The method must fit the question.

Are there some questions which are not answerable by THE TEST methodology?

A while ago I posted a message " Blending and PCT ". No one replied.

In my doctoral dissertation, I was interested in two questions:

   Q1: Are skills X and Y necessary for learning to read as some leading
researchers were claiming?

   Q2: What impact does learning to read have on the skills X and Y which
were known to undergo developmental change in the early school years?

The research design was as follows:

        Pretest on skills X and Y
        Divide sample randomly so that the E and C groups were matched in
skills X and Y and an IQ measure.
        Teach E to read. Read to C, same materials.
        Posttest on skills X and Y.
        Posttest on reading skill.

From this sort of design I was able to say:

        With respect to question Q1: Neither skill X or Y are absolutely
necessary for learning to read. However, skill X predicted (correlation,
independent of IQ, was about .80) how much reading skill a child would
acquire.

        With respect to Q2: Both skills X and Y improved after learning
to read. Learning to read seems to be an early experience which changes
the children mentally, other than teaching them to read.

I don't see how I could have answered the questions I was interested in using
THE TEST. I am sure there are all kinds of questions for which the traditional
methods are the most appropriate.

While I am convinced that THE TEST can help answer the question: What is a
person's goal in a particular situation, there may be other methods for doing
the same thing.

For example, Clinical Psychologists use projective methods to accomplish the
same thing. A person is asked to say what they see when shown an inkblot
(Rorschach Test). A person is asked to tell what is happening in a picture (TAT
test). There is no correct answer. When given a number of such cards, themes
emerge which suggest what the person may be controlling (what are his/her
goals).

My point is that the research question is the important thing.

···

From: David Goldstein
Subject: The method must fit the question.
Date: 11/12/95