Validita a reliabilita výskumných nástrojov: princípy a reálna prax
Validity and reliability of research instruments: Principles and actual practice
Languages of publication
The aim of the paper was to explore how principles of estimating validity and reliability of research instruments, as described in respected methodology textbooks, are satisfied. The initial parts of the paper delineate the theoretical framework and describe the concepts of validity and reliability. The following sections of the paper explain the process of analysis and its findings. The Journal of Educational Research was chosen as the research focus. A sample of 56 randomly selected articles from it has been inspected. The analysis revealed that a large majority (91 %) of research instruments used in these articles was scales and tests, the rest were questionnaires, observation schemes and interviews. Surprisingly, validity was calculated only with 26 of instruments; the rest of instruments were standardized tests or they were face-validated. As far as scales are concerned, construct validity was documented by means of factor analyses. Content validity and face validity were used in tests, questionnaires and interviews. We consider the infrequent use of combination of two sources of validity (e.g., construct and concurrent or discriminant) to be a weak element in the validation processes in the sample of studies. Reliability was documented with 80 % of research instruments. The most frequent method of calculation was Cronbach’s alpha. Inter-rater reliability was used in observations and tests; test-retest reliability was used to control the stability of the pretest-posttest measuring instrument. The size of reliability coefficients in most of studies exceeded 0.80. Throughout the analysis it was corroborated that when judging validity and reliability one has to critically consider the specific conditions of each research study before expressing an evaluation statement.
Publication order reference