Commitment development in software process improvement: critical misconceptions

It has been well-established in the software process improvement (SPI) literature and practice that, without commitment from all organizational levels to SPI, the initiative will most likely fail or the results will not be far-reaching. In this paper, the 'commitment' construct is explored, and three forms of commitment are introduced: affective, continuance and normative commitment. Analysis shows that current models of commitment development lack scientific validity and are based on four misconceptions: (1) the assumption of linearity of the human cognitive process (i.e. commitment in this case), (2) the controllability of this process, (3) the notion of a singular commitment construct, and (4) the sole utility perspective on the commitment phenomenon. Implications of these findings for SPI research and practice are discussed.