Image: People at Barack Obama's inauguration
A group of people celebrate the official start of Barack Obama's term as president on Inauguration Day in Washington, D.C. on Jan. 20. A happiness meter found that the overall happiest days of the last few years were election day (Nov. 4) and President Obama's inauguration (Jan. 20).
updated 7/31/2009 3:14:49 PM ET 2009-07-31T19:14:49

A "hedonometer," or a device that measures happiness, has been created by scientists in Vermont.

The software created by Peter Dodds and Chris Danforth collects sentences from blogs, and is now analyzing Tweets, to zero-in on the happiest and saddest days of the last few years.

"We wanted to capitalize on the explosion of blogs and now Twitter to build an instrument that would give us some measure of the emotional signal from a large collective of people," said Peter Dodds, a researcher at the University of Vermont and co-author of a new paper in the Journal of Happiness Studies.

"All this new data is basically helping us gain insight into social phenomena that didn't exist a few years ago."

The scientists started at the Web site, which combs through 2.3 million blogs looking for sentences that begin with "I feel" or "I am feeling."

The words that came next were ranked on a happiness scale of 1 to 9. A total of 1,034 words were ranked, with "triumphant" registering at 8.87 on one side of the scale and "hostage" measuring at 2.20 at the other end.

Using 10 million of these sentences from the last several years, the scientists calculated the general level of happiness for each day.

The consistently happiest days are, not surprisingly, vacation days and holidays. The overall happiest days of the last few years were election day (Nov. 4) and President Obama's inauguration (Jan. 20). On these days, people typically typed sentences with words like "pride" and "proud."

Some of the saddest days over the last few years have been the anniversaries of 9-11 and even the day before that solemn anniversary, Sept. 10. Michael Jackson's recent death also caused a drop in the average national happiness.

Analyzing blogs only measures the general climate of happiness across the entire United States (90 percent of the analyzed blogs were from the United States.) To find the level of happiness in your neck of the woods the researchers are now applying their software to 140-character Tweets.

"There are something like 1,000 tweets a minute," said Dobbs. "That's really a lot of data that we can analyze."

There are other ways to measure happiness. Many of them involve someone with a clipboard on a street asking something to the effect of "Are you happy today? Yes, No, or Unsure?" The method gathers data, but there well documented problems with the method; people tend to overestimate their level of happiness.

Using people's own writings largely avoids this on-the-spot bias, giving scientists another tool to understand the roots of happiness.

Still, there are some problems with the software. For example, a sentence that reads "I am not happy," will register as a happy statement. Also, most people who blog and Twitter are younger than the average population.

Regardless of these issues, the new software gives happiness researchers a new way to analyze happiness.

"I think this tool offers researchers some really fascinating opportunities to ask great questions," said Robert Biswas-Diener of Positive Psychology Services, LLC and author of a book on happiness. "It would be a great real-time measure that could chart the emotional chatter related to all kinds of real-world events."

At this point the research is more descriptive then prescriptive; it doesn't provide any clues as to how people can increase their happiness. A general increase in the happiness would be a good thing, says Dobbs, but people still need a bad day here and there to appreciate the good ones.

"We don't want people to be happy all the time; that would be Brave New World," said Dobbs. "We clearly need a balance of good and bad days to keep us healthy and balanced."

© 2012 Discovery Channel


Discussion comments


Most active discussions

  1. votes comments
  2. votes comments
  3. votes comments
  4. votes comments