<![CDATA[This study examined gender differences in students' prosocial behavior using Rasch measurement. Although gender differences in prosocial behavior have been widely investigated, previous studies have primarily relied on conventional statistical approaches, providing limited evidence regarding individual response patterns and item functioning. A quantitative cross-sectional survey was conducted involving 526 vocational high school students in West Sumatra, Indonesia, selected through proportionate stratified random sampling. Data were collected using a validated 24-item Likert-type Prosocial Behavior Scale and analyzed using the Rasch Rating Scale Model in Winsteps. Rasch analysis included psychometric evaluation, Wright Map analysis, item hierarchy, and Differential Item Functioning (DIF) across gender groups. The instrument demonstrated excellent psychometric properties, with person reliability of 0.93, item reliability of 0.98, person separation of 3.58, and item separation of 6.83, confirming high measurement precision. The Wright Map revealed substantial heterogeneity in students' prosocial behavior, indicating considerable variation in prosocial tendencies despite generally moderate-to-high person measures. The item hierarchy showed that behaviors involving pure altruism, personal sacrifice, and emotional concern were more difficult to endorse than simple situational helping behaviors. Only one item exhibited significant DIF, indicating that female students were more likely than male students to endorse helping without expecting anything in return. These findings demonstrate that gender differences in prosocial behavior are limited to specific behavioral indicators rather than reflecting broad differences across the construct. The study highlights the value of Rasch measurement in identifying behavioral heterogeneity and gender-related item functioning and provides practical implications for gender-responsive guidance and counseling services, character education, and school-based social-emotional learning programs.]]>
