No matter what action is taken in conducting research there are assumptions being made.  This holds especially true in defending the theoretical models and experimental procedures being used in a project.  These need to be identified to either address the limitations of your work, or to help you identify what may be going counter to your original assumptions as your work progresses.

These assumptions could relate to the plausibilty of a particular research tasks.  For example, if one were to use a sensor network to monitor particulate matter concentrations in different zip codes across a city, then an assumption would be that such concentrations would vary signficantly on the scale of a few kilometers so that such fine-grained monitoring would be worthwhile.  The assumptions can also fit into the category of being sure about what you're actually measuring.  If you're measuring the atmospheric concentration of carbon dioxide, and relating it to local industrial activity, then one assumption being made is that the carbon dioxide being produced above background is only coming from industrial activity and not from another source such as a distant volcanic eruption or the breakdown of methane from a local seep into carbon dioxide and water vapor.

There are many types of assumptions which will come about in a project, and which need to be identified, but they all stem from addressing the question: how do I know what I know?