Problematic research practices, such as publication bias where only positive results are published, have been pointed out in the scientific literature for over half a century. Recently, large scale replication projects have suggested that not all published scientific research is as reliable as we want it to be. Psychological science has been at the forefront of improving research practices, due to a traditionally strong expertise in statistics, combined with an interest in how people change behavior and respond to reward structures. In this presentation I will talk about some of the problematic research practices that have limited knowledge generation in the past, how to recognize them, their consequences for the reliability of research findings, and ongoing efforts towards better research practices that have been developed in the last eight years. I will summarize some easy to implement improvements in designing and analysing experimental studies.