Inattention over time-on-task: the role of motivation in mitigating temporal increases in media multitasking

The study examines whether enhancing motivation can mitigate temporal increases in media multitasking and corresponding declines in sustained attention and performance as time-on-task increases. Traditional vigilance research shows performance deteriorates over prolonged tasks, often attributed to resource depletion. In modern digital contexts, attention lapses manifest as increased media multitasking during lectures or lab tasks. Emerging accounts (e.g., opportunity cost model) posit that attention shifts result from dynamic cost-benefit evaluations, with rising boredom and perceived alternative rewards over time. The authors hypothesized that increasing motivation would elevate the perceived value of the primary task, attenuating increases in media multitasking and performance declines over time.

The paper reviews classic vigilance decrement findings (Mackworth, 1948; Warm et al., 2008; Helton & Russell, 2011) demonstrating declines in attention/performance over time, alongside self-reported attention decreases (Risko et al., 2012; Thomson et al., 2014; Brosowsky et al., 2023). It surveys media multitasking dynamics during lectures and lab tasks, often increasing over time and impairing performance (Wammes et al., 2019; Drody et al., 2024; Ralph et al., 2020, 2021; Tam & Inzlicht, 2024). Non-linear fluctuations have been observed (Rosen et al., 2013; Ragan et al., 2014), suggesting context dependence. The opportunity costs account (Kurzban et al., 2013) proposes that as time-on-task increases, perceived value of alternatives rises (boredom as a signal), promoting shifts from exploitation to exploration (Wiradhany et al., 2021). Prior interventions to curb media multitasking (education, awareness, blocking access) show mixed success, possibly because they do not sufficiently alter cost-benefit variables. Motivation has been linked to improved attention and reduced mind wandering (Engelmann & Pessoa, 2007; Seli et al., 2015, 2019; Brosowsky et al., 2023). Ralph et al. (2021) showed motivating instructions (early exit contingent on performance) reduced media multitasking in aggregate, but temporal dynamics were not examined.

Across two studies, participants completed a 1-back sustained attention task with an optional, task-irrelevant video (TED talk: Brain Magic) they could turn on/off, operationalizing media multitasking as the number of trials with the video playing. Trials were partitioned into nine blocks of 52 trials (468 experimental trials; letters B, F, K, H, M, Q, R, X, Z; 500 ms stimulus, 2000 ms fixation). Performance metrics: proportion hits (correct target responses) and proportion false alarms (incorrect responses to non-targets). Motivation manipulation: Control received standard instructions; Motivated were told they could leave early if performance was sufficiently high (in reality, task length was 20 minutes for all). Motivation was probed via Likert scales. Study 1 (re-analysis of Ralph et al., 2021): Participants: 166 recruited (17–35 years; M_age=19.43, SD=2.47), lab setting; after pre-processing (exclude <30% hits or >20% false alarms), N=157 (78 Control; 79 Motivated). Measures: Pre- and post-task motivation (1–7 Likert); familiarity with video. Analyses: Media multitasking across blocks via Generalized Linear Mixed Model (GLMM) with Poisson distribution; fixed effects: Condition (Control reference), Block (centered); random intercepts and Block slopes per participant. Performance over time via mixed factorial ANOVAs (Condition × Block) for hits and false alarms; sphericity checked (Mauchly) and Greenhouse–Geisser corrections applied. Study 2 (online replication with in-the-moment motivation probes): Recruitment: 428 online participants (University of Waterloo SONA; 218 Motivated; 210 Control; 17–44 years, Mage=20.04, SD=3.36). Pre-processing yielded N=284 (154 Motivated; 130 Control; final age Mage=20.00, SD=3.40). Motivation assessed pre- and post-task and during task via nine pseudo-random thought probes (one per block; 1–7 Likert). Post-task measures: enjoyment and interest (1–7), and for Motivated group, perceived motivation due to early exit; post-task questions about engagement in other activities and whether they involved multimedia devices; familiarity with the video. Analyses mirrored Study 1: GLMM (Poisson) for media multitasking (Condition, Block, and their interaction; random intercepts/slopes per participant); mixed ANOVAs for hits and false alarms over blocks; sphericity testing and Greenhouse–Geisser corrections. Subsample analyses (Supplementary) excluded those reporting other activities or focused only on those who did.

