This was a great video - thank you! Only nit is that the brackets at 14:29 are a little misleading - to echo the comment below, you may want to clarify the following: ln(E(y)) - ln(t_i) = ln(E(y) / t_i) = ln(y/10000) = -3.29665
Very clear tutorial! I understand what poisson regression is now Quick question: By doing all those different tests in the end, isn't that a bit fishing? Just doing all kind of different tests and in the end report the one that show the most interesting results?
im a bit unclear on exactly what you are referring to, but here are some general comments that may help. first, you should definitely avoid "fishing" and just randomly poking around and reporting the most significant result. but in building a model and examining a research question, you often want to hypothesize some relationship, and then test to see if it is there statistically. you may also hypothesize an interaction/effect modification, and then wish to test statistically is it is significant (as well as biologically plausible). so, it is ok to be doing a series of tests, etc, as long as they are considerations you have made ahead of time, and you are checking to see if they are there statistically...but it is not a good practice to just "test everything" and then report on that...there should be some conceptual basis for what you are doing. hope that helps...
I dont mean to be so offtopic but does someone know a method to log back into an Instagram account?? I stupidly lost my password. I appreciate any assistance you can give me!
@Dean Cruz Thanks for your reply. I got to the site thru google and Im trying it out now. Seems to take quite some time so I will get back to you later with my results.
Amazing tutorial! But I think there is a little mistake in the formula for the connect rates to counts part. The formula should be ln(E(yi))-ln(ti)=ln(E(yi)/ti) instead of ln(E(yi))-ln(ti)=ln(E(yi))/ln(ti)
if Poisson regression (at basics) is a method to model counts then how is person-time coming into play here. For example: what if I want to run Poisson regression for a die throw. The counts on the upper face of the die could also be modelled this way right? In die throw there is no person-time involved. So, my basic question is - how is Poisson regression transitioning from simple models to these person-time involving models. Pls let me know. Thanks
Can you kindly add the Overdispersion and excessive zeroes and two Part videos of Poisson regression of Mary Clare Kennedy also a part of the video lecture series of Poisson regression then it will be helpful for viewers.
Clear & straightforward. Thank you
This was a great video - thank you!
Only nit is that the brackets at 14:29 are a little misleading - to echo the comment below, you may want to clarify the following:
ln(E(y)) - ln(t_i) = ln(E(y) / t_i) = ln(y/10000) = -3.29665
same question
Good explanation. What is the name of the data set used in the presentation?
amazing class!
12:20 how to get that confidence interval?
Does anyone know if the R code is provided for this series? I really appreciate the help!
Thank you for the tutorial - very clear and concise. How do you get access to the R code and data files?
can we have access to the database used?
Very clear tutorial! I understand what poisson regression is now
Quick question: By doing all those different tests in the end, isn't that a bit fishing? Just doing all kind of different tests and in the end report the one that show the most interesting results?
im a bit unclear on exactly what you are referring to, but here are some general comments that may help. first, you should definitely avoid "fishing" and just randomly poking around and reporting the most significant result. but in building a model and examining a research question, you often want to hypothesize some relationship, and then test to see if it is there statistically. you may also hypothesize an interaction/effect modification, and then wish to test statistically is it is significant (as well as biologically plausible).
so, it is ok to be doing a series of tests, etc, as long as they are considerations you have made ahead of time, and you are checking to see if they are there statistically...but it is not a good practice to just "test everything" and then report on that...there should be some conceptual basis for what you are doing.
hope that helps...
Explanation is AMAZING! Thank you very much! :)
I dont mean to be so offtopic but does someone know a method to log back into an Instagram account??
I stupidly lost my password. I appreciate any assistance you can give me!
@Maddux Kade instablaster :)
@Dean Cruz Thanks for your reply. I got to the site thru google and Im trying it out now.
Seems to take quite some time so I will get back to you later with my results.
@Dean Cruz It worked and I actually got access to my account again. I am so happy!
Thanks so much you saved my account :D
@Maddux Kade you are welcome =)
Thanks, great explanation
Amazing tutorial! But I think there is a little mistake in the formula for the connect rates to counts part. The formula should be ln(E(yi))-ln(ti)=ln(E(yi)/ti) instead of ln(E(yi))-ln(ti)=ln(E(yi))/ln(ti)
Hi Qianwen, actually this is not a mistake. And both of you are correct. You can write it either.
Thanks for the tutorial
where Can I get the data analysed here?
if Poisson regression (at basics) is a method to model counts then how is person-time coming into play here. For example: what if I want to run Poisson regression for a die throw. The counts on the upper face of the die could also be modelled this way right? In die throw there is no person-time involved. So, my basic question is - how is Poisson regression transitioning from simple models to these person-time involving models. Pls let me know. Thanks
Phenomenally clear. thanks
I loved this. Thanks!
You are welcome Kevin! glad you found the video useful
very clear, thank you
Sorry can you please give me the R code for the crude rate? I can't find it
Can i use poisson regression for the data containing cases and deaths due to a disease year wise
thank you for this video!
This is very good!
simple to follow and didn't waste any of my time! I just want to take this chance to say thank you and have a wonderful day! You are amazing
Where we can get the database?
Can you kindly add the Overdispersion and excessive zeroes and two Part videos of Poisson regression of Mary Clare Kennedy also a part of the video lecture series of Poisson regression then it will be helpful for viewers.
All of the tutorial videos in the series are there
amazing!
Thank you!
I doubt that those kind of STATISTICS MODELS COULD BE USEFUL IN PARTICULAR WHEN YOU HAVE TO UNDERSTAND THE EFFICACY OF PHARMAS
these are exactly the kinds of models used for that.