The objective of this study was to examine globus pallidus internus deep brain stimulation (GPi-DBS) outcomes in primary and secondary dystonia, derived from blinded ratings using two scales and two raters. Twenty-five patients with variable presentations of dystonia were evaluated with videotaped standardized dystonia rating scales at preoperative baseline and at 6 and 12 months following GPi-DBS implantation. These 75 examination videos were retrospectively evaluated, independently and in random order, by two movement disorder neurologists who were blinded to the treatment status. Both neurologists scored each videotaped evaluation using the Burke-Fahn-Marsden Dystonia Rating Scale-motor part (BFMDRS-M) and the Unified Dystonia Rating Scale (UDRS). A final score for each video was assigned by averaging the raters' scores. An intra-class correlation coefficient was used to calculate inter-rater reliability. A linear mixed model was fitted to investigate the time effect and its interaction with type of dystonia (primary versus secondary) for each rating scale. Inter-rater reliability was excellent. Intraclass correlation coefficients ranged from 0.994 to 0.997 for both scales at baseline, 6 and 12 months. The average motor improvement scores after GPi DBS for the entire heterogeneous group of dystonia patients after 6 and 12 months of stimulation was 21.32% (p = 0.0010) and 28.95% (p = 0.0017), respectively, when the UDRS score was used. Similar levels of improvement 20.46% (p = 0.0055) at 6 months and 27.39% (p = 0.00197) at 12 months were found using the BFMDRS-M score. Analysis using unblinded scores from our database revealed a 32.99 and 37.27% UDRS improvement at 6 and 12 months, and an improvement in UDRS score of 38.5 and 43.7% when the analysis was limited to only primary dystonia. If the data were further segregated to include only cases of DYT-1 primary generalized dystonia, the UDRS benefit increased to 48.24%. Our primary dystonia group was diluted by the presence of both old- and young-onset patients, as well as focal, segmental and generalized dystonia. In conclusion, (1) evaluating motor outcomes of DBS therapy for dystonia using independent, randomized retrospective rating by blinded raters' results in lower improvement scores than when outcomes are rated by unblinded treating neurologists. Blinded methodology may be superior and might produce a more realistic assessment of motor outcomes after DBS in patients with dystonia; (2) outcomes were similar whether the BFMDRS-M or UDRS was utilized; (3) GPi-DBS was effective in treating sustained involuntary motor co-contractions in medication refractory dystonia patients, more so in primary dystonia.