Shodhyantri Engineering Services
Research · Analysis · Editing · Publication Support
Get the statistics of your meta-analysis right and reported clearly, from extracting effect sizes to pooling models, heterogeneity, forest plots, bias checks and sensitivity analyses.
Tell us about your meta-analysis manuscript support need. We reply within one working day.
Meta-analysis manuscript support at Shodhyantri focuses on the quantitative core of your paper. Once you have a set of eligible studies, the hard part is turning their reported results into comparable effect sizes, choosing a pooling model, judging heterogeneity and presenting it all so reviewers can trust it. Small errors here, such as mixing standard errors with standard deviations or double-counting a control group, can change the conclusion.
We work alongside you on the analysis and its write-up. We help convert reported statistics into a common metric, run fixed or random-effects models in R, Stata, RevMan or JASP, interpret I2 and tau-squared, produce forest and funnel plots, and plan subgroup analyses or meta-regression. You remain the author who decides what the evidence means. We make sure the numbers and the reporting support that interpretation honestly.
A meta-analysis statistically combines results from several studies addressing the same question. Each study contributes an effect size, such as a standardised mean difference, odds ratio, risk ratio or correlation, weighted by its precision. Under a random-effects model, the analysis also estimates how much true effects vary between studies, which is reported as tau-squared, often with I2 and a prediction interval.
The manuscript usually sits inside a systematic review and follows PRISMA 2020, but the statistical sections need extra care. Readers expect the effect measure, model, estimator, software, heterogeneity statistics, handling of multiple outcomes per study, small-study effect checks and sensitivity analyses. Forest plots carry the main result, so their labelling, ordering and scale matter.
We help convert means, medians, t values, F values and proportions into the correct effect size with documented formulas.
We explain fixed versus random-effects in plain terms and help you choose and justify an estimator such as REML.
We report I2, tau-squared and prediction intervals together and help you discuss what the variation means for practice.
Forest and funnel plots are produced with clear labels, sensible ordering and consistent scales that journals can print.
When we use R or Stata, you receive the script and data file so reviewers' requests can be rerun quickly.
We help you avoid overstating pooled results, especially when studies are few, small or at high risk of bias.
Ten stages, each with a clear output. Nothing moves to the next stage until you have reviewed the last one.
— You review every stage first, then decide —
Every extracted value is traced to a table or page in the source study, with units checked.
All effects for one outcome use the same measure and direction, so higher always means the same thing.
The pooling model and estimator are stated with a reason that fits the research area and data.
Heterogeneity is described with several statistics, and the discussion reflects how much effects differ.
Conclusions hold, or their fragility is reported, after sensitivity and small-study effect analyses.
Abstract, text, tables and plots show identical estimates, intervals and study counts throughout the manuscript.
Every stage is shared with notes explaining what was changed and why. You and your supervisor check it, and the next stage starts only after your go-ahead.
| Common problem | How Shodhyantri helps |
|---|---|
| Pooling studies with very different designs, populations or outcome definitions. | Check clinical and methodological similarity first, and pool only where the combined estimate would mean something to readers. |
| Confusing standard error with standard deviation when calculating effect sizes. | Read each study's table notes carefully and convert standard errors to standard deviations using the sample size. |
| Choosing the fixed-effect model because it gives a narrower confidence interval. | Choose the model before seeing results, based on whether studies plausibly share one true effect. |
| Interpreting a high I2 as proof that the analysis is invalid. | Report tau-squared and a prediction interval alongside I2, and explore sources of variation instead of abandoning pooling outright. |
| Running funnel plot tests with only a handful of studies. | Note that small-study tests have low power with few studies, and avoid strong conclusions about publication bias. |
| Running many unplanned subgroup analyses and reporting only significant ones. | Limit subgroups to those planned in the protocol, report all of them and label any post hoc analyses clearly. |
List subgroup and moderator analyses in the protocol with a reason for each. Reviewers treat unplanned ones with suspicion.
A prediction interval shows where the effect in a new setting may fall, which is often more useful than the pooled mean.
Decide whether higher scores mean improvement or harm, and reverse signs where needed before pooling any results.
Keep a sheet showing source values, formulas and outputs. It answers most reviewer questions about your numbers.
State whether you used REML, DerSimonian-Laird or another estimator and any adjustment such as Hartung-Knapp.
Sort studies by year, weight or subgroup so readers can see patterns rather than a random sequence.
When results are missing, email study authors early. Their replies take time and can add valuable data.
Many journals welcome analysis scripts as supplementary files, and they help reviewers verify your results quickly.
Your work is handled by people who read your field's literature, not by a single generalist.
A PhD-qualified specialist from your field who checks the research logic, gap and contribution.
Handles statistics, models and qualitative coding, and documents every analytical decision.
Improves structure, flow, clarity and academic tone without changing your meaning.
Builds and validates MATLAB, Python, ANSYS or COMSOL work where your study needs it.
Checks grammar, consistency, numbering, citations and the small errors examiners notice.
Applies your university or journal template, reference style and front matter exactly.
We commonly work in R with the metafor or meta packages, Stata, RevMan and JASP. The choice depends on your field, journal expectations and what you can run later. We share scripts or project files so you can reproduce and update the analysis during revision.
Technically two studies can be pooled, but estimates of heterogeneity are unreliable with very few studies, and tests for publication bias need many more. We help you judge whether pooling adds value or whether a narrative synthesis would describe the evidence more honestly.
I2 estimates the share of observed variation due to real differences between studies rather than chance. A high value signals inconsistency but does not tell you how large the differences are. Tau-squared and the prediction interval describe that more directly, so we report them together.
We explain what the statistics show and where caution is needed, but the scientific interpretation and conclusions are yours as author. We help you check that the wording in the abstract and discussion matches the strength of the pooled evidence.
No. Acceptance depends on peer reviewers and editors, who judge the question, methods and contribution. We help make the analysis sound, transparent and reproducible, which addresses the statistical concerns reviewers raise most often, but we cannot control the journal's decision.
Yes. Established methods estimate means and standard deviations from medians, ranges or interquartile ranges. We apply a suitable method, record it in your methods section and, where helpful, run a sensitivity analysis excluding those estimated values.
Yes. Systematic review support covers the protocol, searches, screening, appraisal and PRISMA reporting. This page covers the statistics once studies are selected: effect sizes, models, heterogeneity, plots, bias tests and sensitivity analyses. Scholars often use both services on one project.
We provide research guidance, statistical analysis, editing, formatting and training support. All research content remains the scholar's own work, in line with UGC and university academic integrity norms. We do not guarantee publication, acceptance, indexing or academic outcomes: journals decide through peer review and universities award degrees.
Share your topic, draft or data. We will tell you honestly what it needs, and what it does not, before you pay anything.