Childhood anxiety is the most prevalent mental health concern facing Canadian children but often goes untreated. Parent-administered, Internet-delivered cognitive behaviour therapy (ICBT) has been suggested as an approach to improve treatment access. For this approach to be effective, however, it is important to ensure that parents will use and remain engaged with the treatment intervention. ICBT researchers frequently include measures of engagement in efficacy studies, but the measures tend to be uni-dimensional and inconsistent across studies. Perski, Blandford, West, and Michie (2017) developed a conceptual model of engagement, designed to address limitations in the literature by capturing two broad dimensions of engagement: objective (i.e., intervention usage) and subjective (i.e., attention and interest). The aim of the present study was to use the Perski model to examine engagement within parent-administered ICBT. Ninety-one parents of children with anxiety participated in the nine-week, therapist-guided, parent-administered ICBT program, Anxiety treatment for Children through online Education (ACE). Measures of population characteristics (e.g., demographics; child and parent mental health; parent beliefs) and intervention features (e.g., credibility; satisfaction; working alliance) were administered throughout treatment. Measures of objective engagement (i.e., number of logins per week; number of messages sent by participant to their coach) and subjective engagement (i.e., the Digital Behaviour Change Interventions Engagement Scale) were collected at post-treatment. Forty-seven parents completed the entire intervention, while 44 dropped out. Analyses revealed that 20.3% of parents who dropped out did so during the first lesson, while the remaining parents dropped out at varied timepoints. There were no differences in population characteristics between families who ii dropped out early in the program and those who dropped out later. Higher pre-treatment child anxiety symptoms, lower pre-treatment parent mental health symptoms, and higher levels of parent education were associated with a greater likelihood of program completion. Primary analyses focused on engagement among parents who completed the entire intervention. Findings suggested that more negative parent beliefs about child anxiety predicted higher objective engagement but were also associated with higher child anxiety scores across treatment. Higher credibility and content satisfaction were associated with higher subjective engagement. The working alliance was the only factor associated with both objective and subjective engagement (i.e., a stronger working alliance predicted higher engagement). Importantly, child anxiety symptoms significantly improved from pre- to post-treatment. Neither objective nor subjective engagement was associated with treatment outcome, although this was likely due to methodological issues with the measurement of engagement. Findings have important implications for the development of parent-administered ICBT, highlighting a need to examine the working alliance, parent beliefs about anxiety, and content satisfaction as potential avenues for enhancing engagement and outcome. Findings also have implications for the systematic investigation of engagement in low-intensity interventions; most notably, findings underscore the importance of including multi-dimensional measures of engagement. Ultimately, the present study serves as a valuable step forward in understanding and maximizing engagement in low-intensity interventions, which can lead to the refinement of more accessible treatment options for children struggling with anxiety.