Hi. Really good explanation of the procedure. I am not sure how to report the result of this test. I would provide the t-value of the regression and the probability associated with the intercept, but what about the degrees of freedom? Any comment will be highly appreciated. Thanks, Alf
That’s great, I have tried one Different example but it didn’t work well. This new way is weaker and less reliable than the paired-sample T test. Because the R squared is what determines the power of the model not the standard error. Look at the R square of the model from the model summery its only 0.289, which mean the Time1 centered can only explain 29% of the difference b/w the times. On the other hand, at the paired Samples the Correlation between the variable is strong. Meaning they are highly related. Can’t understand how this works 😎
Thanks for you comments. I'm also not entirely sure how this regression t-test works, but it does have a linear mixed-effects modeling sense about it. Consequently, I suspect the properties of the covariance matrix may influence the power of the test (so, not only the covariance between time 1 and time 2, but also the ratio of the variances). Again, though, I'm not sure. I'll have to research it more.
Hello, I have to do a cox proportional hazard model and build a prediction model: can you please give a lecture on the cox proportional hazard model here on youtube? Thank you. * Literature* Norton N, Olson RM, Perez EA. Association studies of Fcγ receptor polymorphisms with outcome in HER2+ breast cancer patients treated with trastuzumab in NCCTG (Alliance) Trial N9831. Cancer immunol Res 2014;2(10): 962-969
Hi, I want to mention you in the acknowledgement section of my thesis but haven't been able to found your name and affiliation. I'll be happy if you could text me those. :)
Hi. Really good explanation of the procedure. I am not sure how to report the result of this test. I would provide the t-value of the regression and the probability associated with the intercept, but what about the degrees of freedom? Any comment will be highly appreciated. Thanks, Alf
That’s great, I have tried one Different example but it didn’t work well. This new way is weaker and less reliable than the paired-sample T test.
Because the R squared is what determines the power of the model not the standard error. Look at the R square of the model from the model summery its only 0.289, which mean the Time1 centered can only explain 29% of the difference b/w the times.
On the other hand, at the paired Samples the Correlation between the variable is strong. Meaning they are highly related. Can’t understand how this works 😎
Thanks for you comments. I'm also not entirely sure how this regression t-test works, but it does have a linear mixed-effects modeling sense about it. Consequently, I suspect the properties of the covariance matrix may influence the power of the test (so, not only the covariance between time 1 and time 2, but also the ratio of the variances). Again, though, I'm not sure. I'll have to research it more.
Does the std error not force the impression of power? Can you expand the context of std error to remove or limit bias
شكرا أستاذ Thank you Prof.
Could you make a video on how to use Mixed Models in SPSS please! You are a very good presenter!
I have the slides from my epidemiology master.
Thanks Professor!!!
Hello, I have to do a cox proportional hazard model and build a prediction model: can you please give a lecture on the cox proportional hazard model here on youtube? Thank you.
* Literature*
Norton N, Olson RM, Perez EA. Association studies of Fcγ receptor polymorphisms with outcome in HER2+ breast cancer patients treated with trastuzumab in NCCTG (Alliance) Trial N9831. Cancer immunol Res 2014;2(10): 962-969
Hi, I want to mention you in the acknowledgement section of my thesis but haven't been able to found your name and affiliation. I'll be happy if you could text me those. :)
Can I get your email plz? I need a paid little help on SPSS. Thank you.