jonathandroth
commented
9 months ago pretrends power reports the magnitude of the slope of a linear trend
for which power is 50%, so it is always positive by construction.
When visualizing hypothesized violations of trends, you can change the sign of the slope if that is appropriate in your context.
On Fri, Dec 1, 2023 at 11:59 AM Xuanli Zhu @.***> wrote:
Thank you very much for creating package.
My question is that in my practice, it seems that the generated linear pretrends at 50% or 80% are always upward-slopping curves. For those cases where the treatment effect is negative, shouldn't it be more reasonable to test a downward-slopping pretrend? Please correct me if I am wrong. Thanks!
— Reply to this email directly, view it on GitHub https://github.com/mcaceresb/stata-pretrends/issues/3, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFH757PCTHTNIDJUXLLYHIEFLAVCNFSM6AAAAABADD6CZGVHI2DSMVQWIX3LMV43ASLTON2WKOZSGAZDCMRVHE4TEMY . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Alalalalaki
Thank you very much for creating package.
My question is that in my practice, it seems that the generated linear pretrends at 50% or 80% are always upward-slopping curves. For those cases where the treatment effect is negative, shouldn't it be more reasonable to test a downward-slopping pretrend? Please correct me if I am wrong. Thanks!