By Alison Wood Brooks

While I was getting my PhD at the Wharton School, I came to a surprising realization: social scientists—people who spend their lives studying how human beings interact with each other—often don’t study real people interacting with each other. My reaction to this realization was something akin to: SAY WHAT?! Though psychologists, sociologists, and economists had mastered the art of survey design—asking individuals about their thoughts, feelings, attitudes, and beliefs in isolation—they hadn’t figured out how to get people to talk to each other while completing studies, except in logistically-challenging, face-to-face interactions in behavioral lab settings (e.g., asking participants to negotiate in a lab), in hard-to-control naturalistic settings in the world (e.g., observing people in a restaurant), or using confederates (e.g., using a pre-programmed set of responses from a fake person in a survey).

To me, the roadblock holding social scientists back was (relatively) straightforward: technology. Most social scientists today use survey platforms like Qualtrics to conduct their research, and they run many of their studies online. Survey platforms like Qualtrics have amazing, customizable features that allow researchers to do an enormous array of things. However, existing survey platforms do not allow for real human interaction.  

Rather than kvetching about this roadblock, I decided to create a solution. Enter a new resource called ChatPlat (www.chatplat.com). Short for “chat platform,” ChatPlat is a research tool I designed for my own research with the idea of sharing it with other researchers who also study human interaction. ChatPlat allows users to design, administer, record, and analyze chats between real people online.

Here’s how it works:

  1. Create an account at www.chatplat.com.
  2. Build ChatPlats. Each chatplat has many customizable options like: How many people do I want to chat in each group? How long should each chat last? Should the chatters be able to send emoticons to each other? Should they be able to see when their partners are typing? Do I want participants to interact with each other or with a research assistant/confederate? Do I want them to receive a warning message that the chat is about to end? Etc. Once you have designed a ChatPlat chat window according to your research needs, you can embed your ChatPlat into any survey (e.g., Qualtrics) using the simple HTML code provided.
  3. Run your study. Participants are automatically placed into chat groups based on their arrival time at the chatplat window. The first arriver is matched with the second arriver, etc. This matching protocol means that you can run chatstudies online, too (e.g., on MTurk), as long as you pay participants enough that arrival speed is sufficiently fast.
  4. When the study has ended, you can download all of the chat text data from your account at www.chatplat.com to analyze. Et voila! Real human interaction with the data to prove it.

I have used ChatPlat in many of my own research projects. I will describe three examples. First, in this paper published in Organizational Behavior and Human Decision Processes in 2011, I induced people to feel anxious (or neutral) by watching video clips, and then paired them to negotiate in a chatplat window I embedded in Qualtrics. We found that people who feel anxious tend to give lower first offers, exit the negotiation earlier, and earn less profit overall. Second, more recently, using ChatPlat, I paired hundreds of MTurk participants in groups of three to “get to know each other.” English-speakers from all over the world chatted for fifteen minutes, and we found that people who asked more questions during the chats were better-liked by their chat partners and learned more information overall. Third, to supplement data from Facebook about emoticon usage, I paired adults online to chat and measured their tendency to send each other emoticons during their conversation. Mirroring field data from Facebook, I found that sending more emoticons (and using a higher diversity of emoticons) causes people to feel happier and more satisfied with their lives.

Other researchers at major universities like UPenn, Stanford, Columbia, and the University of Minnesota are also using ChatPlat to conduct their research on topics like romantic relationship development, dyadic decision making, negotiations, group and teams, consumer behavior, and organizational behavior. If you are interested in human interaction, I urge you to check out a free one-week trial atwww.chatplat.com or to contact the ChatPlat team. I hope you love it as much as I do.


Alison Wood Brooks is an assistant professor of business administration in the Negotiation, Organizations & Markets Unit at Harvard Business School. She teaches Negotiation in the MBA elective curriculum, Micro Organizational Behavior in the PhD curriculum, and is affiliated with the Behavioral Insights Group at the Harvard Kennedy School’s Center for Public Leadership. In her research, Professor Brooks studies topics at the intersection of emotion, cognition, and behavior. Much of her research has focused on anxiety, one of the most pervasive emotions experienced in the workplace. Professor Brooks’ work on anxiety — and other emotions such as excitement, anger, calmness, envy, and admiration — has led her to examine how individuals navigate interpersonal interaction more broadly. For example, she also studies interpersonal topics such as trust, humor, authenticity, gender, advice seeking, and question asking. Her work highlights how people can regulate their emotions and interact with others to make themselves and their organizations more effective. ChatPlat is a for-profit business for which Allison is a founder and stakeholder. Follow Allison on twitter: @ProfAWBrooks