The purpose of this study was to validate a new questionnaire for assessing students' beliefs about the nature of science (BANOS). Existing instruments have limitations in terms of psychometric validity. A new questionnaire termed "BANOS" was developed to address such limitations. The BANOS is based on dimensions of the nature of science as a theoretical framework. The BANOS was administered to 860 Grade 12 students in Namibia, using the paper-and-pencil method. Data analysis employed reliability analysis, exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and parallel analysis. The reliability of the BANOS was [alpha] = 0.87. EFA revealed a final interpretable five-factor structure, and the factor solution accounted for 67.73% of the total variance. However, parallel analysis revealed that only four factors had eigenvalues that were statistically significant and the resultant scree plot also supported the retention of four factors. CFA results showed that the measurement model had a poor statistical fit for the data. These findings indicate that the eight-dimensional framework could not be confirmed at EFA level. However, the BANOS had adequate construct validity and reliability. Results are discussed in terms of intricate similarities among the dimensions of nature of science.