Note the focus on preferences and self-declarations, not behavior.
This is about how people self-described when they feel alert and productive. It doesn't necessarily mean there aren't benefits (or conversely, detriments) to "early to bed, early to rise."
It's possible that the work of those who feel the most alert in the morning is more physical -- for example, manual labor. Maybe they have less of a need and opportunity for cognition, so their scores are poorer. Maybe, too, the work makes them feel differently by evening from how night owls feel.
It's also possible that those with a high need for cognition tend towards work that can be done at any time of the day or night, and they are tired in the morning because they (whether optimally or not) stayed up late in the night working on it.
So, a follow-up question is how this relates to behavior. If night owls consistently go to bed around the time of peak melatonin production and wake up early, do their scores improve? Do their preferences change? And the same question in the inverse for larks.
