I am trying to make a risk assessment type scale that measures the likelihood/risk of someone experiencing a psychological disorder in a setting. There are about 10 factors (e.g., gender, sleep quality, self-esteem, resilience, loneliness) that increase the likelihood and I would like to include these in the new scale. There are existing validated scales that examine each of these factors in the literature, but these are relatively long and including all these scales would make the new scale too long.

My plan at the moment is to pick 4-5 best questions (for confirmatory factor analysis later on) from each of these existing scales and adapt them to suit the setting I am interested in, and run the new scale with their full-length counterparts. As 11 scales (10 existing and 1 new one) would be difficult to complete in one sitting, I was thinking of breaking it down to 2 or 3 different testing sessions, preferably with the same participants, but if not possible, with different participants. Is this a reasonable approach to take?

I am aware that there are many books/articles about scale developments (e.g., Scale development: theory and applications by DeVellis), but I feel that these either target single construct (unidimensional) or multidimensional construct in which the subscales sort of “form” the multidimensional construct (e.g., personality with openness, neuroticism, agreeableness, conscientiousness and extraversion), and it is difficult to find the information on the type of scale I am trying to make. If anybody could recommend a good reference for the type of scale I am trying make, it would be most appreciated.

More Marcel Juventin's questions See All
Similar questions and discussions