Fig. 11

Bias versus the nonadherence rate for the null and non-null effect using DAG 2. The naïve 2SLS and 2SRI share the same line, the first stage adjusted 2SLS and 2SRI share the same line, and both-stages adjusted 2SLS and 2SRI share the same line as they produce the same amount of bias. The bias for the ITT and baseline-adjusted ITT are out of the bound [-0.1,0.1] for the non-null effect. Abbreviations: RD: risk difference, ITT: intention-to-treat, PP: per-protocol, AT: as-treated, IP-weighted PP: inverse probability-weighted per-protocol, 2SLS: two-stage least square, 2SRI: two-stage residual inclusion