Bounded measurement error estimates: their properties and their use for small sets of data