I don't think it's a straw man. If it is true that the delta matters and it is also true that at the time when this metric was showing the most positive results and trending upwards, online communities such as this talk about the existence of the loneliness epidemic https://news.ycombinator.com/item?id=20468767 then one must ask oneself whether this is a property of the online communities in question.

At the time when the gallup poll showed an upward trend towards its peak this community was talking about the loneliness epidemic. When the gallup poll shows a downward trend toward its lowest, this community is talking about the loneliness epidemic. And it's the change in satisfaction that is the most significant. So there are two changes in opposite directions causing the same conclusion.

If this were happening to me, I would ask myself "Am I sure this is a general property and not just a property of me?". Do you find this not convincing to move your estimate of the likelihood of the loneliness epidemic actually existing? If you don't, it's all right. We can leave it here.