This study examined the differences in the use of {ne} and {je} in Korean dramas from the 1990s and 2010s in terms of gender and discourse situations. The research focuses on how male and female characters use {ne} and {je} differently in public and private contexts for each era, and how these patterns have changed over time. The findings are summarized as follows. First, in the 1990s, men primarily used {je} in public settings and alternated between {ne} and {je} in private contexts, indicating a context-dependent usage pattern. Women, on the other hand, mainly used {je} in both public and private situations, but {ye} appeared more frequently in public settings, showing that women also adjusted their usage according to the context. Second, in the 2010s, men used both {ne} and {je} regardless of the discourse context, while women predominantly used {ne} across contexts. Third, compared to the 1990s, the use of {ne} in public situations increased for both men and women in the 2010s, indicating a notable change, while no change was observed in private settings.

Given the view of tests as mechanisms within a language policy framework (Shohamy, 2006), Korean proficiency tests are not only used as a way to measure language knowledge, but rather as tools to impose national ideologies about langage use and diversity. The field of Korean language testing needs to embrace sociopolitical dimensions, and engage what is going on value and consequence in test development and validation. By drawing upon a combined approach of Messick’s (1989) validity framework, Fairclough’s (2001) Critical Discourse Analysis (CDA), Bachman and Palmer’s Assessment Use Arugment (AUA), this study aims to illustrate how the practice of testing can be discursively interpreted as a multilayered phenomenon, constituted through discourse. It discussed the applicability of AUA’s two claims (decisions, consequences) to the validation of Korean language proficiency or related certification testing for immigrants with foci of value implications and consequences. The interconnectedness of test validation and CDA is modelled through the procedure of Faircough’s (2001) analytic methodology. In an effort to illustrate that language testing is discursively value-laded, this study not only offer a theoretical and methodological addition to the current inquiry of test validation, but also re-emphasize that language testing is sociopolitically driven from a discursive angle.