Take Your Medicine: Write Online!

By on

Decades of research have shown that writing about our deepest, darkest emotions makes us feel better. But what happens when we write about those feelings online?

Patients with life-threatening illnesses must regularly process heavy emotions. Many patients and their caregivers turn to online health communities to get support and share their journeys. We partnered with CaringBridge.org – a large journaling platform that lets patients write about their health journey – to explore how expressive writing affects people’s engagement with their online community.

Haiwei demonstrates posting a journal update on CaringBridge.


We know that when people first join a community, their early experiences affect how long they stick around. We hypothesized that people who write more expressively will stay longer. But how do we quantify expressive writing?

Fortunately, past research gives us a few hints. In older experiments, researchers instructed patients to write expressively using pen and paper. It was almost like a prescription: “Write down your deepest emotions and thoughts, three times per month, with a minimum of 15 minutes per session.”

These studies found that patients who “took their medicine” (i.e. followed the instructions) had better health outcomes. In particular, the more they wrote, the better they felt. Also, the more emotional words (like “happy”, “sad”, etc.) and the more cognitive words (like “accept”, “think”, and “know”) they used, the better the outcomes.

In our study, we quantified these characteristics during users’ first month as CaringBridge members. We measured how frequently they wrote journal updates, how long those updates were, and what types of emotional and cognitive words they were using. Using techniques from Machine Learning, Natural Language Processing, and Statistics, we collected some fascinating results about how the behavior of expressive writing predicts that CaringBridge users will stay engaged on the site longer.

We were surprised to find out that emotional words don’t seem to impact user engagement. In fact, positive emotional words don’t affect outcomes, while using more negative emotion words is (slightly) harmful to continued engagement. On the flip side, using cognitive words helps engagement. The strongest predictor, however, is “taking a bigger dose of medicine” (i.e. writing longer and more frequently).

Check out our full CSCW 2018 paper [PDF File], “Write for Life: Persisting in Online Health Communities with Expressive Writing and Social Support.” And next time you want to engage community members, consider designing mechanisms that encourage expressive writing online!


Ramen is More Photogenic than Chicken Wings: A Winter Break Externship Report

By on

GroupLens externs with some of their favorite foods.


Guest written by Maryam Hedayati, Steph Herbers, Sophia Maymudes, and Anna Meyer.


Christmas is almost here. Do you know what most people won’t be doing on December 25th? Writing online restaurant reviews. Let’s dive deeper into the world of online restaurant reviews to learn more about this and other interesting trends. (more…)

SqueezeBands: Hugging Through the Screen

By on

A woman raises her hand towards a webcam during a videochat with a friend. Her hand is encased in a cloth device with shape memory alloy springs.
Lucy and Jackie demonstrate using SqueezeBands to send a high five! The camera detects mutual gestures like this one and creates a sensation of touch by squeezing and heating each person’s hand band.


When I Skype with my family, I really wish that I could reach through the screen to give them a hug! Instead, we sometimes have to pretend—we lean forward “hugging” the monitor or bring our hands towards the camera to do a virtual “high five.” What if you could actually feel some of that touch instead of just having to imagine it?


Your feelings of connecting to a group can predict your future behavior

By on

Maybe you’ve joined a group recently could be a Taekwondo group, a wine tasting club, a fantasy football league, or whatever. Do you know that how people felt “connected” to a group before they joined can predict their future behavior in the group? Social psychologists have identified two conceptually distinct ways a member can connect with a group — identity-based attachment (e.g., “I feel connected to the Taekwondo group because I started to learn Taekwondo when I was a kid!”) and bonds-based attachment (e.g., “I feel connected to the wine tasting club because my best friend Daniel is a club member!”) — and worked to understand their causes and consequences. What we have done is study how connections between a person and an online group can predict that person’s future behavior.


The more they try, the more they are likely to come back!!

By on

Are you trying to launch an online site for customers? Do you know that on an average, 60% of users do not return after using the site once?

In this research, we discover factors that predict whether first-time users return to MovieLens, our movie recommendation site.  A model based on these factors successfully predicts 70% of returning users (and non-returning ones).  Notably, the best single predictor of user return is the diversity of features explored in the user’s first session!  Along the way, we develop a process and a metric for activity diversity — one that can be applied to any site or context. Interested in further details?