The author’s argument and the analysis that he cites are both only about a person’s mood throughout a single day. The pattern of people’s moods throughout the whole week is irrelevant.

The author’s evidence only looks at those people who do use words associated with positive or negative moods on social media. The fact that many people don’t use these words is irrelevant.

The use of mood words in other forms of communication is irrelevant; the author only addresses language on social media. He does overlook the possibility that social media language isn’t necessarily indicative of people’s actual moods, but this is not what (C) points out.

The author doesn’t overlook this possibility. He’s focused on the frequency of positive mood words used in social media messages, not the number of messages posted at different times of the day.

If most evening messages are posted by different people than the morning messages, the author can't conclude that a person's mood follows the same pattern as the messages. The messages came from different people— this tells us nothing about one person’s mood throughout the day.