In development practice today, when you ask ‘How do you improve governance systems in developing countries in order to improve the lives of the poor?’ the so-called hard skills dominate the discourse. But what are these so-called hard skills? At their most mind-numbing these are number-crunching skills derived from a variety of quantitative social science disciplines. Beyond that these are skills in technical analysis and solution-finding. So, if you want to curb corruption in Country XYZ you find the technical experts on building world class procurement and other systems send in accountants and the like and so on. You design systems, set up an Anti-Corruption Commission. You deploy your notion of ‘best practice’ in the relevant technical field. All this is well and good but will that blow a corrupt public political culture away and with it the broader tolerance of corruption by the population at large? The difficulties in getting results and making them stick being experienced around the world are becoming too big a pile of inconvenient facts to be ignored. Yet the fields of learning and the skills that can help to make governance reform initiatives work are often dismissed as ‘soft’. Those who study values, beliefs, norms, attitudes and public opinion, and what to do about them in specific situations – all significant influences on the outcomes of reform initiatives – are told they have nothing meaningful to contribute to the great efforts needed to improve governance systems around the world. And all those inconvenient facts? They are going to keep piling up until so-called hard sciences find the humility to expand their paradigms.