ABSTRAKTujuan penelitian adalah menguji kecocokan model teoretikal kepekaan retorikal (rhetorical sensitivity) berbasis struktur mentalitas budaya komunikasi pada guru tersertifikasi dalam konteks masyarakat multietnis di Kabupaten Simalungun, Provinsi Sumatera Utara. Model penelitian terdiri dari 47 variabel manifes (indikator), enam variabel laten dan 15 lajur hubungan kausal. Penelitian berangkat dari paradigma klasik (positivis).Disain penelitian adalah cross-sectional survey.Data dikumpulkan melalui wawancara tatap muka. Alat pengumpul data yang utama adalah kuesioner.Jumlah sampel sebanyak 339 orang.Teknik penarikan sampel yang digunakan adalah pencuplikan acak sistematis. Marjin kesalahan pencuplikan sampel sebesar + /- 4 %. Data dianalisis dengan metode Structural Equation Modelling (SEM) dengan bantuan piranti lunak LISREL 8.5. Terdapat tiga temuan utama yang dihasilkan penelitian ini.Pertama, tiga ukuran kecocokan absolut model keseluruhan berkategori baik.Semua ukuran kecocokan inkremental model berkategori baik.Semua ukuran parsimoni model keseluruhan berkategori baik. Satu ukuran lain keseluruhan model berkategori baik. Kedua, 45 butir variabel manifest (indikator) memiliki validitas dan reliabilitas yang baik. Terakhir, terdapat 12 lajur hubungan kausal yang berkategori signifikan.Penelitian ini menyimpulkan bahwa keberlakuan model teoretikal kepekaan retorikal (rhetorical sensitivity) berbasis struktur mentalitas budaya komunikasi pada guru tersertifikasi dalam konteks masyarakat multietnis di Kabupaten Simalungun, Provinsi Sumatera Utara bersifat terbatas.

---

ABSTRACT

The main aim of this research is test how fit theoretical model of rhetorical sensitivity based on communication mentality structure on certified teachers in multiethnic context. This model research was shaped by 47 indicators (manifest variables), six latent variables and 15 causal paths. Paradigm used to this research is positivism. All concepts that building theoretical model are measured in interval levels and operationalized in self-report format with 7 points Likert-type scale. All indicators were translated from English to Indonesia. Unit and level analysis is individual. The target population is certified teachers as listed until 2013 at District of Simalungun, Province of North Sumatera. Sample size is 339 people that selecting by systematic random sampling technique with margin of error ± 4%. Sampling frame is available. This research design is a cross-sectional survey in which data was collected by face to face interview that guided by structured questionnaire. Data was analyzed by structural equation modelling (SEM) method that supporting with LISREL 8.5 for Windows.

This research found that three of six absolute fit measures were classified good fit. All incremental fit measures were classified good fit. All parsimonious fit measures were apparently classified good fit. There 45 0f 47 indicators were valid and reliable. Twelve of fifteen casual paths were significance. This research generally conclude that data which are taken from certified teachers in multiethnic context did not all fit to the theoretical model.