CHALLENGE:
Marketers are increasingly using behavioral targeting and harvesting information on individuals found on the internet, such as Facebook and MySpace. But are social media participants creating false personas in cyberspace and are marketers creating invalid strategies due to their reliance on these data? Are marketers basing strategies on imagined personas? A study comparing Facebook personas with their real-world authors found a strong correlation between impressions created by personas and their real-world counterparts. Personas who were liked, were also liked in the real-world.
HOW THEY RESEARCHED IT:
Professors at the psychology department of Tufts University conducted the study comparing Facebook personas to real-life counterparts using 37 undergraduate volunteers. Undergrads were individually paired with one of six confederates in a four-minute “get-to-know-you” meeting. The confederates rated the undergraduate subjects on a number of dimensions related to likability. Undergraduates from another university viewed videos of the meetings and coded specific cues about the subjects’, non-verbal expressivity. Undergraduate subjects’ Facebook page was accessed by the researchers with the permission of the undergraduate subjects. Undergrads from a private university then rated the undergraduate subjects based on his or her Facebook page in terms of likability, and these ratings were combined to reveal a final “Facebook liking” score.
WHAT HAPPENED?:
Results from the survey indicate that impressions made through social media are similar to those attained through real-life interaction. The correlation between confederate liking and Facebook liking was statistically significant, and males and females were perceived similarly across all variables. Positive first impressions based on Facebook correlate to increased webpage expressivity and likewise positive first impression based on dynamic behavior correlate to increased non-verbal expressivity. In accordance with previous research the study finds no evidence of a linear relationship between self-disclosure and impressions of liking with relation to face-to-face interaction or personal web pages respectively. Other concluding evidence finds that online behavior is similar to that of personal interaction, for example people who disclose personal information and are expressive in face-to-face interactions tend to display the same traits on web pages. An interesting caveat of the study is that self-disclosure and expressivity are unrelated to one another, both online and in dynamic interactions. Overall concluding evidence shows that while there is some divergence between online and real-life personas, for the most part the two are very highly correlated.
WHY MANAGERS SHOULD CARE:
Marketers in today’s technologically driven business environment rely increasingly heavily on social media to provide a glimpse of what consumers want. The issue of whether or not social media accurately portrays consumer personalities is therefore an important concern for anyone using the Web to glean information about the consumer market. For marketers this study reveals positive news: searching Facebook sites is not for nothing. Nearly all Gen Y consumers update information on at least one social networking site, and older consumers are increasingly joining the trend. A high correlation between real-life and online personalities means that marketers can use the social networking sites to gather lifestyle information from consumers who may otherwise be unwilling to share, or instead of using costly and time consuming methods primary data collection methods. Social networking sites should not be the sole source of consumer research as there is some differential between dynamic interaction and online personas.
CAN YOU HELP?
Comment back by sharing how you have implemented data gained from social networking sites to improve marketing and promotional strategies. Does this study make you feel like you can rely on Facebook profiles or do you think more research is needed before you will rely on social website data? If you have been relying on social website profiles, have these profiles been efficacious?
From Max Weisbuch, Zorana Ivcevic, and Nalini Ambady “On being liked on the web and in the ‘‘real world”: Consistency in first impressions across personal webpages and spontaneous behavior,” Journal of Experimental Social Psychology 45 (2009) 573–576
Alternate Shot 
Interesting research but I wonder:
1. Would the results be different for non-college students? Older adults?
2. Similarity in likability and willingness to share info does not necessarily mean other attributes would be consistent, does it? The fact that Jack is equally willing to share some personal info online and in-person does not mean he is also equally likely to be candid about planned purchases, brand preferences or price-sensitivity.
I’d love to see research that compares stated brand preferences and purchasing behaviors off to online.
Good points.
1. Use of students for research has always been problematic in assessing the generalizability of published research. In this case, given that students are such a high and dispportionate share of Facebook subscribers, I’d give the researchers a break on sample generalizability. I’d be more concerned over their small sample.
2. You’re right that we cannot assume all personal trait inferences would be similarly consistent and as you suggest, the further we move from the traits researched in the reported study, the less certain we should be about generalizing all traits, such as purchase-related behaviors are part of another set of personal characteristics. I’d like to see more research on the validity of social network profile information, but we’re still early in the life cycle of such research. I’m sure we’ll see more publications soon.
What can we generalise from this study?
If we are charitable about the methodology, we can say that it suggests that US undergraduates when assessing video interviews of students group them in similar ways to other US undergraduates assessing Facebook pages of the same students.
The value of this study is that it could have been a disconfirming case. If personas are generally going to be similar online and offline, then it is plausible that they should be for students, when assessed by students, given that they will share a substantial set of common codes and are assumed to be heavy users of Facebook.
If the personas had turned out to be poorly associated, then this would have been an important finding. I think it is really useful that Tufts have published this as it means that it is at least worth looking at in a more scientific way. I am sure the people at Tufts would not try assert that this study implied any warranted assertabiltiy that the phenomena of online/offline accordance of personas existed outside the population of US undergraduates.
BTW, if I read the study correctly, 37 is the number of objects to be assessed, the sample sizes are the number of people who viewed the interviews and the number who viewed the Facebook profiles. Presumably the statistically significant result related to the similarity between the two sets of observations.