Objective: Modifiable risk factors for chronic diseases are plausible targets for intervention. However, psychosocial risk factors, including markers of perceived and biological stress, are understudied. Mobile sensing data, which are continuously collected from a person’s naturalistic smartphone use, may be used to examine the effect of both acute and prolonged stressors. We conducted a pilot study to test a mobile sensing collection tool, and to validate naturalistic text collection from smartphones as an objective behavioral indicator of stress.Methods: We assessed 25 undergraduate students in 2016 - 2017 (mean age = 20.64 years, S.D. = 2.74, 13 males, 12 females, 13 men, 12 women). We collected affective text language use via a custom keyboard, self-reported questionnaires of depressive and anxious symptoms, perceived and subjective stress, sleep, and lifetime cumulative stress, and the biological stress markers salivary C-reactive protein (CRP) and interleukin-1β.Results: Due to the pilot and exploratory nature of the study, only effect sizes are reported. We observed large effect-size associations between three measures of affective language (total positive words, total negative words, and total affective words) and cumulative lifetime history of stressful events. We also observed a large inverse correlation between negative words and reported hours slept. There were medium effect-size associations between affective language measures and self-reported subjective and perceived acute stress, as well as CRP.Conclusions: The current study shows initial promise and justification for using mobile sensing measures, especially text, in future studies.