Internal validity, partly statistical conclusion validity. Part of external validity is called conceptual validity.
Meaning:
Refers to the degree of certainty of causality between independent variables and dependent variables in research.
Influencing factors:
History; Choose; The subject died. Mature; The growth of testing experience; Fatigue factors (Hawthorne effect, placebo effect, required characteristics). Accuracy and reliability of experimental instruments and measuring means, statistical regression factors. The interaction between selection and maturity.
Improvement method:
Conduct strict experimental design-control independent variables, dependent variables and irrelevant variables. Strictly monitor the implementation plan, process (accident) and links of the experiment.
Meaning:
It refers to the extent to which research results can be generalized and popularized in other similar situations, that is, the universal representativeness and applicability of research results.
Influencing factors:
The reaction effect of the test; Reaction effect of experimental arrangement; Interference of repeated experimental treatment. Interaction between selection bias and experimental variables.
Improvement method:
Increase the representativeness of the sample, overcome the over-artificial situation of the experiment, and ensure the effectiveness of the measuring tool.
# Statistical regression effect, the students who did poorly in the first exam may do better in the second exam, and the students who did better in the first exam may do the opposite. This is the so-called statistical regression effect.
Meaning:
The accuracy of the concept of relationship between variables, the consistency of the operational definition of experimental variables and the reasoning definition.
Influencing factors:
Conception: the analysis of conception before operation is not complete enough. Confuse conception and conception level. The improvement between conception and conception is limited.
Subjects and subjects: required characteristics. The subject's anxiety about the evaluation. Experimenter expectation effect.
Procedure and operation: deviation of single operation; Deviation of single method.
Interaction: the interactive influence of different treatments; Interaction between testing and experimental treatment.
Meaning:
The validity of statistical conclusions caused by the appropriateness of statistical methods mainly reflects the relationship between statistics and overall parameters.
Influencing factors:
Statistical method: low statistical verification (sample size, significance level A, error variation). The degree to which the assumptions on which the selected statistical methods are based are satisfied. Multiple comparisons and error variation.
Subjects: random changes of subjects.
Tools: measure the reliability of tools.
Procedure: the reliability of experimental treatment, that is, the degree to which the experiment meets the design requirements.
Environment: In the experimental environment, there are no qualitative unrelated accidents.