Table 2.

Simulation Results for Functional Regression Predicting Outcome from Mediator

Sample size	250	500	1000
Intercept β0
Bias	0.0277	0.0236	0.0301
Mean squared error	0.0592	0.0284	0.0145
Root mean squared error	0.2432	0.1685	0.1203
Mean estimated SE	0.2379	0.1668	0.1174
Coverage	0.9492	0.9504	0.9488
Treatment βx
Bias	−0.0161	−0.0176	−0.0241
Mean squared error	0.0835	0.0410	0.0203
Root mean squared error	0.2889	0.2024	0.1424
Mean estimated SE	0.2833	0.1986	0.1399
Coverage	0.9486	0.9434	0.9504
Mediator conditional effect function βM (t)
Bias	−0.0186	−0.0180	−0.0219
Mean squared error	0.1329	0.0651	0.0339
Root mean squared error	0.3645	0.2552	0.1841
Pointwise coverage	0.9603	0.9577	0.9552
Familywise coverage	0.8598	0.8492	0.8364