Empirical research is vital if software engineering technology is to be meaningfully evaluated --- without it only guesses can be made at the competing merits of different approaches. Unfortunately, as this thesis demonstrates, much of existing empirical work contains some kind of weakness...