Hi.
I have 2 studies that are causing me problems.
One has the same values pre and post - as such the changes in means is zero.
Another doesn't report the raw values, but states the change is means is o +/- 0.
However, RevMan doesn't compute zero as mean or SD.
What are the recomendations for this situations? CAn I insert 0.1 for instance?
Thanks
Jorge
Hey Jorge.
Mean can be 0 in RevMan.
SD is rarely 0. Perhaps add a very low SD?
Sincerely
Martin
Hey Martin Bjørn Stausholm . Thanks for replying.
Indeed RevMan accepts mean of 0 - it happened to be one of those cases where SD was also 0 and it wasn't computing it. Thanks!
It is the second recomendation that I have received for using 0.00001. However, is there any reference to support that decision? If don't support it, maybe I will have problems in the review process?
Regards
Jorge
Hey Jorge. I was too quick in my suggestion about imputing a low SD. I don't think it is a good idea. Sorry.
Is there other variance data such as Standard Error of Measurement, confidence interval, P-value or interquartil range? They can be used to estimate SD.
Eheheh
Pretty much nothing. I just know that P>0.05.
In the draft I am currently doing I have been using a lot of p intervals to estimate the SD. I have used the default values of P non significant = 0.051; p
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis