Table 1.
Comparison of ADO and Staircase (SC) methods in their reliability, precision, and efficiency (see “Methods” for their definitions) of estimating temporal discounting rates (log(k)).
| Measures | ADO | Staircase (SC) | |
|---|---|---|---|
| Reliability: Maximum test–retest reliability (TRR) | Experiment 1 (College students), Visit 1 | 0.961 | 0.903 |
| Experiment 1 (College students), Visit 2 | 0.982 | 0.946 | |
| Experiment 2 (Patients w/ SUDs) | 0.973 | 0.892 | |
| Experiment 3 (Amazon Mturk)a | 0.965 | N/A | |
| Precision: Within-subject variability (SD of individual parameters) | Experiment 1 (College students), Visit 1 | 0.122 (0.105) | 0.413 (0.252) |
| Experiment 1 (College students), Visit 2 | 0.098 (0.070) | 0.537 (0.409) | |
| Experiment 2 (Patients w/ SUDs) | 0.073 (0.063) | 0.371 (0.180) | |
| Experiment 3 (Amazon Mturk)a | 0.339 (0.262) | N/A | |
| Efficiency: Trials required to reach 0.9 test–retest reliability | Experiment 1 (College students), Visit 1 | 7 | Failed to reach 0.9 even after 42 trials |
| Experiment 1 (College students), Visit 2 | 6 | 39 | |
| Experiment 2 (Patients w/ SUDs) | 11 | 27 | |
| Experiment 3 (Amazon Mturk)a | 11 | N/A | |
aExcept for Experiment 3 (Amazon Mturk participants), all experiments used 42 trials per session. In the Amazon Mturk experiment, there were 20 trials per session.