Good morning,

Did the "percent" papers have anything in common or provide some salient reason for the change?  If not, I would convert the values back using the inverse formula and see how they measure up to other means.  It sounds quite odd, kudos to you for catching it!

Rich

Dear Dr. Richman, and Dr.Richard,

Thanks for your response. 

@Dr.Richman - Thanks for giving your feedback on the unit conversion, suggest me if there is any other alternate ways to unify the units.

Regards

Vishwanath

I'd probably go with Richard's suggestion (i.e. try and convert your one % study to absolute values and then meta-analyse all as mean differences of absolute values). Mara's suggestion is (I believe) to use either of the common standardised mean difference formulae to combine all the results and I am not entirely sure that the assumptions for that approach apply (i.e. that the only contributing factor to variation in the standard deviations between studies is the use of different scales). Probably not too much difference between the options though, try both and see how/if they differ?

Dear Dr.James,

Thanks a lot for your inputs.. Will do both and see..

Regards

Vishwanath

It is not acceptable to convert the Mean ± SD values into %Mean ± %SD? Using %Mean = control mean/ treatment Mean*100 and % SD = SD of test/Mean of test*%Mean of test.

The mean is a measure of central tendency. The standard deviation is a measure of dispersion. You can have the same mean for a data set/population but with a very different SD and vice versa. As suggested by Dr.Richard, try and convert your % study to absolute values.

Dear Dr.Ravichandran,

Thanks for your inputs.

Regards

Vishwanath GL