By Deborah Zabarenko
WASHINGTON (Reuters) - Hate mornings, especially on Mondays? You may be surprised to know that much of the world doesn't share that grumpy feeling.
Twitter shows people are more cheerful in the morning, get gloomier as the day wears on and rebound in the evening, with a peak right before bedtime. They're also happier from December to late June, when days gradually lengthen in the Northern Hemisphere.
Using the micro-blogging site Twitter as a gauge of global sentiment, social scientists studied 509 million tweets from 2.4 million users in 84 countries between February 2008 and January 2010, according to research reported on Thursday in the journal Science.
Bad news, even if it happens to a complete stranger, brings Twitter users down, research by scientists from Cornell University suggests.
Twitter, the five-year-old site that lets users communicate 140-character posts, offers an unprecedented chance to study human behavior and social networks, the scientists said.
Those brief posts get down to the nitty-gritty, showing that Twitter users prefer bacon to sausage and Cheerios to Frosted Flakes, Cornell sociologist and co-author Michael Macy said in a telephone interview on Friday.
"Twitter offers an unprecedented opportunity for social and behavioral scientists to study social behavior and interaction in real time with high temporal granularity -- hour by hour, day by day, over the course of the year -- and to do this at population scale with millions of people all over the world," Macy said.
DON'T BLAME WORK
It may seem intuitive that most people's moods go downhill through the day, but Macy said previous studies have been inconclusive, possibly because they relied on unrepresentative samples, often college students.
The daily mood swing was the same on weekends as during the week, Macy said, "which suggests that it's not caused by work, because on the weekend most people are not working."
This rhythm was the same across the globe, he said, from India, Africa, Australia and New Zealand to Britain, Canada and the United States.
Going forward, Macy said his group will focus more on behavior and not just on feeling, and will "decompose" the mood into more specific emotions such as anxiety or depression, and even home in on such issues as references to body image.
To get a picture of the global mood, the scientists searched tweets for about 1,000 words that reliably indicate positive emotions -- "agree," "fantastic," "super" -- and negative ones, such as "afraid," "mad," and "panic."
The researchers considered up to 400 tweets from each person, and excluded those with fewer than 25 tweets.
The scientists noted a recent survey of U.S. Twitter users found they are 51 percent white, 24 percent African American and 17 percent Hispanic. People with college and advanced degrees and with higher household incomes show up in higher numbers. The researchers said this still makes their study more representative than earlier ones on college students.
In another study cited in Science, researchers at the University of Vermont created a timeline to track the last year in tweets, ending in early September 2011, showing the world's average happiness slid from last October.
The death of Osama bin Laden was the apparent low point, followed closely by the earthquake in Japan, the London riots in August, the 10th anniversary of the September 11 attacks and the death of Amy Winehouse.
Emotional highs, as tracked on Twitter, were the Christmas and New Year holidays, followed by Valentines Day, Thanksgiving and Easter. The wedding of Britain's Prince William to Kate Middleton was a bright spot, comparable to Easter and Father's Day.
(Editing by Michele Gershberg and Anthony Boadle)