Fudging is the process of altering data to agree with predictions and desired results. It is of no value in science, since the results have been modified from their factual state and thus cannot be considered to be valid and reliable. Most scientists agree that it is a dishonest practice, akin to altering written records bit by bit until they say something completely different than before. Rather than drawing conclusions from the data, the conclusions are pre-established before the experiments. The data is then pushed and prodded to fit into the assumed results. When this doesn't work, full-scale fudging often occurs, where the data is falsified to meet expectations.