I just did; I was replying to your exact question in the comment. Questions like "Instagram can make me laugh when I am feeling down" do not look like the kind of questions that a good independently designed survey would do; in general, the would tend to be more of the kind that focus on a concrete event and their surroundings - it may look like they loose the perspective, you are looking at the only bad time that person had, but, done across a lot of people, it is more informative (would statistically represent the most likely state for the population, even if it is totally wrong for some individuals).