In addition to the item content covering the range of mild to severe symptoms, it is also necessary to consider the response scale of the observed items. As an example, consider attention problems in the general population. Hay et al. (2007) compared the Australian Twin Behavior Rating Scale (ATBRS) and the Strength and Weaknesses of ADHD-Symptoms and Normal Behavior (SWAN) scale. The response format of the ATBRS items ranges from ‘never’ to ‘very often’, and resulted in observed skewed data in the Hay et al. study. The format of the SWAN items is designed to cover gradual difference both above and below average, and ranges from ‘far above average’ to ‘far below average’, and resulted in more normally distributed observed data. If the ATBRS and the SWAN data would be used for an analysis of the nature of the latent structure, then the results could be very different because the ATBRS neglects the gradual differences at the lower end of the normal distribution that are detected with the SWAN (Hay et al. 2007). If the response format of the observed
