Hi dears, I am trying to conduct a meta-analysis on the prevalence of stunting among adolescents. However, I have faced difficulties in selecting the population for the study. Because, the studies are different in some aspects. For example, some studies are conducted among school adolescents, only early adolescents , late adolescents, adolescents in the community, among adolescents with HIV etc. This shows that the study population is different among studies. Therefore, can I pool the results of these studies even though there are special populations among the studies?
Many thanks dear Henry!
However, in sub group analysis still there is a very high statistical herogeneity I2 more than 90%. Therefore, can I proceed irrespective of the outcome of I squired values?
Ayelign Mengesha Kassie, I would recommend the website of Cochrane Training (http://training.cochrane.org/), where you can find an excellent course "Exploring heterogeneity" (see link below).
http://training.cochrane.org/authors/intervention-reviews/core-topics/heterogeneity
Statistical heterogeneity in your results does not mean to abandon MA but that you should interpret your results with caution. An answer to the question of how to explore or to interpret high heterogeneity is given in the above mentioned Cochrane training course.
I hope this will help you.
It is better to include all population group and you can do sub-group analysis, category based on school adolescents, only early adolescents, late adolescents, adolescents in the community, among adolescents with HIV etc.
I appreciate Halder's answer above.
By doing subgroup analysis, you will find out the individual effect based on each population &, you will appreciate the different findings or nearly same. then, your discusion will be clear & easy
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis