What's wrong with "the results are not significance different"? Assuming your methods are correct you've demonstrated that 'defence enzyme activity' (whatever that is) doesn't have an effect. That's a useful and worthwhile finding.

Your goal isn't to find a statistically significant difference (at least, it shouldn't be). It's to find out 'the truth'.

I do not understand. Changing the scale of a measurement should not affect the statistical findings regardless of whether you are talking about a meta-analysis, t-test, one-way ANOVA, etc.

What's wrong with "the results are not significance different"? Assuming your methods are correct you've demonstrated that 'defence enzyme activity' (whatever that is) doesn't have an effect. That's a useful and worthwhile finding.

Your goal isn't to find a statistically significant difference (at least, it shouldn't be). It's to find out 'the truth'.

Changing scale will not affect findings. Findings will be the same even if you change the scale in as much as you are using the same date set. In Meta-analysis, everything should be clear because you may have to publish the study along side the forest plots and that can really tell if you're using valid studies. However, the purpose of any research is to find out results, whether significant or not. So if you didn't find a significant result, you shouldn't be bothered because that is an objective finding.

Gordon L Warren Musa Sani Dan-Azumi Andrew Cook

Thanks too much for your advise, yes you are all correct. actually,

the result of each single study through forest plot is upregulated mostly (not all). but the result of forest plot showed NS as I explained to you. (the raw data from treatment and the control are very close ). if this is correct then I can present the result like this.

what is your advise ?

Saba, I think you have been given excellent advice already in the 3 answers above. Report the results as calculated, regardless of statistical significance.