A non-randomized triangular design has been shown to be more efficient than the conventional random response model in estimating the prevalence of sensitive attributes in surveys. Since most surveys focus on estimation, herein we derive sample size formulas for estimation of prevalence and a...