DISEÑO DE INVESTIGACION PSICOLOGICA
Es posible calificar los diseños de investigación, caracterizar las dimensiones básicas en las que difieren y analizar las implicaciones de esas diferencias para la validez de los resultados de investigación.
From those years there is a
rapid development and
Gulliksen (1950) carries out the canonical synthesis of this approach.
Later
Lord and Novick
(1968) carry out a reformulation of the classical theory and open the way to the new IRT approach.
CLASSIC LINEAR MODEL
Spearman proposes a very simple, common sense model for people's test scores,
and which has come to be called the classical linear model.
Consists in
Assume that the score a person gets
on a test, which we call its empirical score,and
usually designated by the letter X, is made up of
two components.
The true score
of that person in that test (T), whoever it is.
For another
an error (e), which can be due to many causes that
they escape us and that we do not control.
To finally be like this:
X = T+e
In order to advance Spearman
adds three assumptions to the model.
Which are:
- There is no correlation between true scores and measurement errors.
- True (T) score is equal to the mathematical expectation of the observed score.
3.Measurement errors affecting two different tests are independent of each other.
T = E (X)
R (T, E) = 0
R (E, Ek) = 0
THEORY OF GENERALIZABILITY AND TESTS
REFERRED TO THE CRITERION
This classical approach has generated various variants on
all depending on the treatment given to the measurement error
LIMITATIONS OF THE CLASSICAL APPROACH
There have been numerous attempts to estimate the different components of the error, trying to decompose it into its parts.
The best known is
The Generalizability Theory (GT) proposed by Cronbach and his collaborators (Cronbach, Gleser, Nanda and Rajaratnam, 1972)
Is about
A complex usage model, which uses analysis of variance for most of its
calculations and estimates.
Another psychometric development are
Criterion-Referenced Tests (CRT).
These are tests used mainly in the field of
education and assessment in work contexts
Your goal is
Determine if people master a specific criterion or field of knowledge.
what measurement problems
were well resolved within the classical framework for
What new theories are proposed?
There was
two basic questions that did not find a good solution in the classical theory
And
They made the measurement
psychological was not homologous to the one exhibited
other empirical sciences.
1. Within the classical framework, the measurements are not invariant with respect to the instrument used.
2. The absence of invariance of the properties of the tests with respect to the people used
to estimate them.
If a psychologist assesses the intelligence of three
different people with a different test for each person, the results are not comparable.
We cannot strictly say which person is more intelligent.
This is so because the results of the three tests are not
on the same scale, each test has its own.
With more clarity
Important psychometric properties of tests.
Such as
The difficulty
of the items, or the reliability of the test
They were on duty
of the type of people used to calculate them.