A good test instrument is required to measure students’ academic abilities. For final-year students, this ability is measured by USBN tests. This study aimed to analyses the characteristics of biology test items in the National Standardized School Examination (USBN) and analyses item tests categorized as difficult and poor. Type of study was descriptive exploratory with quantitative and qualitative approaches. Quantitative data analysis was performed using the classical test and item response theories with 1 logistic parameter model. Qualitative data analysis described why such question items are categorized as difficult and poor according to the classical test theory. The data were collected by using documentation of the answers of 193 students. The study results show that, according the classical test theory, Biology USBN item tests had the average difficulty index of 0.525 (moderate category) and the average of discrimination index obtained is equal to 0.124 (good category). Based on the item response theory, the mean difficulty index was -0.00029 (moderate category). The reliability test estimation result was 0.51, categorized as moderate. The qualitative analysis result showed that items categorized as difficult are the food web in the ecosystem, nervous system, and human excretion system.

This study aims to develop an instrument that is valid and reliable to measure the character of junior high school students. The developed instrument is a character assessment based on the school culture, used to facilitate teachers and schools in assessing student characters. This study is development research. The trial subjects involved in this study were 128 grade VIII students of state junior high schools in Bantul. The developed instrument was validated for the content by experts. The construct validity was analyzed using the confirmatory factor analysis (CFA), and the reliability was estimated using the omega reliability formula by McDonald. The result of the study is a character assessment instrument comprising six aspects with 12 indicators and 22 statements. The developed instrument has fulfilled the validity and reliability requirements. The content validity acquired an Aiken index of 0.87, while the CFA test acquired a loading factor score of each item being valid with λ 0.5 and t-value 1.96. The student character measurement fits the theorized model (χ2 = 230.719; (p) = 0.08; RMSEA = 0.040; and CFI = 0.964) and McDonald’s composite reliability estimation result 0.907. The score of students’ characters in the developing category were 63.3%, 18.8% for the early development category, 11.7% for the requiring assistance category, and 6.3% for the cultured category