Using a Telegram chatbot as cost-effective software infrastructure for ambulatory assessment studies with iOS and Android devices

Michael Barthelmäs,Johannes Keller,Marcel Killinger

doi:10.3758/s13428-020-01475-4

Abstract

In this work, we present an innovative and cost-effective approach to run ambulatory assessment (AA) studies on participants’ smartphones via Telegram Messenger. Our approach works both for Android and iOS devices. The population of potential participants in a given country or region consists of all individuals who (a) are in possession of a smartphone, (b) are willing to install Telegram Messenger, and (c) live in an environment providing constant connection to the Internet. In our new approach to AA, participants are asked to subscribe to a Telegram chatbot that provides them with links to brief surveys at specified points in time in their everyday lives via short notifications. We developed a user-friendly Python script that allows for the flexible editing of the chatbot’s settings, e.g., the number of surveys per day. All common survey software designed for mobile devices can be used to present surveys to participants. This means that data collection takes place exclusively via the selected survey software, not via Telegram. With our approach, AA studies can be carried out among iOS and Android users cost-effectively and reliably while data security is ensured. Initial data from a pilot study show that studies of this kind are feasible, and the procedure is accepted by participants. Our Python script is licensed under General Public License (GPLv3) and therefore freely available and editable: https://github.com/Raze97/Telegram-Survey-Bot

Highlights

Ambulatory assessment (AA) studies can test psychological theories in everyday life
We describe the sequence of actions using the example of a daily survey, i.e., a survey link that is sent to participants multiple times throughout a study
We describe the steps to set up an AA study when the server that controls the actions of the chatbot runs with Linux

Summary

Introduction

Ambulatory assessment (AA) studies can test psychological theories in everyday life. This offers the opportunity to evaluate the generalizability of theories and to examine boundary conditions that cannot be examined in the laboratory (cf Fahrenberg, Myrtek, Pawlik, & Perrez, 2007; Trull & Ebner-Priemer, 2014). AA studies are one important tool to improve the understanding of human experience and behavior by capturing self-reports (e.g., Engeser & Baumann, 2016), observations (e.g., Ellis-Davies, Sakkalou, Fowler, Hilbrink, & Gattis, M., 2012), as well as motoric (e.g., Tryon, 2005) and physiological signals (e.g., van Lier et al, 2020) in the field. AA studies that focus on self-reports are typically implemented via participants’ smartphones (e.g., Killingsworth & Gilbert, 2010).

Objectives

Methods

Conclusion