Parameter description | Notation | Values |
---|---|---|
Total number of randomised participants | n | 100, 200, 500, 1000 |
Set 1 | ||
 Generation of intermediate variable |  |  |
  Intercept (log(odds) of the event in the control group) |  | loge(0.2) |
  Treatment effect on the intermediate variable | αR | loge(1, 1.05, 1.1,…2, 5) |
  Effect of unmeasured confounding on intermediate variable | αu | loge(0.8) |
 Generation of outcome variable—continuous outcome study |  |  |
  Intercept (control group mean) |  | 3300 |
  Treatment effect on the outcome | βR | 0, 0.1, 0.2,…,2,5 SDs |
  Effect of unmeasured confounding on the outcome variable | βu | − 0.2 SDs |
 Generation of outcome variable—binary outcome study |  |  |
  Intercept (log(odds) of the event in the control group) |  | loge(0.1) |
  Treatment effect on the outcome | βR | loge(1, 1.05, 1.1,…2, 5) |
  Effect of unmeasured confounding on the outcome variable | βu | loge(1.2) |
Set 2 (differences from set 1) | ||
 Generation of intermediate variable |  |  |
  Interaction between treatment and unmeasured confounding | αRU | loge(0.8) |