Dear friends, I am doing meta-analysis and the results of N2O emissions from different incubation experiments are reported in different units, shall I do unit conversion or does not need to made unit conversion??
With regards,
It depends on the effect size you are using in your meta-analysis; raw scores vs. standardized effects. If you are using raw scores (e.g., mean of N2O emissions in a specific population, mean difference between different groups), then you need to ensure that the scores have the same meaning (units). However, no conversion is necessary if you use standardized effect sizes (e.g., standardized mean difference, correlation coefficients). When using standardized effects, the meta-analytic results will be the same or very similar* with/without conversion.
*There could be minor differences due to rounding errors when making/avoiding a conversion.
If doing a meta-analysis it seems likely that your aim is to synthesize, compare, and/or contrast results. In all cases, it seems likely that it will help readers if the results that they look at are all measured in the same units if if they are all measures of the same thing. Unless I am missing something here the question could perhaps be rephrased - is there any reason NOT to convert all results onto a common scale? From what you have said I find it hard to imagine one.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis