The validity of quant itative researches in the education field

Dr. Janez Sagadin

Summary:  The author considers a study as valid if it can provide valid and credible results. The conditions for the validity of quantitative empirical research in the education field are the following: the validity of data collecting procedures / measurements as a condition for the validity of collected data, the validity of data analysis results, the validity of generalization and the validity of causal explanations or conclusions. He points out how to satisfy these conditions, explaining the causes of validity disturbance, in the framework of the particular condition. He finds out that we can only more or less approach the absolute fulfilment of conditions and consequently also the absolute validity of the research. In connection with the first condition, he particularly emphasizes the advantage of the measurement model of the confirmatory factor analysis over the measurement model of the classical test theory. When treating the second condition, he points out the validity of statistic procedures of data analysis (statistic validity of the research). In the framework of treating the third condition, he considers the conditions for the validity of generalization, from probability samples to target populations (statistic induction), pointing out also the contents view of generalization. Since the employment of statistic induction is not applicable to improbability samples, he suggests, in that case, the direction to the internal validity of findings. In addition, he shows the arbitrariness of concluding about target populations. The discourse on the validity of generalization, regarding experimental studies, is organically included into the treatment of the validity of causal explanations. The author states the conditions for the existence of the causal connection between two variables. He also states the assessment criteria if there exists such a connection. He explains that the inductive – deductive approach or procedure is typical of quantitative empirical research. First, there is a deductively achieved and theoretically checked (verified) causal hypothesis (hypothesis about a causal connection) and then the empirical testing of this hypothesis. It verifies if the hypothesis (including its causal explanation) is supported by empirical data, properly collected and analysed. In that case, the causal explanation is considered as valid. In the article, the empirical testing of causal hypotheses is more specifically presented for experimental studies (for random and non-random experiments, especially for those, which are more applicable to the field of education). Regarding the testing, he emphasizes the role of internal validity of experiments. Concerning non-experimental studies, the above mentioned testing is presented in a shorter form, as well as the procedures regulated by SEM (Structural Equation Modelling).