unless you are assuming WAIS is the gold-standard so you are trying to figure out where/how X is deviating!

 

[spunky]

And I am assuming reliability is established via a gold standard. Unless you are assuming WAIS is the gold-standard so you are trying to figure out where/how X is deviating!

It seems to me (OP will let us know) that this more of a classroom-type exercise t where they want students to recognize how to use the correction-for-attenuation formula from Classical Test Theory

\(r_{x'y'} = \frac{r_{xy}}{\sqrt{r_{xx}r_{yy}}}\)

 

[UnnieE]

It seems to me (OP will let us know) that this more of a classroom-type exercise t where they want students to recognize how to use the correction-for-attenuation formula from Classical Test Theory

\(r_{x'y'} = \frac{r_{xy}}{\sqrt{r_{xx}r_{yy}}}\)

Hi! And thank you all for such quick answers. I will try my best to answer all your questions.

- The reliability of test X is unknown, this is a exam practice question and we are only provided with the details i listed in the start unfortunately

- And yes i think "spunky" is right that this is more based on learning the correction for attenuation formula, but i find it hard to use here since we have only been provided with the reliability for WAIS-IV word understanding (test Y) r = 0,85
and then the rest of the question goes on to say "And we assume that the observed correlation between scores on these tests is r = 0,73."

 

[UnnieE]

Important update! mystery solved

I contacted the teacher, and she sent me a sav file with the data sett of test X, so i think she wants me to calculate the reliability from that datasett by myself which is great because then it all makes sense. So i will calculate the r for test X, and i already have the r for test Y, and then i will just use the formula and everything should be just fine

Thank you everybody that commented and tried to help me! it is much appreciated.