"Only 8% [of interrupted time series design studies] reported confidence intervals (CIs) or standard errors (SEs) along with the relative slope effects" 🫠 https://t.co/pRPgSFkdS2
555 followers
473 followers
@ZFusfeld I’d recommend a mixture model or another longitudinal design. Typically you’d want >1 data point for pre period if using ITS https://t.co/U6jozyEZGH
1,608 followers
@dgurdasani1 This paper has serious methodology issues. I advise the authors to follow this as a guide/checklist https://t.co/KssoOghxus to fit and report time-series (TS) models. Currently I can see no discussion of autocorrelation and stationarity, the t
3,838 followers
RT @spenafajuri: What do you think of this quote #epitwitter? "Interrupted time series (ITS) is considered one of the strongest quasi-expe…