Table 1.

Design characteristics investigated by simulation studies

	First author, year
Design characteristic	Street28 2002	Kanninen27 2002	Demirkale42 2013	Graßhoff47 2013	Louviere24 2008	Crabbe40 2012	Vermeulen48 2010	Donkers41 2003	This study
Number of choice tasks	8–1120*	360		Varied to achieve optimality	4,8,16,32*	16	9		2–20*
Number of alternatives	2	2,3,5*	2,3*	3	2	3	5	2	2–5*
Number of attributes	3–8*	2,4,8*	3–12*	1–7*	3–7*	3	2,3*	2	2–20*
Number of levels	2	2	2–7*	2	1,2	3		2	2–5*
Number of blocks			5
Sample size					38–106*	25, 250*	50
Outcome type	D-efficiency	D-optimality	Number choice sets to achieve d-optimality	D-efficiency	D-efficiency	D-error	Relative d-efficiency	D-error	Relative d-efficiency
Comments	Only 38 designs presented.	Attribute levels described by as lower and upper bound	Evaluate different components of blocks	Locally optimal designs created. Compared binary attributes with 1 quantitative attribute, swapped alternatives within choice sets	Variation of levels is referred to as level differences	Authors compared designs with and without covariate information	Compared best-worst mixed designs with designs that were: (1) random, (2) orthogonal, (3) with minimal overlap, (4) d-optimal and (5) utility neutral d-optimal design	Designs compared with a binary attribute with an even distributed vs a skewed distribution	Characteristics were individually varied, holding others constant, to explore their impact on relative d-efficiency

*Design characteristic has been investigated.