Lecture
One of the criteria for the objectivity of a study is its reproducibility.
The possibility of conducting repeated studies that yield the same phenomena and regularities. That is, the possibility of comparing data obtained for different participants or populations.
In this context, the concept of the reliability of data is introduced.
Reliable data, when obtained again under the same procedural conditions, produce insignificant deviations from the original values. This requires a probabilistic evaluation of the data on the basis of statistical decisions.
- The problem of the intersubjectivity of data, i.e., the possibility of obtaining the same data by different researchers.
The problem of the unreliability of the results of psychological research is directly related to the planning of the study's samples – their representativeness and the number of individuals (or observed cases) included in them.
The unreliability of data may be determined by the instability of psychological reality, by its being mediated through the subjective world of another person, by the unreliability of the methodological procedure, and so on.
The unreliability of data is their variability across repeated trials as a result of the non-systematic influence of extraneous factors and the variability of variables.
That is, the source of the unreliability of data is non-systematic confounding – when any of the extraneous variables, or their combinations, irregularly intrude into the dependence under study.
The source of an extraneous variable conditioned by the time factor may be either internal causes (changes in the participant's states) or external ones (accidental distraction by noise, a phone call, etc.).
To control such confoundings, one must bring the experiment closer to an infinite one – increasing the number of trials.
This reliability is related to the assessment of internal validity.
2) The reliability of methods – an aspect of operational validity
3) Reliability as the stability and reproducibility of results (an aspect of external validity).
Comments