|
One Sample T-Test
|
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 95% Confidence Interval | |||||||||||
| t | df | p | Lower | Upper | |||||||
| EitherVScontrol | 0.385 | 39 | 0.702 | -0.469 | 0.689 | ||||||
| Note. Student's T-Test. | |||||||||||
|
Bayesian One Sample T-Test
|
|||||
|---|---|---|---|---|---|
| BF₀₊ | error % | ||||
| EitherVScontrol | 4.241 | ~ 9.476e -8 | |||
| Note. All tests, hypothesis is population mean is greater than 0 | |||||
|
Descriptives
|
|||||||||
|---|---|---|---|---|---|---|---|---|---|
| N | Mean | SD | SE | ||||||
| EitherVScontrol | 40 | 0.110 | 1.811 | 0.286 | |||||
A wider prior means that larger effect sizes are probable. The figure suggests that a wider width does not change the results much.
This shows how the evidence changes as subjects are added. There are slight variations as subjects are added, but the result does not change for the last 10 subjects. However, if you were not convinced by these results, it is completely fine to add more subjects until you perceive the results to be robust. Note that such a sequential analysis is problematic in NHST (Neyman-Pearson) because an important prerequisite is that the total number of subjects (sampling plan) is predetermined; otherwise, the alpha is not valid. Further, the order in which subjects are entered may change how the evidence develops. Nonetheless, the final results will be identical irrespective of the order.