Raloxifene increases bone mineral density (BMD) and decreases vertebral fracture risk; the effects on quantitative ultrasound (QUS) variables, however, have been less well studied. We aimed to further evaluate the effectiveness of QUS for monitoring raloxifene treatment and withdrawal effects. Osteopenic, postmenopausal women (age = 50–80 yr, n = 125), who completed a 96-wk study (phase A) evaluating treatment compliance or monitoring, were invited to participate in a 96-wk raloxifene withdrawal study (phase B). Those originally receiving treatment were then randomized to continue on raloxifene (60 mg/d) + calcium (500 mg/d) (n = 23) or to discontinue raloxifene and take placebo + calcium (500 mg/d) (n = 23). Previously untreated women remained untreated (n = 12). Yearly QUS and BMD measurements were performed. At the end of phase A, lumbar spine BMD ( p = 0.005), amplitude-dependent speed of sound (Ad-SoS) ( p = 0.006) and average SoS ( p = 0.040) decreased in untreated women but remained stable in treated women. Significant changes in Ad-SoS and ultrasonic bone profiler index had occurred in treated women by the end of phase B ( p < 0.01). All variables, except bone transmission time, were higher for those receiving any raloxifene treatment ( p < 0.05). Until further knowledge has been acquired, QUS measurement variables should only be used in conjunction with BMD when assessing changes in bone because of raloxifene therapy.