“…Based on the discussion so far, it is clear that when constructing a forced-choice measure resistant to SDR, it is essential to identify item-combinations for which differences in desirability evaluations between items are sufficiently similar across test takers. Typically, item desirability matching relies on empirically obtained item desirability ratings (e.g., Chernyshenko et al, 2009;Christiansen et al, 2005;Converse et al, 2010;Drasgow et al, 2012;Guenole et al, 2018;Lee et al, 2019;Naemi et al, 2014;Usami et al, 2016;Vasilopoulos et al, 2006;Watrin et al, 2019). In one common approach, a "desirability sample" is asked to explicitly rate desirability of each item under consideration (e.g., Chernyshenko et al, 2009;Christiansen et al, 2005;Usami et al, 2016).…”