Assessing the Attitudes Towards Computing Scale: A Survey Validation Study

Students who have positive attitudes towards computing are more likely to intend to persist in computing and ultimately persist in computing. Thus, this study describes the development and assessment of the Attitudes Towards Computing Scale, which was developed based on Williams et al. [20] Computer Science Attitude Survey. Assessment of the survey involved testing its reliability, dimensionality, and validity. Internal consistency reliability of the subscales and the total scale were strong. However, dimensionality tests using factor analysis did not support a five-factor structure; rather, the factor analyses suggested either using one item per subscale or mean composites of the subscales to form a total score of attitudes towards computing. This suggests there may not be meaningful differences among subscales. Tests of criterion validity show that the short-form of the scale, using a mean composite score of either one item per subscale or the composites of the subscales themselves, predict students' computing knowledge and intentions to persist just as well as using the subscales. Furthermore, an example is shown why using the subscales may be problematic due to multicollinearity issues. Recommendations for improving scales, such as not using reverse-coded items and having a sufficient number of items that differentiate between subscales, are provided. Ultimately, this study provides the computer science field with a scale of attitudes towards computing that demonstrates moderate validity and offers an example of how to assess other scales used in computer science education.