Method evaluation
Psychological evaluation targets either internal psychological processes (e.g. affects), or
specific behaviours. Since the definition of the field may be more or less accurate we ask
ourselves whether the aptitudes / characteristics measured by the test correspond to the
field of evaluation. Thus, for an evaluation instrument to be considered a good measure of
a field, we need to know that it actually measures what it set out to do by design. These
requirements yield the analysis of its psychometric properties.
Psychological tests are an objective and standardized measure of a behaviour sample
(Anastasi, 1979) and represent one of the frequently used techniques in psychological
evaluation. In order that on the basis of test results to be able and formulate correct
conclusions on a person’s behaviour, the following characteristics of the test must be
evaluated: reliability, validity, construction norms.
Reliability
Reliability is given by the relative absence of measurement errors in a psychological test.
More often than not the reliability of a test refers to two aspects: internal consistency, and
stability / constancy in time of the results.
Internal consistency of a test represents the extent to which all items measure the same
variable. The coefficient of internal consistency is calculated by reporting the variance of
each item and the variance of the total score. The internal consistency of BTPAC has
been analysed on the basis of Alfa Cronbach internal consistency coefficients (e.g.
Analogical transfer test, Spatial aptitude evaluation tests), and on the basis of the halving
method (e.g. Reaction time evaluation tests). The values obtained for each test (0.50-0.85)
indicate a good internal consistency; therefore it can be said they reliably measure the
constructs.
Taking into account the factors influencing the calculation of the test-retest reliability –
stability – (the time interval elapsed between test and retest, the degree of item difficulty,
the changes in the performance of the subject influenced by the first testing) the study of
the stability in time of the BTPAC test results was carried out at a two week test-retest
interval. The interval was determined because as strict an analysis was desired of the
result stability, considering the sources that can distort the coefficient. The BTPAC tests
have shown test-retest coefficient that varies between 0.50 and 0.80. The value of these
coefficients indicates score stability.