It seems unsatisfactory that much teaching practice is based on ideas with only weak or sometimes even no documentation for their effect. Many resources in terms of money and time have been lost on implementing ideas that after a short while must be dropped because they did not function well in practice, or had no relevant importance for student outcomes. In education we have quite often witnessed introduction of strategies and methods based on personal beliefs, habits, fancy and fads; not research findings and evidence. Fields like education that frequently are occupied with ill defined problems and at the same time are oriented at gaining improvements are especially exposed to fads. Fads are tricky because they seem promising, appear suddenly, grow to extensive popularity, but then just as sudden as they appeared disappear again. And further more they are very often forgotten just as fast as they rise to the surface (Best, 2006). In order to avoid fads, fancy and personal bias in education the science of teaching has gained ground over the last decades. Today we have from research and especially from syntheses of research results quite much evidence on what works and to what degree it works. This presentation will give a brief introduction to meta-analyses and syntheses of educational research related to student achievement (Hattie, 2009, 2011). And then point to teaching methods that are manageable in classes of any size, are engaging to students, and qualified for increasing and developing students’ abilities and competences.