Can I change statistical results (mistakes) a discussion section with new results in a major revision stage? I corrected all the comments of reviewers but I don't know what should i do in this case. Might this situation affect the editor negatively?
In addition to making changes according to the reviewers' comments, you may also make other corrections and changes that improve the quality of the paper. However, you should indicate to both the editors and the reviewers that you have made such changes, and explain why they have been made. Given that it's a major revision, the paper would go back to the reviewers anyway, so it makes sense to correct everything possible.
As a caveat, however, such corrections should not be made in an attempt to skirt around flaws raised by the reviewers. I recently reviewed a manuscript where there were several anomalies in the data, and I mentioned these in my review. The authors completely removed the offending data, and replaced them with new results that better "fit" their conclusions. Such behavior is not acceptable, as they did not explain why the previous results were wrong; they just replaced them.
A major revision can potentially involve quite a large magnitude of change. As long as it remains essentially the same paper, you can add whatever makes sense to strengthen it and make it a better work than before. You can also drop material if, for example, you discover a mistake or you decide that it is distracting. Dropping material is much rarer than adding material, however, except in the case when you are up against a length limit and you have to drop something to make room for new material.
The test that I would propose for determining how much is to much for a major is as follows:
- If the changes might require different reviewers, it's probably too much.
- If the subject of the main result changes, it's probably too much (though the main result might be substantially strengthened or weakened).
As @aeismail says, however, make sure that you explain all of your changes clearly in your response letter, so that the reviewers understand what you have done and why.