There are some overlaps between external validity and research ethics. For example, how to handle raw data so that they can be processed by some algorithms and methods while making sure the handling is not misconducted. Or, to what extent the data should be shared to allow the readers to replicate the experiments while keeping sensitive data credential. To understand those problems, this work firstly presents several alternative methods, then uses a combined systematic process to analyse several cases. We can see that one problem often has more than one solution and they should be carefully considered to select a suitable one if any. The combined process is working well and should be considered to engage when analysing research ethical problems.