In this dissertation, we focused on two alternative approaches to evaluate the hypothesis of interest more directly, i.e. informative hypothesis testing and model selection using order-restricted information criteria. These approaches have shown to be more ‘powerful’ than NHST. The main implication is the possibility to reduce costs. Data collection in the social and behavioral sciences is usually the most expensive part of conducting research. Since the outcome of this dissertation ensures that researchers can use smaller samples, the costs of data collection can be reduced. In addition, researchers who are dealing with inevitable small samples in particular may benefit from these alternative approaches. Finally, we hope that this dissertation gives applied researchers a push to employ more
informative hypotheses