It depends on what the objective is. You are certainly entitled to make choices based on your own personal experiences but when we're talking about broad social science dynamics around the country or around the world, you would be in grave error to use your personal experience as a stand in for everyone else.

For example, the actual divorce rate in Canada is about 50%, not nearly everyone, as you've experienced. And, basing the beliefs of all women off of YouTubes meant to enrage you so that you watch more and they make more money is hardly an actual metric of anything either. You're right, bland responses aren't all that interesting or saleable. But it still matters if you are trying to understand and actual social science dynamic to not just listen to YouTubers for your data. That's why researchers learn a whole host of data gathering and analysis techniques meant to minimize pre-existing bias so that we can see the truer picture.