The widespread view of science as consisting of experiments that test hypotheses obscures what is in reality a multi-stage logical process. Erroneous or vague resolution of issues at any stage can lead to inconclusive or contradictory results. Deriving a prediction from a theory or understanding a datum or experiment in relation to a theory is fraught with difficulties. Thus posing a well-formed scientific question and designing a proper study in relation to this question is difficult in all but the simplest cases. Delineation of objects and system boundaries is subject to definitional ambiguities and requires specification of experimental frames. Proper study execution, data evaluation, and logical inference of the meaning of results all have their own difficulties. In this study, each of these stages of a research project are evaluated to clarify the sources of ambiguities and uncertainties and suggestions are offered for reducing errors and speeding scientific progress.