UBC Theses and Dissertations
On estimating the causal effects in interrupted time series design with phase-in period and gradual implementation : a simulation study Zheng, Tianyi
Interrupted time series design has been widely applied to assess the causal effectiveness of an intervention. This thesis investigates potential estimation and modelling challenges that arise from interrupted time series (ITS) studies. Specially, for studies that may involve delayed effects and phase-in periods. A phase-in period is a special form of delayed intervention effect where the full effect occurs sometime after intervention. Through a simulation study, this thesis proposes various applicable analytical strategies for phase-in periods and highlights the different casual effects that they referred to. This thesis concludes that multiple counterfactual assumptions may exist, and different analytical strategies lead to different causal effects. Researchers should be cautious about forming their counterfactual assumptions and picking the appropriate analytical strategy accordingly.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International