To draft the results for a meta-analysis, you need to systematically report your study selection process, summarize the characteristics of the included studies, present the synthesized effect sizes using forest plots, and address both statistical heterogeneity and publication bias.
Writing the results section of a meta-analysis or systematic review can feel overwhelming because of the sheer volume of data. By following a structured approach—and adhering to standard reporting guidelines like PRISMA—you can present your findings clearly and logically to your readers.
1. Outline the Study Selection Process
Always begin your results by detailing how you arrived at your final pool of papers. State the total number of initial records found during your literature search, the number of duplicates removed, and the number of full-text articles excluded (along with the reasons for exclusion). This narrative text should directly correspond to your PRISMA flow diagram, which is a standard figure expected in published meta-analyses.
2. Summarize Study Characteristics
Next, provide a descriptive overview of the included studies. You should include a comprehensive "Table of Study Characteristics" that outlines authors, publication years, sample sizes, participant demographics, and study designs. To speed up this tedious extraction phase, you can use WisPaper's My Library to organize your final PDFs and chat with your uploaded papers via AI to instantly pull specific demographic data or methodologies without re-reading every document. Accompany this table with a brief narrative paragraph highlighting the most common traits and any notable differences across the studies.
3. Present the Main Effect Sizes
This is the core of your meta-analysis results. Report the overall pooled estimate (such as Cohen's d, Hedges' g, or odds ratios) along with its 95% confidence interval and p-value. You must also reference your forest plot here. Walk the reader through the forest plot by highlighting the overall trend and weight of the studies, rather than listing the specific results of every single individual paper in the text.
4. Report Heterogeneity and Subgroup Analyses
You need to tell the reader whether the results of the included studies were consistent. Report your heterogeneity statistics, typically using Cochran’s Q and the I² statistic. If you find high heterogeneity (e.g., an I² over 50%), report the findings of your subgroup analyses or meta-regressions to explain what variables (like age group, geographic location, or intervention type) might be causing the variance.
5. Address Publication Bias and Sensitivity
Finally, evaluate the robustness of your findings. Describe the results of your publication bias assessments by referencing a funnel plot and reporting the results of statistical checks like Egger’s test or the trim-and-fill method. Conclude the section by mentioning any sensitivity analyses, such as a "leave-one-out" analysis, to prove that your overall pooled effect isn't being skewed by a single outlier study.
