We examine the effects that may resolve the disagreement between strin
g-scale unification and low-energy data. ?These include string thresho
ld corrections, nonstandard hypercharge normalizations. Light supersym
metric thresholds, intermediate gauge structure, and extra matter beyo
nd the minimal supersymmetric standard model. We explicitly evaluate t
hese effects within the flipped SU(5), SO(6) X SO(3), and various SU(3
) X SU(2) X U(1) string models, and surprisingly find that the only wa
y to reconcile string unification with low-energy data is through cert
ain types of extra matter. Remarkably, many of the realistic string mo
dels contain precisely this required matter in their massless spectra.