Educational research is a social science that has greatly developed over the past years. However, it does not enjoy the high reputation the natural sciences or other social sciences such as economy have traditionally enjoyed. This presentation is the answer to Lagemann´s call for action (2002) to strengthen educational research and to historian Carl Kaestle´s recommendations (1993) to increase the reputation of educational research. We believe that one of the problems in our field is the lack of valid scientific instruments which help us obtain reliable scientific data. It is critical that educational researchers can count on these instruments. However, these instruments are not always available for all the different educational areas to study. In this presentation we argue in favour of designing and conducting our own scientific instruments.
The main objective of this presentation is to offer a model for the creation and the validation of instruments to obtain scientific knowledge in the educational field. In order to do so, we will describe the endeavors of creating a quantitative instrument that helps researchers build qualitative data. We will also explain the process of getting the instrument validated.
The instrument we have designed is a Likert scale questionnaire, one of the most commonly used instruments in social and educational science research (Croasmun and Ostrom, 2011; Echauri et al, 2014). If we want to link research and practice, we need a method that collects quantitative data and that can help researchers draw qualitative findings and conclusions. Likert scales take into account the human aspect of education – the educators´ opinions and views- and they are excellent tools to measure parameters for groups of people.
We have validated our questionnaire by two procedures: statistical analysis and experts´ judgment. The former assesses our instrument´s overall coherence by calculating its internal reliability; the latter assesses the instrument´s content by having a group of experts analyze each of the items of the questionnaire individually.
RESULTS AND DISCUSSION
In the results section of this presentation we explain the measures obtained in the validations and how they helped us build the final version of our questionnaire. This final version of our instrument has been tested on two groups of teachers from the USA and from Spain, as part of a doctoral dissertation. Through the process of creating a solid questionnaire, getting it validated, and testing it on two international groups of teachers, we have gained invaluable insight into the steps that need to be taken in such a complex process.
Being involved in this process has made us come to the conclusion that, while educational research needs to use scientific methods that ensure reliability, researchers cannot forget the complexity of the educational science, which is larger than quantifiable data. We believe that quantification needs to be a tool that helps educational researchers understand our complex field of study, and not the other way around.
Documentación de apoyo a la presentación ONLINE de la ponencia
Documento Complementario de Apoyo
Jaime Fauré Niñoles
Comentó el 11/12/2020 a las 19:02:35
Muchas gracias por su excelente presentación. Mi duda es si los resultados con los que cuentan en la actualidad ya les permiten saber qué grupos de profesores tienen puntajes más o menos altos en cada una de las dimensiones.
Irene García García
Comentó el 12/12/2020 a las 13:37:23
Como respuesta a su pregunta podemos decirle que sí. Ya tenemos las puntuaciones en cuanto a las 3 hipótesis que planteamos y creemos que son resultados muy satisfactorios. Aún estamos en proceso de analizar cada dato y sacar las conclusiones y discusión de los mismos, pero ya podemos afirmar sin duda que nuestro instrumento ha dado los frutos que esperábamos y que satisface con creces las necesidades de la tesis que estamos escribiendo.
Gracias por su interés,
Ángel Alberto Magreñán Ruiz
Comentó el 11/12/2020 a las 14:16:16
First of all, thanks for the work, it is really useful.
I have some questions about the content of your talk. How did you validate youir instrument statistically? By means of using the Cronbach test? On the other hand, how many items did you recommend in a Likert scale? and how man scale items?
Thanks in advance for your work.
Irene García García
Comentó el 11/12/2020 a las 15:22:51
Thank you for the interest you have shown on our paper. Yes, we calculated our instruments' internal reliability through statistical analysis, more especifically using the method developed by Cronbach (1951) which is integrated into the statistical package SPSS.
The number of items you should include on a LIkert scale depends on how many hypothesis you need to test. In our case, we had 3 constructs we needed to measure, and we dedicated 10 items to each of the hypothesis. In total, we had 29 items (two of the original items were merged into one after the experts validated our questionnaire). However, we have seen questionnaires in other scientific articles with 80 or more items. In the end, it will depend on the characteristics of your study. However, if you are studying a population which volunteers to answer your questionnaire (as in our study), we recommend you keep the questionnaire short, since a very long questionnaire may discourage people to answer it.
Finally, we used 6 scale items. Normally people use 5 items ('strongly disagree', 'disagree', 'neutral', 'agree' and 'strongly agree') or 7 items (the ones mentioned before plus 'somewhat disagree' and 'somewhat agree'). We decided to add the option 'does not apply' due to the specific characteristics of our population.
I hope I have clarified all your doubts.
Deja tu comentario
Lo siento, debes estar conectado para publicar un comentario.