From an evolutionary point of view, it seems that the quadrant 1 and 3 messaging mediums (see chart) represent a form of r/K selection classification. Each trades between quantity and quality. The focus in quadrant 1 is to increase the quality of the content with higher expense per message. The focus in quadrant 3 is to increase the quantity of the content with a corresponding reduction in quality and lower expense per message.
Algorithms for determining text sentiment, theme, writer’s sex, age, and education are only effective on large and well formatted compositions. They were designed to work on structured essays that are around 1000 words long. The likelihood of accurately determining any of these characteristic from a 140 character tweet or a blog posting that is riddled with expressions like LOL or OMG is as good as a coin toss.