<![CDATA[Purpose – This study aims to compare self-report scores for verbal communication ethics with observed communication practices at one adab-based school and one mainstream school selected purposively. The comparison is limited to these two institutional contexts and is not intended to represent all adab-based or mainstream schools. Design/methods/approach – This study employed a convergent mixed-methods design with a comparative quantitative component. Quantitative and qualitative data were collected during overlapping periods, analyzed separately, and then integrated through a joint display at the interpretation stage. The analytical sample comprised 80 students in Grades IV through VI, with 40 students from each school. Data were obtained through a 50-item four-point Likert questionnaire, nonparticipant observations, and semi-structured interviews. Quantitative analysis included descriptive statistics, the Shapiro-Wilk test, and the Mann-Whitney U test, while observational and interview data were analyzed using descriptive thematic analysis. Findings – Students’ self-report scores were higher at the mainstream school (M = 179.50; SD = 14.394) than at the adab-based school (M = 167.15; SD = 17.696), U = 426.00, Z = -3.601, p < .001, r = .40. However, SBA-OBS documented requests for permission, low tone of voice, controlled verbal responses, and some coarse words; SU-OBS documented coarse words, raised tones, shouting, and polite expressions in certain situations. The qualitative findings extended and, in some respects, contradicted the ordering of the self-report scores, but did not permit cross-school frequency comparisons, causal inference, or claims of sustained behavioral change. Research implications – The findings indicate that self-reported ethical verbal communication behavior was not fully parallel with the communication practices observed at the two schools. Data integration is necessary to distinguish among self-reported behavior, observed practices, and institutional contexts. Interpretation is limited by the two-school sample, the absence of demographic matching, limited documentation of observations and data cleaning, and the absence of direct measurement of social desirability bias.]]>
