In order to do a meta-analysis for a particular topic of interest, it is necessary to extract data from past studies. The purpose of my study is to generalize the climatic trend magnitude across geographical space. However, I'm facing a lot of problems regarding data extraction. What data from the literature should I extract? Is that the Z values or the amount of climatic trend magnitude only or both? I'm expecting a practical answer.
Thank You
You need to account for bias in methodology, design and duration of the different studies before you combine the data. Try to find out what data the studies have in common and use appropriate tests for comparison. A z test is good if the data is normal in distribution. Also take into account effect size estimation, confidence intervals estimation and power analysis.
See: Comparing and Combining Data across Studies: Alternatives to Significance Testing, Eduardo Fernandez-Duque, Oikos Vol. 79, No. 3 (Sep., 1997), pp. 616-618, DOI: 10.2307/3546906
You need to account for bias in methodology, design and duration of the different studies before you combine the data. Try to find out what data the studies have in common and use appropriate tests for comparison. A z test is good if the data is normal in distribution. Also take into account effect size estimation, confidence intervals estimation and power analysis.
See: Comparing and Combining Data across Studies: Alternatives to Significance Testing, Eduardo Fernandez-Duque, Oikos Vol. 79, No. 3 (Sep., 1997), pp. 616-618, DOI: 10.2307/3546906
@Henry Ratul Halder can you share any article regarding the exact procedure...
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis