We introduce a new approach to monitoring the activity of smartphone users based on their physical interactions with the interface. Typical events are taps, scrolling and typing, carried out to interact with apps. As compared to other measures, this directly encapsulates potential problematic physical smartphone behaviour as a signal. The approach contrasts against conventions such as self-reporting or timing activity sessions, and it focusses on active rather than passive smartphone activity. Using this alternative method, we collected all user interface interaction events from a sample of 64 participants over a period of 8 weeks, using a bespoke monitoring app called Tymer. User Smartphone Addiction was seen to significantly correlate with high levels of interaction with Lifestyle apps, particularly for female users. Interactions with Social apps in general were also associated with Smartphone Addiction. In particular, user interactions with Snapchat correlated with Smartphone Addiction, represented across all types of interface interaction. This is significant given the widespread usage of Snapchat by teenagers, and we hypothesise that the app’s design provides a particularly strong pathway in support of Smartphone Addiction.

Introduction

Smartphone usage is now ubiquitous across much of the global population, with the smartphone offering a vast range of applications (apps) that help to extend human cognition (Clark & Chalmers, 1998). Typical daily usage levels are high (De-Sola Gutiérrez et al., 2016) and there is evidence of dependency and attachment to smartphone technology (Hoffner, Lee, & Park, 2016), combined with the potential disruptiveness of incoming notifications (Turner et al., 2015, 2017) and “checking habit” formation (Oulasvirta, Rattenbury, Ma, & Raita, 2012). It is now acknowledged that such uncontrolled and problematic behaviour can become harmful, being recognised as Smartphone Addiction (SA) (Pearson & Hussain, 2016). Problematic smartphone use has been linked to teenage depression and anxiety (Ha, Chin, Park, Ryu, & Yu, 2008; Lemola et al., 2014), and more widely, various relationships have been found concerning stress, depression, sleeping problems, anxiety, subjective well-being, and loneliness (Demirci, Akgönül, & Akpinar, 2015; Elhai, Dvorak, Levine, & Hall, 2017; Lee, Chang, You, & Cheng, 2014; Lemola et al., 2014). Despite this research progress, detecting indicators of problematic smartphone behaviour is a challenge for two main reasons. Firstly, the range of utility that smartphone apps provide means that usage levels are generally high and that high usage is socially acceptable (Chotpitayasunondh and Douglas, 2016). Therefore behaviour correlating with SA can easily be hidden. Secondly, apps increasingly allow the smartphone to be passively used for large periods of time as a substitute for other devices (e.g., GPS navigation, TV, music player) meaning that as smartphone usage gets more diverse, high level metrics, such as time on the smartphone, may not represent the strongest indicator of problematic behaviour. These issues contribute to the invisibility of SA (Roberts et al., 2014) and the challenge of encouraging behaviour change to avoid it.