Measurement Invariance of the CESD-R-10 Among Adolescents over the Transition from Paper to Online Administration

Despite the prevalence of online surveys, research validating paper-based scale measurement properties for online use is lacking. We assessed the measurement invariance of the 10-item CESD-R-10 scale in adolescents transitioning from paper to online administration. We analyzed 2-year linked data from 2,831 Canadian secondary school students during 2019/20 and 2020/21. Using structural equation modeling, we examined measurement invariance across configural, metric, scalar, and strict stages due to the COVID-19 pandemic shift to online surveys. Baseline mean depression score was 8.2 (SD = 5.9), with 33.7% (N = 953) reporting clinically relevant symptoms. We found measurement invariance between paper and online administration. The results support measurement invariance in males, while partial scalar invariance for females when thresholds of items 6 (I felt fearful) and item 9 (I felt lonely) are freely estimated between the two administrations. Our findings demonstrate that the CESD-R-10 remains invariant even with a shift in data collection mode from paper to online surveys. This study supports the validity of conducting meaningful comparisons of depression symptoms when changing the mode of data collection in survey research among adolescent samples.