From replicability to generalizability

Scientific findings are taken as reliable when comparable results can be obtained after replicating the original study with similar methodologies. In her thesis Maria Arroyo Araujo explored whether using the same experimental protocol in different laboratories can give comparable results; in addition, she also tested how small changes in environmental variables of the protocol could influence the results across laboratories to make them more comparable.

Arroyo Araujo: 'From these studies we learned that results become more comparable when the different labs follow the same protocol than when each lab follows their own protocol for the same task; however, this scenario is not always feasible in animal science. In another study, we summarized findings published in the literature that showed great variability of results among studies that measure the same behavioral construct in rodents. These differences couldn’t be explored in relation to the diverse experimental protocols used as few articles had a complete and clear report of their methodology; therefore, another lesson from this thesis is the importance of complete and transparent reporting to accurately assess the replicability of results and the factors that might affect it. A way to improve the planning, performance and reporting of preclinical studies is by using the openly available guidelines or to use a quality system, such as the one discussed in this thesis. The EQIPD quality system guides the researcher through the scientific process in such a way that promotes research practices that to enhance the quality of their research and thus, the value of it.' 

Dissertation: https://hdl.handle.net/11370/c9188f3c-22ca-44b8-a4a5-d4567fd718bd