Table 2.

Estimated effect of self‐training versus no training on the paper 2 RQI, from the primary and various sensitivity analyses

Analysis	Estimate	Standard error	p‐value
Primary analysis, MAR	0.237	0.070	0.001
Multiple imputation, MAR	0.234	0.071	0.001
Multiple imputation, copy no training†	0.172	0.069	0.013
Multiple imputation, expert opinion	0.195	0.132	0.145
δ k∼N(−0.21,0.462)
Multiple imputation, δ‐method with	0.189	0.072	0.009
δ=−0.21†

^†Information‐anchored sensitivity analysis.