A guide to design of experiments in education
Experimental Statistics. Quality Education. Intervention. Randomization.
Education in Brazil has presented and continues to present challenges of various kinds. During the 20th century the main educational challenge was access to schools and combating the high rate of illiteracy. The construction of public schools as a way of facilitating access to basic education led to other challenges that persist to the present day, such as the low quality of the education offered, a constant and relevant theme throughout the world, since the quality of education is a critical issue for the development of a country. Several studies have been carried out with the aim of identifying ways to improve teaching and learning and one way to identify best practices is to conduct a statistical experiment as a way to establish a causal relationship between different factors and students' academic performance. Furthermore, these studies make it possible to evaluate the impact of public policies and government programs aimed at improving education, helping to direct resources more efficiently and effectively. As important as performing these interventions is conducting them properly so that it is possible to obtain reliable data. With the wide variety of existing statistical techniques, it is common for researchers from other areas to innapropriately use of experimental statistics when conducting an intervention. In this work, a literature review is carried out as a way of mapping studies that use experimental statistics in the educational context, in order to identify the difficulties found by researchers, as well as good practices carried out, in addition to carrying a case study to identify possible practice difficulties that were not noted in the review. This work aims to disseminate the correct use of experimental statistics techniques in a clear, objective and understandable way for researchers from other areas who are not familiar with that subject. To achieve this goal, an accessible and simplified language is adopted when explaining complex concepts, using concrete examples and graphics that help to visualize the statistical concepts. Moreover, the importance of using controlled experiments and randomization are emphasized, which allow inferences followed up on the basis of quantitative evidence. Additionally, as part of the proposed work, a guide will be developed that will present the steps and important points to plan an experiment properly, guaranteeing the validity and reliability of the results obtained. This guide will include guidance on designing the experiment, choosing sample size, selecting experimental and control groups, randomization, selecting appropriate statistical tests, and other aspects crucial to successful experimental research. It is hoped that this guide will provide a useful reference for researchers from other fields who wish to use experimental statistics correctly and effectively in their research.