Skip to main content
. Author manuscript; available in PMC: 2009 May 2.
Published in final edited form as: Stat Med. 2008 Nov 10;27(25):5195–5216. doi: 10.1002/sim.3361

Table III.

Ratio of standard errors for standard regression calibration (RC1), calibration study RC (RC2), moment reconstruction (MR) and multiple imputation (IM) relative to efficient regression calibration (ERC), as a function of correlations between the true covariate X and outcome variable Y and between X and the measured covariate W.

ρXY* ρXW RC1 RC2 ERC MR IM
0.20 0.20 1.41 1.41 1 1.35 1.32
0.20 0.40 1.10 2.36 1 2.29 2.12
0.20 0.60 1.04 3.81 1 3.69 3.04
0.20 0.80 1.01 6.61 1 6.40 4.00
0.40 0.20 2.36 1.10 1 0.95 0.93
0.40 0.40 1.41 1.41 1 1.21 1.14
0.40 0.60 1.16 1.99 1 1.72 1.50
0.40 0.80 1.05 3.21 1 2.77 1.94
0.60 0.20 3.81 1.04 1 0.76 0.76
0.60 0.40 1.99 1.16 1 0.85 0.84
0.60 0.60 1.41 1.41 1 1.03 0.99
0.60 0.80 1.15 2.04 1 1.49 1.24
0.80 0.20 6,61 1.01 1 0.74 0.74
0.80 0.40 3.21 1.05 1 0.77 0.77
0.80 0.60 2.04 1.15 1 0.84 0.83
0.80 0.80 1.41 1.41 1 1.03 0.93
*

ρXY is the correlation between X and Y.

ρXW is the correlation between X and W.