PURPOSE: This study aimed to examine whether a physical therapy intervention, designed to reduce pain and improve the neuromotor control resulted in greater improvements in stance-phase knee flexion during stair ambulation in individuals with patellofemoral pain. The relationship between changes in stance-phase knee flexion and changes in pain, disability, and onset timing of individual vasti activity was also examined. METHODS: Forty participants aged 40 yr or younger diagnosed with patellofemoral pain were randomly allocated to a physical therapy (N = 21) or placebo (N = 19) treatment group. Stance-phase knee flexion was measured in two dimensions using a PEAK movement analysis system during stair ambulation. Individuals were divided into those with improvements in onset of vastus medialis obliquus (VMO) activity relative to that of the vastus lateralis (VL) of more or less than 10 ms. RESULTS: Groups were similar at baseline. After the 6-wk intervention, individuals in the physical therapy group had significantly greater changes in knee flexion at heel strike (mean difference 4 degrees, 95% CI = 2-7 degrees) and peak stance-phase knee flexion (mean difference 9 degrees, 95% CI = 5-12 degrees) than those in the placebo group. No differences were noted during stair ascent. Individuals with greater change in the onset timing of the vasti had greater improvements in stance-phase knee flexion. Changes in usual pain in the week before testing and change in the vasti onset timing were independent predictors of change in stance-phase knee flexion during stair descent, together accounting for 27-40% of the variability in knee motion. CONCLUSIONS: Physical therapy intervention resulted in significantly greater changes in knee joint motion than a placebo treatment, and these changes in knee motion were partly related to changes in pain and changes in onset timing of the vasti.