Erregistroa posta elektronikoz bidali: Estimating the Causal Effect of Randomization Versus Treatment Preference in a Doubly Randomized Preference Trial