How to improve research methodology in gender care: a non-binary choice

Discussions about the evidence base for the Dutch gender care model, specifically puberty blockers, easily culminate in a binary choice between randomized controlled trials (RCTs, called 'not ethical/feasible' by some) and the pre-post design which compares patient outcomes after treatment with measurements before treatment within a single group of treated patients (called 'scientifically weak' by others). The RCT has two distinguishing features: First, an RCT compares a treated group with a control group that has received no, or another, treatment. Second, an RCT assigns patients to treatment or control by randomization to ensure that both groups are comparable before treatment. To make the discussion non-binary, this paper focuses on the design with a control group but without randomized assignment, known as a quasi-experiment in psychology. Its pros and cons are discussed, as are some improvements to it and statistical methods that partly make up for the lack of randomization.