Mapping the evidence with R

Learning objectives

By the end of this practical, you will be able to:

Count studies, not rows, when describing an evidence base.
Build an evidence heat map that shows clusters and gaps.
Cross a third variable from your codebook (organisms, validity…).
Say what an evidence map can and cannot tell you.

From codebook to map

This morning you designed a codebook. Each coded variable can become an axis of a map.

Today’s data: a published systematic map on diversified farming and biodiversity (Jones et al. 2021).

237 studies · 4,076 comparisons
48 countries · 1986–2021
105 coded columns in the full database (32 kept for today): practice, comparator, taxon, functional group, validity…

Our question Where is the evidence — and where is it missing — on the effects of farm diversification on biodiversity?

Coding shapes the map The published dictionary lists “cover crop”; the data contain none. The data contain “Associated plant species”; the dictionary does not define it.

Quick quiz

Show of hands

The dataset has 360 rows from 1992. How many studies is that?

360
About 36
2

One study can report dozens of comparisons (taxa, sites, dates). Rows are not studies — and they are not independent. Count with n_distinct(study_id).

Same data, two maps

What an evidence map tells you — and what it does not

It tells you

Where evidence accumulates (clusters)
Where it is missing (gaps)
How it was produced (designs, metrics, validity)
Whether a meta-analysis is feasible, and on what

It does not tell you

Whether a practice works, or how much
Whether results are consistent
Whether a gap is a priority (it may be agronomically irrelevant)

Many studies ≠ strong effect. Do not “vote count”.

Source: James et al. (2016); Collaboration for Environmental Evidence (2022).

Your 30 minutes

Time	Task	Format
14:08	Open the pack, load the data, count studies	Alone
14:13	Exercise 1 — first evidence map (fill the `___`)	Alone
14:25	Exercise 2 — add a third dimension: organisms or validity	Pairs
14:35	2–3 pairs show their map	All
Bonus	Where (world map) and when (cumulative curves)	Later

Open meta_analysis_course.Rproj (the course kit), then td_evidence_map.R. Instructions and folded solutions: the practical web page.

Key messages

Count studies, not rows — rows are non-independent comparisons.
Show the zeros: in an evidence map, the gap is the result.
Sequential colours and numbers in tiles: readable by everyone.
A map describes the evidence; it does not say what works. Your codebook sets its axes.

In the online notebook

Each point of this session developed further, with the same data, more pitfalls and exercises with solutions:

literaturesynthesis.github.io/notebook

References

Collaboration for Environmental Evidence. 2022. Guidelines and standards for evidence synthesis in environmental management. Version 5.1. (A. S. Pullin, G. K. Frampton, B. Livoreil, and G. Petrokofsky, Eds.). Collaboration for Environmental Evidence.

James, K. L., N. P. Randall, and N. R. Haddaway. 2016. A methodology for systematic mapping in environmental sciences. Environmental Evidence 5:7.

Jones, S. K., A. C. Sánchez, S. D. Juventia, and N. Estrada-Carmona. 2021. A global database of diversified farming effects on biodiversity and yield. Scientific Data 8:212.