Background. There is limited evidence for the effects of constraint-induced movement therapy (CIMT) in the early stages of stroke recovery. Objective. To evaluate the effect of a modified CIMT within 4 weeks poststroke. Methods. This single-blinded randomized multisite trial investigated the effects of CIMT in 47 individuals who had experienced a stroke in the preceding 26 days. Patients were allocated to a CIMT or a usual care (control) group. The CIMT program was 3 h/d over 10 consecutive working days, with mitt use on the unaffected arm for up to 90% of waking hours. The follow-up time was 6 months. The primary outcome was the Wolf Motor Function test (WMFT) score. Secondary outcomes were the Fugl-Meyer upper-extremity motor score, Nine-Hole Peg test (NHPT) score, the arm use ratio, and the Stroke Impact Scale. Analyses of covariance with adjustment for baseline values were used to assess differences between the groups. Results. After treatment, the mean timed WMFT score was significantly better in the CIMT group compared with the control group. Moreover, posttreatment dexterity, as tested with the NHPT, was significantly better in the CIMT group, whereas the other test results were similar in both the groups. At the 6-month follow-up, the 2 groups showed no significant difference in arm impairment, function, or use in daily activities. Conclusions. Despite a favorable effect of CIMT on timed movement measures immediately after treatment, significant effects were not found after 6 months.