General. The closeness with which separate observations (or parameter values calculated from observed data on the basis of a model) agree with each other. According to Bevington and Robinson (2003) a measure of how carefully a result is determined without reference to any "true" value (in contrast to accuracy)