Bayes factor tests for intervention effects

When people are suffering from mental issues like depression or anxiety, they can seek help from an intervention against these mental issues. When an intervention is new, researchers typically want to investigate whether the intervention has the desired effect on relevant outcome variables. Also, in recent decades, insurance companies have increased the pressure on mental health care workers to systematically monitor the progression of their clients before, during, and after the intervention, in order to avoid the funding of ineffective interventions. The psychological scales that are used to measure the mental states of the clients over time are typically indirect, not completely reliable scales for psychological constructs like depression. That is, the observed scores on these scales typically consist of a mixture of true scores and random measurement error. Hence, observed changes in scores collected before and after the intervention may be the result of measurement error rather than true changes over time. This raises the question of how much evidence observed data contain for or against a true intervention effect, as compared to no true intervention effect where observed differences are solely due to random measurement error.In this thesis I developed several hypothesis tests for intervention effects which quantify the relative evidence in data for or against hypotheses regarding intervention effects. The hypothesis tests are so called Bayes factors, which are built from within the Bayesian statistical approach. I developed Bayes factors for mean and trend change after the intervention for single-subject designs as well as group designs