Table 7.

Mean Parameter Estimates for the Two Best-Fitting 1T-Models in Experiment 1

Model		Ter		st	decay	c2	c3	c4	σ
A(1T; 3C; eS)		0.356		0.124	0	0.577	0.870	1.004	0.433
LA(1T; 3C; eS)		0.353		0.114	0.175	0.579	0.841	0.967	0.460

	${\rho }_e^2$	${\rho }_m^2$	${\rho }_d^2$	${\rho }_e^3$	${\rho }_m^3$	${\rho }_d^3$	${\rho }_e^4$	${\rho }_m^4$	${\rho }_d^4$
A(1T; 3C; eS)	.976	.826	.702	.881	.773	.610	.875	.718	.554
LA(1T; 3C; eS)	.974	.859	.724	.920	.795	.615	.905	.763	.602

Note. Average of the parameter estimates across all 4 participants. T_er = nondecision time (in s) for any number of alternatives; s_t = range of variability in t_er; cⁿ = criterion for the corresponding n number of alternatives; σ = SD in Gaussian noise added to the accumulation process; ${\rho }_{\lambda }^n$ = input strength at the λ level of difficulty (e = easy; m = medium; d = difficult) for the corresponding n number of alternatives. Model variant labels abbreviate the models’ structure: A = accumulator (without decay); LA = leaky accumulator (with decay); T = t_er; C = criterion; eS = equal starting point.