This is a very interesting article. I like that you are measuring actual behaviors, as opposed to sentiment. Question is whether these behaviors can be considered as accurate proxies, that indicate perceptions of DT.