Abstract:

In recent decades, science education has undergone significant transformations, incorporating tools to enable students to develop as scientists themselves. This approach aligns with modern theories that propose applying a constructivist perspective to the experimental learning process. In line with this, changes have been implemented in the course of Experimental Physics 1 at the Faculty of Engineering of the Universidad de la República, a public university in Uruguay.
Teachers have modified their approach to experimental activities by varying the level of support and guidance provided to student groups through practice guides. Students are now required to make informed decisions and actively participate in the development of the experiments. To evaluate the effectiveness of these changes, a study was conducted using written productions from practice guides and reports submitted by student groups.
Category models were constructed to evaluate the level of guidance in the practice protocols and the level of demonstration of knowledge of the groups of students.
The practice guides were evaluated based on the level of explanation of the guidelines. To evaluate the students' approaches and depth of analysis, categories were constructed using Bloom's hierarchical taxonomy. These categories served as a tool for assessing outcomes at a general and generational level in this and other courses.
The development of these categories represented an initial step towards their future application as a comprehensive tool for assessing outcomes across various courses and generations.
The evaluation of results requires the search for and recognition of approaches corresponding to the different levels of the categories in the textual productions. This study focuses on optimizing this procedure by comparing the results obtained with and without the assistance of programming languages for categorization of results. In particular, the recognition of the categories corresponding to the students' performance was carried out based on the identification of keywords associated with the different levels of Bloom's taxonomy in the reports.
Analyzing the frequency of occurrence of different categories revealed that the modifications made in the course between 2019 and 2020 yielded positive results, as the level of statements made by student groups in the reports increased.
Furthermore, a correlation was found between the degree of support and guidance in the practical proposals and the students' level of response. This highlighted the importance of providing clear instructions that incorporate the degree of freedom and detail expected, while leaving space for reflection and clarifications that are considered relevant so that all students can develop the proposed activities.
By employing assistance in identifying categories, positive results were obtained, consistent with the manual identification process opening up a new line of research concerning the recognition of categories through natural language processing and pattern recognition in students' proposals, which would constitute an important tool in the qualitative analysis carried out for course evaluation.

Keywords:

Science education, laboratory, higher education, Bloom's taxonomy.