Psychometric Evaluation of a Contextualized Literacy Instrument for Primary Students in Papua

Abstract

National literacy tests often fail in Indonesia’s 3T regions (Frontier, Remote, and Disadvantaged) because of cultural bias and limited contextual relevance, resulting in distorted measurements of children’s actual abilities. This study addresses that gap by psychometrically evaluating a contextual literacy instrument composed of images and texts, designed specifically for primary school children in Papua. The instrument assesses functional literacy skills—literal, inferential, and evaluative—using familiar local contexts to enhance readability and engagement. This is among the first literacy assessments in Papua to integrate both visual and textual elements with systematic psychometric validation, thereby directly challenging the urban bias embedded in national literacy assessments. A descriptive quantitative methodology was employed with sixth-grade pupils in Papua, examining item validity, reliability, difficulty levels, and discriminative power. Results show that 18 of 20 items demonstrated strong validity (r > 0.30), with a KR-20 reliability coefficient of 0.79, indicating satisfactory internal consistency. The difficulty distribution was 65% moderate, 20% easy, and 15% difficult, while 70% of items effectively distinguished between high- and low-performing pupils. Incorporating local visual and contextual features enhanced item relevance without compromising psychometric quality. These findings suggest that the instrument is effective for assessing literacy in low-performing regions such as Papua, although results remain preliminary due to the small-scale pilot with only 32 students.