“…For controlling the type I error rate of similarity decisions based on 2 several approaches exist: Zhai, Mathew, and Huang (2016) present tolerance limits for the distribution of 2 . Several papers suggest the construction of bootstrapping confidence intervals (Ma, Wang, Liu, & Tsong, 2000;Ocana, Frutos, & Sanchez, 2009;Stevens, Gray, Dorantes, Gold, & Pham, 2015;Mangas-Sanjuan, Colon-Useche, Gonzalez-Alvarez, Bermejo, & Garcia-Arieta, 2016;Paixão, Gouveia, Liva, & Morais, 2017, among others) to control the type I error rate when using 2 for the similarity decision. A lower 95% confidence bound is determined by means of bootstrapping techniques (for details see Efron & Tibshirani 1993) and equivalence is concluded if > 50.…”