How good is good?

Whether a result is a number, a fact, or a subjective impression, it is difficult to interpret if it cannot be compared to anything else

What sources of comparison are likely to exist, against which a result can be compared?

Can we engineer a comparison situation where none formerly existed?

In a comparison of, say, two software systems, is it better to use two groups of individuals such that one group tests one system and the other a different one, or have both groups test the two systems at different times?

©1994-2003 Kevin Boone, all rights reserved