There is an even simpler answer to your question than the ones other, very good, comments are pointing out. All of the scientists who unquestioningly believed that dogs could tell the difference didn't do this study. All of the scientists who unquestioningly believed dogs couldn't, didn't do this study. All studies are done by whichever members of the scientist population don't believe the affirmative or the negative.

I don’t think this is right either.

All studies are done by scientists who are desperate to get published and have their work referenced.

Even a cursory glance at that corner of funny-cute-etc-animals YouTube videos will convince that animals, especially mammals, have at least very complex internal and social lives.

This study isn't treating dogs as if they were not conscious or intelligent.

It's isolating a particular aspect of intelligence to determine if it exists in a particular fashion. In this case, the question of if and how dogs are aware of human agency. This is a very interesting question.

The exact same falsifying process is used with studies about humans. Using these processes we've significantly changed our understanding of human cognition in the last century so it seems like it's working!

No one's suggesting that dogs or people aren't intelligent. The question is exactly how this intelligence functions, because there are many ways to reach a particular conclusion.