Home Publications About

Messaging Study

We are looking for participants to evaluate auto-responses in mobile messaging. Participants would need to install an android application (CAR) on their smartphone which would predict availability and automatically send auto-responses to some incoming messages. An example can be seen below.

In addition to the auto-response seen above, there are other categories of responses as well such as sharing the ringer mode, sharing that the participant is in a scheduled event etc. The type of response that is sent would be based on the prediction using Machine Learning. Not all incoming messages would have an auto-response. Only new messaging sessions and those predicted as unavailable would be auto-responded to.

To participate in this research, you must:

  • Be 18 years or older
  • Own an Android device with Android 7+ (Nougat or greater)
  • Communicate frequently through SMS or messaging apps like WhatsApp, Signal, Facebook Messenger etc.

Participation in this study involves:

  • Installing and keeping an android application running in the background
  • Answering a questionnaire daily in the app
  • Taking an end-of-study survey and giving an interview

The study would run for two weeks. For the first week of the study, the application would collect and learn from the participant's messaging data. This does not include any messages exchanged or with whom, and only device usage data such as current activity, calendar event, ringer mode etc. would be used for learning. For the second week of the study, the application will then detect any incoming messages and try to predict participant's availability. If predicted as unavailable, an auto-response would be sent on the participant's behalf. During the second week, a questionnaire would be sent to the participant each day to evaluate the auto-responses sent that day. The questionnaire would take about a minute to complete each day.

Participants would be compensated up to $30 through a virtual MasterCard based on their participation. Please contact the researcher for more information about the payment process.

Some other important points

  • The sent auto-responses also include a survey link which is for the recipients of the auto-response.
  • Other than the initial setup and daily questionnaire, the application requires minimal interaction
  • No auto-responses would be sent to group conversations, if that happens please let the researcher know
  • The application does not support Slack, please let the researcher know if you use Slack on your phone along with your Android version.
  • In some cases, the message notification might get removed if an auto-response was sent. In those cases, a new notification would be generated to inform the participant of the auto-response. Tapping that notification will open the application to which the auto-responses is sent.

You can find the complete list of all permission needed by the application as well as all the data collected by the application on the sidebar. No personally identifiable information would be collected other than the form filled at the start of the study, which would also be used for contact and payment purposes. No messages or information about participant's contacts would be sent to the server.

For any questions, comments or concerns and to participate in the study please contact the researcher at pranut@cs.pitt.edu.