sorry, my question wasnt clear. I am wondering how you back transform data where you have applied a constant to each point before log transforming. How do you back transform this? do you minus 1 from the back transformed mean? and does it affect your CIs? many thanks for your responses, you're so helpful!
Thank you very much for your prompt reply, much appreciated. So I will just simply multiply it with itself and after that I will get the descriptive statistics for the mean and CI?
Thanks for your sharing, it is very helpful, but may I know how to cite your works or where can I get the related information on these back-transform techniques? Thank you in advanced!
thank you! I cant find an example of how to back transform anywhere, much appreciated. However, what if you had negative values in your original data? Also if you did natural log, would the number be: 2:718281? many thanks!
thank you so much for this video,i got a question about transforming back a negative skewed variable and 0 skewd variable.I want to test the normality of my variable , wich contains some number negative and other 0,for exemple var(35,60,45,0,-20,-18,3),what to do in this case: log (var+1) or log((max+1)-var).and I have transformed the result before or after the construction of my modelle, know that I want to build a discriminant model z score. please help me please
Is it possible to back transform the standard deviation and standard error? If no, isn't it weird to then report a, for example, independent sample t-test without the right standard deviation?
Do you know how to back transform the STDDEV and SE, or do you know where I can find an explanation on how to do it? I have been searching all over the net and only found statements that back-transformation of the SE and STDDEV is not possible or no statement about it at all..
Would this method of back transformation work for data that was transformed using the reflective method, which you use to transform negatively skewed data?
Hello, I got a question about transforming back a negative skewed variable, to transform this I used log10((max+1) - variable) = for my data log10(11-variable). I used this tranformed variable in my regression analyses and my β=2.427. How do I tranform this β back?
Thank you for this video. I am just confused as to how you arrived at .334 when you mention in the video .344 when calculating the log transformed CI?
Great, thank you for your quick reply. much appreciated!
First, thank you very much for your helpful videos and all your effort in them. My question is that how do we report square root transformation?
sorry, my question wasnt clear. I am wondering how you back transform data where you have applied a constant to each point before log transforming. How do you back transform this? do you minus 1 from the back transformed mean? and does it affect your CIs? many thanks for your responses, you're so helpful!
Thank you very much for your prompt reply, much appreciated. So I will just simply multiply it with itself and after that I will get the descriptive statistics for the mean and CI?
i wanna report the standard deviation beside the mean of variables. how can i back standard deviation result to the normal standard deviation?
thanks for sharing!
Thanks for your sharing, it is very helpful, but may I know how to cite your works or where can I get the related information on these back-transform techniques? Thank you in advanced!
You can cite the channel name as you would a website, per APA guidelines.
thank you! I cant find an example of how to back transform anywhere, much appreciated. However, what if you had negative values in your original data? Also if you did natural log, would the number be: 2:718281? many thanks!
thank you so much for this video,i got a question about transforming back a negative skewed variable and 0 skewd variable.I want to test the normality of my variable , wich contains some number negative and other 0,for exemple var(35,60,45,0,-20,-18,3),what to do in this case: log (var+1) or log((max+1)-var).and I have transformed the result before or after the construction of my modelle, know that I want to build a discriminant model z score. please help me please
It depends on what type of skew is present.
***** in my analysis i have both types of skew
Is it possible to back transform the standard deviation and standard error? If no, isn't it weird to then report a, for example, independent sample t-test without the right standard deviation?
Do you know how to back transform the STDDEV and SE, or do you know where I can find an explanation on how to do it? I have been searching all over the net and only found statements that back-transformation of the SE and STDDEV is not possible or no statement about it at all..
Can we use the same back transformation formula for geometric mean? Thanks.
Would this method of back transformation work for data that was transformed using the reflective method, which you use to transform negatively skewed data?
Yes, it would. You would do the reverse of the steps in the reflective method.
Hello, I got a question about transforming back a negative skewed variable, to transform this I used log10((max+1) - variable) = for my data log10(11-variable). I used this tranformed variable in my regression analyses and my β=2.427. How do I tranform this β back?
If you back-transformed your data to the original units and then ran the regression you won't need to do anything to the B value.
in my analys i have both type of skew !!!!
114kha Then you will have to perform different techniques for each variable depending on the type of skew.