Skip to main content
Figure 1 | BMC Medical Research Methodology

Figure 1

From: A perfect correlate does not a surrogate make

Figure 1

Graphical depiction of incorrect inference based on surrogage endpoints. The graph shows perfectly correlated results (namely a straight line) for the relationship between surrogate and true outcomes for a control group C and experimental group E. The mean surrogate outcome in the E group is smaller than the mean surrogate outcome in the C group . However the mean true outcome in the E group is larger than the mean true outcome in the C group, , yielding the opposite conclusion for the effect of experimental intervention.

Back to article page