Nowadays, there were many data published online and with papers, such as PLoS journal. I am wondering can I use those data to perform my analysis and if so, do I need permission of those authors to publish my results? Thanks.
Edits: The data I am trying to use is the supplementary data that provided by the article. The field I am working on is epidemiology. All comments are appreciated.
They've published their data. The way academia works is that people build on other people's work. You're just expected to reference their work properly.
Copyright is not particularly relevant here. Copyright law varies from country to country, but in any case copyright does not prohibit you from, e.g., adding two numbers just because those two numbers appeared in a copyrighted paper. Here in the US, copyright applies to the presentation of a work, not to ideas or data. For example, telephone directories have been found not to be copyrightable in the US. Since copyright is not very relevant, neither is licensing. Copyright law is specifically constructed so that it will not interfere with activities like scholarly commentary.
You will of course want to behave well toward the people whose data you're using. It won't hurt to talk it over with them. They know it intimately, and they may be able to warn you against misinterpreting it. For example, you don't want to be in the position of publishing a "gee whiz, amazing!" conclusion based on their data, only to find later that it was a misunderstanding. It may be appropriate for them to be coauthors on any resulting paper. Even if the purpose of your analysis is to contradict the conclusions they drew from their own data, it still doesn't hurt you to tell them what you're doing, and be above board and respectful.
The whole meta-analysis field would collapse if data extraction from published sources was not allowed, was restricted, or required permissions to be sought from the copyright holder.
It is also important that data cannot be copyrighted. Copyright is restricted to expressive and creative outputs, not measurements. So, if you are drawing from a raw data source, you do not have to ask permission. It is however standard practise to acknowledge the source of the data, which gives credit to the original gatherers of the data and lets those that read your output know where you got the data.
Note, it is possible to get a doi for your data when you post it on platforms such as figshare or zenodo. If the original authors have done so, then cite the data source, not the paper accompanying paper (unless you also refer to the outcome of the work).
The data published in most journals/repositories have licenses - as published articles have. So you have to check individually what the license of the data you want to use allows and what it doen't.
If you want to use a data, you should ask the publisher to get the agreement of use it. This is the same as any figure that you want to use. You should ask for permission because some journals does not give that permission even it is published.
Published data is public. This is essential to the proper functioning of science. Others can examine and re-analyze the data, challenge existing interpretations, and test new theories.
It is nice to let the authors know that you would like to analyze their data, and to show them the results for comment before submitting a paper, but it is not required (formally, or by academic custom) that you do so. The only non-negotiable requirement is that you properly cite the source of the data.
It is indeed a sign of progress that more studies make their data available in easy-to-use formats. I remember the days when we xeroxed plots on graph paper in order to extract numerical data.
