This is the website that I want to use to combine standard deviation http://atozmath.com/CONM/Ch2_CombinedSD.aspx . There is a study that already did the combined SD. In my attempt to double check that the website is correct before using it to combine SD for my data which the author did not calculate it, I have noticed that each of the data that I put to combine the standard deviation on the website, does not correlate to the study which already calculated the combined SD. I thought that perhaps they had changed it to SEM, yet when I calculated the SEM, the figure was close but it was not the same. Please see file. Thanks so much.
You can follow the code or calculations shown here:
https://www.researchgate.net/post/Hi_I_have_Mean_SD_and_N_of_13_different_groups_I_wanna_combine_to_come_up_with_a_single_Mean_and_SD_what_software_calculator_can_I_use_Thanks
To run the R code, you can download R, or run the R code at: http://www.r-fiddle.org/#/
Note that the formula on your website uses N in the calculations, whereas the formula on in the ResearchGate discussion uses (N-1).
ReseachGate isn't letting me edit my response.
The answer I get with this method is a grand sd of 14.715.
Thank you for your response Dr Mangiafico.
Please, I was wondering why it is (N-1) here on ResearchGate. Does it mean that the website above is wrong. I have limited time and was hoping to get a software to do all of the combined SD for me as opposed to having to do it manually. It will also minimise errors.
Currently, I am still trying to understand the link that you sent to me above. I don't really do stats so its been challenging.
I assume that the N form would be used for populations, while the N-1 form is used for samples. The same distinction is used for the calculation of standard deviation. Unless you've measured every individual in the population, you would use the N-1 form.
As I mentioned above, you can use the R code in the other ResearchGate post. You can just copy and paste it into either the R-fiddle site, or into the console of the downloaded R software. Just change the numbers for n, mean, and sd in the first three lines.
Hi, I got this from Luis F. G. (Portugal) I am sure you would find it easy to use. His contact is also attached.
Good luck.
Thanks so much for your response Dr Mangiafico and Mohammed, I really appreciate it. On another note, please how can you work out the "difference in SD"?
As of now, I have tried to work it out using this website https://robertgrantstats.wordpress.com/2012/08/20/how-to-work-out-the-standard-deviation-of-differences-for-a-sample-size-calculation/
but it does not make sense to me. An online tool would be a great help and will help me minimise errors.
Thanks in advance
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling