Test performance or general average? Some questions about measuring academic performance in research
Introduction: Academic performance (AP) is a relevant subject in educational research, but its operationalization is still subject to controversy since there is no unified opinion on considering weighted average (WA) or test performance (TP) as a measure of ap in research, especially with situation-...
- Autores:
-
Dominguez-Lara, Sergio
- Tipo de recurso:
- Article of journal
- Fecha de publicación:
- 2017
- Institución:
- Universidad Cooperativa de Colombia
- Repositorio:
- Repositorio UCC
- Idioma:
- spa
- OAI Identifier:
- oai:repository.ucc.edu.co:20.500.12494/9612
- Acceso en línea:
- https://revistas.ucc.edu.co/index.php/pe/article/view/1712
https://hdl.handle.net/20.500.12494/9612
- Palabra clave:
- Rights
- openAccess
- License
- Derechos de autor 2017 Pensando Psicología
Summary: | Introduction: Academic performance (AP) is a relevant subject in educational research, but its operationalization is still subject to controversy since there is no unified opinion on considering weighted average (WA) or test performance (TP) as a measure of ap in research, especially with situation-dependent constructs such as test anxiety (TA).Objective: The objective was to analyze empirical equivalence between WA and TP.Method: 115 psychology students from a private university with ages between 18 and 31 (M = 22,704; SD = 3,126) belonging to three different academic cycles participated. They were evaluated with the State-Test Anxiety Inventory, also considering test scores and weighted averages of each student. Regarding data analysis, bivariate correlations between tp and wa (rTP-WA) were implemented. Subsequently, the rTP-WA observed in different cycles were compared. Finally, we analyzed differences between ta-tp (rTA-TP) and TA-WA (rTA-WA) correlations observed in each cycle. All analyses were complemented with a magnitude measure of the effect.Results: The correlation between TP and WA is of low magnitude, and the differences found between correlations were not statistically significant.Conclusions: It is not convenient to consider TP and WA as equivalent in the context of studies with situation-dependent variables. The importance of overcoming the limitations observed in this study was suggested. |
---|