Dec. 22, 2011 -- Twitter users may be less happy than they used to be, say University of Vermont scientists.
They analyzed billions of tweets over nearly three years and found that happiness is on a downward slide.
It's unusual for researchers to use Twitter to sense people's moods, though another recent study did so.
In the new study, scientists analyzed the word content of more than 4 billion tweets posted by 63 million Twitter users worldwide. The tweets went out between September 2008 and mid-September 2011. The tweets offer almost an instant look over the "collective shoulder of society" in near-real time, researcher Peter Dodds, PhD, says in a news release. He is an associate professor at the college of engineering and mathematical sciences at the University of Vermont in Burlington.
Volunteers read the tweets and then rated the "happiness" quality of the words in those tweets, using a scale from one (saddest) to nine (most happy).
For example, the word "laughter" got an average rating of 8.5, "food" scored a 7.44, "greed" came in at 3.06, and "terrorist" at 1.30.
The research is published online in the journal PLoS ONE.
Happiness in 140 Characters or Less
Scientists applied this sad-to-happy word rating scale and other mathematical formulas to the billions of tweets they collected. They observed that for happiness, "after a gradual upward trend that ran from January to April 2009 ... there was a gradual downward trend, accelerating somewhat over the first half of 2011," the researchers write.
There were drops in happiness during the bailout of the U.S. financial system in 2008 and when the H1N1flu pandemic hit in 2009. That same year, the death of singer Michael Jackson caused the largest single-day drop in happiness, while Osama bin Laden's death in 2011 resulted in the study's lowest happiness day.
The Twitter data provided other intriguing insights:
- Happiest day of the week: Saturday, closely followed by Friday and Sunday.
- Least happy day of the week: Tuesday.
The happiest hour of the day: 5-6 a.m. That's when tweets contained more positive words and fewer negative ones. Happiness sharply dropped until noon, and then had a more gradual decline until the day's average low point between 10 p.m. and 11 p.m.
Even the use of curse words in tweets had a predictable daily cycle: Swearing reached its high point around 1 a.m. and a low between 5 a.m. and 6 a.m.
"These patterns suggest a gradual, on-average, daily unraveling of the human mind," the researchers write. At least among those who tweet.