Background:Adequately measuring resilience is important in order to support young people and children who may need to access resources through social work or educational settings. A widely accepted measure of youth resilience has been developed by Ungar and Liebenberg which has been shown to work within vulnerable youth [1]. While the measure is completed by the young person on paper, it has been designed to be worked through with a teacher or social worker in case further clarification is required. However, this method is time consuming and when faced with large groups of pupils who need assessing can be overwhelming for schools and practitioners. The current study assesses app software with a built-in avatar who can guide the young person through the assessment and its interpretation. Objective:The primary objective is to compare the reliability and psychometric properties of a mobile software app to a paper version of the Child and Youth Resilience measure (CYRM-28). Secondly, the study will assess the use of the CYRM-28 in a Scottish youth population (11-18 years).Methods:Following focus groups and discussion with teachers, social workers and young people, an avatar was developed by a software company and integrated into an android smartphone app designed to ask questions via the device’s inbuilt text-to-voice engine. Seven-hundred and fourteen students from two schools in North East Scotland completed either a paper version or app version of the CYRM-28. A cross-sectional design was used and students completed their allocated version twice, with a two-week period in between each testing. All participants could request clarification either from a guidance teacher (paper version) or from the in-built software glossary (app version). Results: Test and retest correlations showed that the app version performed better than the paper version of the questionnaire. Paper (r(303)=.81, p<.001, 95%CI [.77, .85]); App (r(413)=.84, p <.001, 95%CI [.79, .89]). Fisher’s r to z transformation found the difference in the correlations to be statistically significant, Z=-2.97, p <.01. Similarly, Cronbach’s alpha in both conditions was very high (app: α=.92; paper: α=.87). Such a high Cronbach’s alpha indicates there may be item redundancy. Ordinarily this would lead to a possible removal of highly correlated items, however the primary aim of the current study is a comparison of app delivery method over a pen-and-paper mode and therefore outside the parameters of this paper. This will be considered in the discussion. Fisher’s r to z transformation found the difference in the correlations to be statistically significant [Z=-3.69, p <.01]. A confirmatory factor analysis [2] supported the three-factor solution (individual, relational and contextual) and reported a good model fit (χ2 (15, N= 541) = 27.6, p=0.24). Conclusions:ALEX, an avatar with an integrated voice guide, increased reliability when measuring resilience compared to a paper version with teacher assistance. The CFA reports similar structure using the avatar when compared against the original validation.