This study aims to explore how experienced English speaking raters score speaking responses in terms of the overall scoring style, strategies for improving scoring consistency, and interpretation and application of scoring rubric for each scoring area. For this purpose, this study conducted retrospective interviews with six experienced speaking raters and three inexperienced speaking raters of NEAT, in order to compare and contrast the two rater groups’ scoring behaviors. The participants were asked to verbally report how they decided the score of each scoring area and what strategies they employed to improve their scoring consistency. The main findings are as follows: First, the experienced raters employed note-taking strategy while they were listening to and scoring each response, in order to make a more accurate decision of the score for each scoring area. Second, the experience raters applied the so-called ‘Golden-Rule’ of NEAT effectively and consistently, while the inexperienced raters were hardly consistent in applying this rule. Third, the experienced raters scored each scoring area independently of the other areas by applying the absolute scoring criteria given for that area without being seriously affected by the overall impression of the response, while the inexperienced raters decided the scores of the five scoring areas interdependently. Lastly, it seems that inexperienced raters apply a narrow range of scale points, that is, give similar scores for most responses if they do not have a clear understanding of the scoring criteria. Based on the results, some suggestions are made for English speaking rater training.