Table 3.
Structure of Models Tested—The 8 models without decay shown below are structurally analogous to the 8 models with decay (not shown)
| Model | Evidence | St. pt. | Decay | Inh. | ter | Criterion |
|---|---|---|---|---|---|---|
| (constraints) | (parameters) | |||||
| A(1T,1C,eS) | ≥0 | = | no | no | 1 | 1 |
| A(1T,1C,cS|bS) | ≥0 | cor | no | no | 1 | 1 |
| A(1T,3C,eS) | ≥0 | = | no | no | 1 | 3 |
| A(1T,3C,cS|bS) | ≥0 | cor | no | no | 1 | 3 |
| A(3T,1C,eS) | ≥0 | = | no | no | 3 | 1 |
| A(3T,1C,cS|bS) | ≥0 | cor | no | no | 3 | 1 |
| A(3T,3C,eS) | ≥0 | = | no | no | 3 | 3 |
| A(3T,3C,cS|bS) | ≥0 | cor | no | no | 3 | 3 |
Note. St. pt. = Starting point; Inh. = Inhibition; ter = nondecision time; cor = negatively correlated. Model variant labels abbreviate the models’ structure: A = accumulator (without decay); T = ter; C = criterion; cS = correlated starting point (Experiment 1 only); bS = biased starting point (Experiment 2 only); eS = equal starting point. The other 8 models tested are identified by LA, meaning leaky accumulator (with decay), otherwise with the same structures.