When researchers at Cornell University discovered that mood levels are similar at corresponding times across the world, they didn't use conventional sources.
They didn't collect the data by conducting an online poll or distributing a survey by mail - they gathered the data by using Twitter.
Cornell graduate sociology student Scott Golder, 30, gathered about 530 million tweets from 84 countries and plugged them into a text-analysis program called Linguistic Inquiry and Word Count.
The program tracks key words from any text format such as email, poems or transcribed speeches, according to the program's website.
Researchers studied what percentage of people tweeted positive mood words such as "happy" and "excited" and what percentage of people tweeted words like "depressed" and "angry" at certain times of the day and week.
"People criticize Twitter for being inconsequential but ... social scientists have always wanted to study large numbers of people in their everyday lives," Golder said.
More people tweeted the word "relaxed" on the weekends, but the beginning of the workweek reported the highest percentage of tweets about being "depressed."
Ruth Michel, 22, a recent UF public relations alumna, said that, to her, Twitter is more of a professional opportunity to follow current trends than it is a personal site.
But her tweet on Tuesday, "Don't be annoying. Background music on a website is an immediate turnoff," still gives insight to her mood at 12:59 p.m.
Public relations junior Haley English, 20, said when she thought about it, she definitely can see a pattern from her friends' tweets.
"On Sundays there is always someone who posts something about feeling terrible and being hungover and wishing Chick-fil-A was open," she said.
UF Area Director for Social Psychology James Shepperd said he was not completely familiar with the study but said the main question to consider when generalizing social research data is the group generating the messages.
"That determines how far you can generalize the facts," Shepperd said.
Researchers in this study knew few specifics about the ages and lifestyles of people who use Twitter, Golder said.
It was not a random sample, but by collecting key words from tweets all across the world, it is diverse, he said.
"Even though there will always be a place for traditional research methods, there's now an opportunity to do really groundbreaking work through the Internet," Golder said.
