Hello,
In case I conclude that a well-known construct should be measured formatively as it actually tests how often I anticipate different kinds of interpersonal misconduct like yelling, ignoration etc which form an incivility measure ( i.e a perception that I was treated in an uncivil manner), can I change it to a formative measure without testing for convergent validity? I understand the need to test convergent validity for newly developed scales , yet is it the case also for existing scales which I only want to measure in a more adequate manner which fits the scale better?
Thank you,