PMAP 8521 • Example: Matching and IPW with R: 5: Inverse probability weighting

Поделиться
HTML-код
  • Опубликовано: 29 янв 2025

Комментарии • 22

  • @Op.Dr.HakanAnil
    @Op.Dr.HakanAnil 11 дней назад

    Thank you for your excellent video. I have watched many videos on implementing propensity score models in R, but yours was by far the most clear and understandable. I truly appreciate it.

  • @TheYepyep3
    @TheYepyep3 3 года назад +5

    Your videos are extremely helpful, thank you so much for making these!

  • @muhammedhadedy4570
    @muhammedhadedy4570 2 года назад

    Wow. That's amazing. This is propaply the best tutorial in explaining propensity score analysis on RUclips.
    Thank you so much for your work. You literally saved my life.
    :D :D

  • @amanielkefleyesus3236
    @amanielkefleyesus3236 2 года назад +1

    Kudos+++ for this clear and well-demonstrated concept of matching and IPW ! Thanks

  • @kjevlevigen
    @kjevlevigen 3 года назад

    Fantastic procrastination before my Epidemiologi exam! Thank you :-)

  • @paolatello8083
    @paolatello8083 3 года назад +1

    I'm veryyyyyy happy with your explanation! Keep doing this videos!

  • @noahcat9064
    @noahcat9064 3 года назад +1

    This is super helpful! Thanks for the video!

  • @gordiballester
    @gordiballester 3 года назад

    Really helpful. Greetings from Spain, Viva España.

  • @yeobin_mingyu
    @yeobin_mingyu 3 года назад

    Thank you SO much sincerely ! 감사합니다 !

  • @kenkoonwong2166
    @kenkoonwong2166 3 года назад +1

    after calculating IPW, do we need to add the confounding variables back to lm to control after the weight has been provided? Thank you for the explanation. the video is very helpful.

  • @TheBjjninja
    @TheBjjninja Год назад

    Thanks you. Around 9 minutes in, I would probably use mutate(ps= predict(model, type='reponse'). Much easier and less typing to just use the model and run predict()

  • @florencenyaguthii5022
    @florencenyaguthii5022 3 года назад

    Thanks for this video very helpful

  • @munthirmansour
    @munthirmansour 3 года назад

    what is the package used for ipw

  • @dr.kingschultz
    @dr.kingschultz 2 года назад

    Very good video

  • @azulchincarini8924
    @azulchincarini8924 2 года назад

    Is there the possibility of combining differences in differences with matching propensity score?

  • @dataanalytics3001
    @dataanalytics3001 3 года назад

    If there are various estimates from adjusted regression, PS, and IPW all of them are significant and in the same direction. Is there any statistical tests to decide which one is correct or closer to the truth?

  • @taiwankyh
    @taiwankyh 3 года назад

    Thanks so much

  • @trueperson22
    @trueperson22 2 года назад

    Can we just add family=binomial without link="logit" at 3:27?
    Thanks a lot for this great lecture

    • @phamnguyenductin
      @phamnguyenductin 7 месяцев назад

      Yes, the default link function for family = binomial is the logit function. Only if we intend to use another link function do we need to specify otherwise.

  • @priyankaroy3686
    @priyankaroy3686 Год назад

    extremely useful video for me. Thank you. However, I would like to ask how did you convert net into a new numeric column called net_num? I have a variable with "yes" and "no" which i had converted into factors with "as.factor" function. But that is not working for generating IPWs. please help

  • @austinpower14
    @austinpower14 2 года назад

    Do anyone know what to do if you have a multinomial distribution? I have 4 potential outcomes.

  • @caviper1
    @caviper1 Год назад

    Gracias