If you are just reading news (without performing a keyword search), you will see a score associated with each news section. A section score is an aggregated score of all news (texts) in that section. It basically tells you the overall sentiment score of all articles combined, so that you can have a feel about the general sentiment of that news domain. However, if you carry out a search, an aggregated keyword sentiment score will just look like a section score on top of the screen. Please remember, it is an aggregated score of all keyword sentiment scores from all relevant news, which is completely different from news (text) sentiment score (including all words, not just keywords). That is because there is a chosen keyword associated with the "section" (e.g. a search result and/or a self-created section).
The sentiment scores of a system provided section and a self-created section are very different in terms of calculation methods. However, they are equally useful in understanding and interpreting the sentiment of each collection of news.