Αναλυτικά στατιστικά μάθησης. Στατιστική ανάλυση των εξ αποστάσεως προγραμμάτων του ΙΝΕΠ

Abstract

Using Educational/Learner Analytics methodologies, this analysis attempts to test the following assumption: Are we expecting a public employee with a type “A” personal profile to underperform when participating in a Government Training Program that on average is underrepresented by type “A” personal profile employees? And if one educational program is being more heavily attended by type “B” personal profiles (type “B”s have a higher percentage in this Program compared with their across all projects average), should we expect a type “B” person to overperform? By assuming that free participation in a Program by a person indicates compatibility/suitability, it is natural to expect that public employees of a type that is being underrepresented (lower percentage of his type than on average) in a specific Program will underperform. After a brief statistical overview of the general trends in “INEΠ’s” multi disciplinary programs, we identify specific subcategories of training programs where the profile of the average participant diverges (with a amin focus on distance learning and digital competence improving courses). An econometric regression has been applied where use has been made of data on personal profiles of 65.000 people that have been trained in the Institute during the last 2 years. The regression provided discouraging results, with statistically insignificant estimated slope parameters for all but one variable. The 11 used (profiling) variables could not identify a predictable component in the grades awarded. Two rational explanations are offered. Lack of reliable data and lack of normality in the distribution of the dependent variable. Almost all of the observations (grades), were suspiciously gravitating towards 10 (full marks). More research is need