All conclusions are meant to summarize the results and implications of a topic under study. This can seem like repetition most of the time, but the discussion is one of the most important parts of a science research paper because it's where a researcher should interpret results, discuss data, raise unanswered questions, be specific and give recommendations for future studies. In short, the discussion or conclusion section provides the most in-depth analysis of the findings in the study and gives readers a sense for what they should do with the information.
Different scientific papers will require varied goals, so it's beneficial to check the specific assignment requirements for each paper. However, in general, the discussion should draw conclusions that are supported by the given evidence, present information in a format that demonstrates the major results, clarify what should not be concluded from the results, explain implications and give suggestions for how to use the information. The discussion section enables the writer to frame how others should understand the study in question. It does not need to be lengthy, just enough to provide a full interpretation.
Some common problems in the discussion include overloading the reader with too much information, ignoring "negative" results that do not support the original hypothesis or the conclusion that the writer wants to draw, making statements that are too broad and unspecific, drawing conclusions that are not supported by the given evidence and focusing too much on problems and not solutions. Most of the time, it's best to stick to the evidence and draw straightforward conclusions, rather than explaining results that the researcher wishes were there, but aren't actually present.
The value of the conclusion is to demonstrate a writer's critical thinking skills, especially the ability to synthesize, analyze and reason well. The concluding argument should be structured logically; all points should flow and transition well from one to the next. Outside theories do not need to be discussed, as in a research paper in another discipline. Scientific research papers are only concerned with the specific topic at hand. Avoid repetition and stick to the main points to be made. The best conclusions will link interpretations directly to findings and hypotheses discussed earlier in the paper.