Evaluación de las competencias investigativas en los estudiantes de maestría de la Universidad Nacional Experimental del Táchira mediante el uso de modelos de regresión multinivel

• Main Author: José Alexy Moros Briceño
• Format: Article
• Language: Spanish
• Published: Universidad Francisco de Paula Santander 2018-01-01
• Series: Ecomatemático
• Subjects: investigative competences; hierarchical structure; students; masters; multilevel regression models
• Online Access: https://revistas.ufps.edu.co/index.php/ecomatematico/article/view/1670/2546

The objective of this research was to apply multilevel analysis for the adjustment of a multilevel model aimed at evaluating the competencies of the masters students of the National Experimental University of Táchira, Venezuela. Theoretically, the model analyzes the variability among the students within each master’s degree (level 1) and among the master’s degrees (level 2), whose initial hypothesis establishes that the students of the same master’s degree possess similar investigative competencies with respect to those who belong to other master’s degrees. It was supported by an explanatory, correlational and quasi-experimental study. The sample consisted of 225 students, applying a structured instrument in two parts: one of 12 items on a nominal scale to measure personal and academic variables, and another on a 40-item Likert scale to evaluate research competencies. It was found that there are significant differences between the masters with respect to the research skills, because the intraclass correlation coefficient (ρ=0.142) of the null model prevents compliance with the hypothesis of independence: Students of the same master’s degree have similar investigative skills in comparison to those who are of another master’s degree, multilevel analysis is appropriate. Then, variables predictors of the levels were incorporated: masters, students and both levels, obtaining the following multilevel model: COINij=1,3376*CMLC+2,5882*CML-1,0031*ST1*CMLC+ uoj+eij