Choosing an appropriate software development method is an important managerial decision problem. It is desirable to base a critical decision like this on proven knowledge. We look at the nature of this knowledge and how it develops in software engineering. A number of fundamental limitations to scientific enquiry in software engineering result in a corresponding body of knowledge that is diverse, complex, incomplete, ambiguous and inconsistent. This naturally limits the scope of formal approaches to software method selection. In this situation software method selection is a complex managerial decision-making activity under uncertainty. The human mind copes with this demanding task well in some cases and badly in others. Researchers have identified a number of human mental processes that can have a detrimental effect on managerial knowledge management and decision making. We look at some of these processes and discuss how they can flaw software method selection. Forewarned is forearmed: From our analysis we derive some advice on how managers can guard against these fallacies.