Research undertaken with athletes has shown that lower-evaluated feedback is related to low self-efficacy levels. However, the relationship between teacher feedback and self-efficacy has not been studied in the dance setting. In sports or dance contexts, very few studies have manipulated feedback content to examine its impact on performers' self-efficacy in relation to the execution of a specific movement. Therefore, the aim of this investigation was to explore the effect of manipulated upper, lower, and accurate grade feedback on changes in dancers' self-efficacy levels for the execution of the "Zapateado" (a flamenco foot movement). Sixty-one students (56 female, 5 male, ages 13 to 22 ± 3.25 years) from a Spanish dance conservatory participated in this experimental study. They were randomly divided into four feedback groups: 1. upper-evaluated, 2. objective and informational, 3. lower-evaluated, and 4. no feedback-control. Participants performed three trials during a 1-hour session and completed questionnaires tapping self-efficacy pre-feedback and post-feedback. After each trial, teachers (who were confederates in the study) were first asked to rate their perception of each dancer's competence level at performing the movement according to conventional criteria (scores from 0 to 10). The results were then manipulated, and students accurate, lower-evaluated, or upper-evaluated scores were given. Those in the accurate feedback group reported positive change in self-efficacy, whereas those in the lower-evaluated group showed no significant change in self-efficacy during the course of the trial. Findings call into question the common perception among teachers that it can be motivating to provide students with inaccurate feedback that indicates that the students' performance level is much better or much worse than they actually perceive it to be. Self-efficacy appears most likely to increase in students when feedback is accurate.