Guidance, Regulator and a Rating System for Trials/Experiments in Social Sciences: Consolidated Conduction, Reporting and Evaluation of Trials Entity (CONCRETE).
Keywords : Assessment and Evaluation, Educational Effectiveness, Research Methods
0 users have rated
This article provides new answers to an old debate, dealing with bias in experimental designs in social sciences and especially the field of Education. To begin with the article provides a short review of studies which have investigated the influence of experimental design on the magnitude of the effect size (ES) in various areas of social sciences. The studies reviewed here all possess one element in common: they illustrate that stricter experimental designs report lower ES. A phenomenon which is argued to be mainly due to multi-level biasing associated with weaker designs, such as quasi or single case trials. Consequently, the aim of this article is to introduce news ideas with the objective of reducing bias for trials/experiments in social sciences while borrowing from the field of Medicine. Specifically three new ideas for trials in social sciences, especially education, are presented: 1) a list of 31 recommendations to serve as guidance for future trials, 2) the recommendation of an independent regulator to supervise trials and work together with multiple researchers and actors, 3) a quality evaluation system to judge the trustworthiness of trials in social sciences. The benefits of these three proposals, which are missing in social sciences, are also discussed.