Retweeted health messages may not be what the patient ordered

A study of the credibility of health messages on Twitter showed that credibility dips when doctors who have a large number of Twitter followers passed on messages, instead of composing their own tweets, said Ji Young Lee, a 2008 Penn State master's degree graduate in media studies.

When non-medical professionals with a lot of Twitter followers forward messages about health on Twitter, however, their followers tend to find those messages more credible.

"Our study results imply that people may perceive tweets and retweets differently depending on the source of the content," said Lee, who is now a doctoral student in communication at Ohio State. "They do care about whether a message is originally created by someone or retweeted by someone, as well as whether the source is a professional and popular."

A tweet is a message that is 140 characters or less that a user posts on Twitter, the popular microblogging site. When a Twitter user forwards a post from another person, it is called a retweet.

The study, which appears online ahead of its publication in the journal Health Communication, shows how people infer credibility and trust based on certain cues, said S. Shyam Sundar, distinguished professor of communications and co-director of Penn State's Media Effects Research Laboratory, who worked with Lee on the study.

"It does show that people are aware of all of these cues," said Sundar. "And they are likely to use all three cues -- bandwagon, authority and proximity -- when they're reviewing online health communications."

The authority cue indicates the source's reputation for expertise and bandwagon is a cue that suggests how popular the source is. Proximity refers to whether the content is original -- a tweet -- or forwarded information -- a retweet.

A total of 63 undergraduate college students took part in the study and were asked to follow the Twitter accounts of either a doctor with many followers, the same doctor with a few followers, a layperson with many followers, or a layperson with a few followers.

The researchers added information to the Twitter accounts of these four fictitious sources to suggest the cues. For instance, they added "MD" on account names to indicate that the sources were doctors. To signal that the source was popular, the researchers increased the number of followers on those accounts from 21 to 983.

Subjects found original tweets from the doctor with a large Twitter following to be the most credible. However, the perceived credibility decreased when that doctor retweeted a message from another doctor, according to the researchers.

The credibility increased when the popular layperson retweeted the same message from a doctor.

"In the social media universe, the number of followers that a layperson has seems to translate into trustworthiness," said Sundar. "While tweets are judged based on the expertise of the person tweeting them, retweets depend on the trustworthiness of the person forwarding the health messages."

Sundar said subjects received Twitter messages that were controversial statements about weight loss that were tweeted to them over a one-week period. For example, one tweet read, "Exercising less than one hour per day can help one lose weight. Exercising more than one hour increases appetite and results in weight gain."

The researchers used controversial statements to make sure subjects would challenge the credibility of the message.

"We selected controversial messages because one of the major outcome variables in our study was perceived credibility of the content," said Sundar. "The question of judging credibility of a particular message will be moot if the message is a well known truism."

The research could help health communications professionals and doctors how best to spread the word about health and medicine, as well as help raise awareness on how web users process online information, said Sundar.

"With many laypeople posting health information, there is a greater need for online users to assess the credibility of what they read," said Sundar. "And, as more and more people go online for health information, we want to learn how they use the unique cues present in online media, so that we can find the best ways to communicate health information."