Study 1: - Motivation: Mixed ANOVA showed main effect of Condition (F(1,155)=15.86, p<0.001, ηp²=0.09; Motivated > Control) and main effect of Time (Pre > Post; F(1,155)=9.21, p=0.003, ηp²=0.06); Condition×Time interaction ns (F=1.35, p=0.25). - Media multitasking: GLMM revealed a significant Block×Condition interaction (Estimate=-0.18, SE=0.08, z=-2.17, p=0.030), indicating more gradual increases in media multitasking over time in the Motivated group; main effects of Condition and Block were not significant. - Performance: Hits showed main effects of Condition (F(1,155)=10.81, p<0.001, ηp²=0.07; Motivated higher) and Block (F(5.55,860.04)=12.23, p<0.001, ηp²=0.07; decline over time), and Condition×Block interaction (F(5.55,860.04)=3.24, p=0.005, ηp²=0.02), with more gradual decline in the Motivated group. False alarms: no significant effects. Study 2: - Motivation (pre/post): Main effect of Time (F(1,282)=237.88, p<0.001, η²=0.46; decline from start to end), no significant Condition effect (F=3.26, p=0.072) or interaction (F=3.78, p=0.053). - Motivation (during task): Mixed ANOVA showed main effects of Condition (F(1,270)=10.60, p=0.001, ηp²=0.04; Motivated higher overall) and Block (F(3.71,1001.49)=138.13, p<0.001, ηp²=0.34; decline over time), and Condition×Block interaction (F(3.71,1001.49)=3.64, p=0.007, ηp²=0.01), with slower declines in the Motivated group. - Media multitasking: GLMM showed a significant Condition effect (Estimate=-2.34, SE=0.57, z=-4.11, p<0.001; Motivated less likely to multitask) and significant Block×Condition interaction (Estimate=-0.13, SE=0.06, z=-2.09, p=0.04), indicating more gradual increases over time in the Motivated group; Block main effect ns. - Performance: Hits showed main effects of Condition (F(1,282)=10.72, p=0.001, η≈0.04; Motivated higher) and Block (F(4.43,1249.87)=60.05, p<0.001, η≈0.18; decline over time), with Condition×Block interaction (F(4.43,1249.87)=5.77, p<0.001, η≈0.02), indicating more gradual decline in the Motivated group. False alarms: at floor; no significant effects. - Experiences: Enjoyment and interest did not differ between conditions (enjoyment: t(282)=0.24, p=0.809, d=0.03; interest: t(282)=0.34, p=0.731, d=0.04). Motivated participants reported being motivated by the early exit instructions (M=5.29/7, SD=1.79). - Outside activities: A substantial proportion reported engaging in other tasks during the online study; subsample analyses were broadly consistent but some effects were smaller or non-significant, suggesting subtlety of manipulation effects. Overall: Media multitasking tends to increase and performance to decrease over time-on-task; motivating instructions consistently attenuate these temporal trends.

The findings support the view that temporal changes in media multitasking and performance are closely linked to motivation dynamics. Increasing motivation elevates the perceived value of focusing on the primary task, leading to delayed or attenuated shifts towards alternative activities (e.g., video watching). This aligns with the opportunity costs account, where rising boredom and perceived rewards of alternatives drive attention away from the current task. The results suggest that sustained attention decrements may reflect dynamic cost-benefit evaluations rather than pure resource depletion. Motivational manipulations can mitigate the vigilance decrement manifested as increased media multitasking and declining performance, with implications for designing interventions in educational and occupational settings.

Across two studies, motivating instructions reduced overall media multitasking, slowed increases in media multitasking over time, and attenuated declines in task performance. Motivation itself decreased with time-on-task, but the decline was more gradual under motivating instructions. These findings indicate that boosting motivation can help sustain attention by increasing the perceived value of attending to the current task versus alternatives, consistent with cost-benefit models of effort and vigilance. Future work should examine diverse motivation manipulations (e.g., monetary incentives, gamification), their feasibility in real-world contexts (classrooms, workplaces), and the robustness of temporal effects across tasks, durations, and environments.

The paradigm involved a 20-minute 1-back task, differing from traditional, longer vigilance tasks with rare signals, limiting direct generalization. In the online study, many participants engaged in activities outside the experimental context, potentially diluting manipulation effects; subsample analyses showed similar patterns but smaller and sometimes non-significant effects, suggesting subtlety of the impact. Media multitasking patterns may be context-dependent (steady increases vs. non-linear fluctuations). The motivating manipulation (early exit contingent on performance) may not be feasible in settings requiring fixed presence (e.g., classrooms, workplaces). Overall, while patterns were consistent across lab and online contexts, the magnitude of effects and generalizability warrant further investigation.