Tutorial · 10 min

How to use Meta-Analysis Engine

Pooled effects, forest and funnel plots

Fixed and random effects models with DerSimonian-Laird and Paule-Mandel estimators, heterogeneity statistics, Egger's test and trial sequential analysis.

Before you start

  • Effect sizes with standard errors, or 2x2 counts per study

Steps

  1. 1

    Open Meta-Analysis Engine

    Sidebar → Meta-Analysis Engine.

  2. 2

    Enter or paste your study data

    One row per study. Extract Data and Literature Synthesis Tables can produce this table for you.

  3. 3

    Choose the model

    Random effects when studies differ in population or design; fixed effect only when they are near-identical.

  4. 4

    Read heterogeneity honestly

    Check I², tau² and the prediction interval before interpreting the pooled estimate.

  5. 5

    Check small-study effects

    Inspect the funnel plot and Egger's test, and use trial sequential analysis to see whether the evidence is conclusive.

  6. 6

    Export the outputs

    Download the forest and funnel plots plus the underlying numbers as PDF and CSV.

Tips

  • Save the model configuration as a preset so every review in the project uses the same settings.

Troubleshooting

The pooled estimate looks implausible

Check that all effect sizes use the same direction and scale — mixing log and raw ratios is the usual cause.

More tutorials

Browse all 79 tutorials →