GLiTRS Dynameta
Introduction
GLiTRS Dynameta is designed for the ingestion and interactive meta-analyses of insect biodiversity effect sizes, oriented around the effect of anthropogenic threats (based on the IUCN threats classification scheme) on biodiversity. If you encounter any issues or bugs while using GLiTRS Dynameta, please raise this with Joe Millard.
The platform is split into 5 main pages:
- Use this 'Introduction' to familiarise yourself with the site and its layout.
- Use 'Summary' to explore our findings in an engaging way.
- Use 'Explore data' to access the data behind the project, visualise its geographic distribution, and run custom meta-analytic models investigating the effect of anthropogenic threats on insect biodiversity.
- Use 'Read more' for any documentation on the appropriate approach for carrying out a GLiTRS meta-analysis.
For any publications using GLiTRS Dynameta, please cite both our original software article and the database publication:
- Millard, J., Skinner, G., Bladon, A. J., Cooke, R., Outhwaite, C. L., Rodger, J. G., Barnes, L. A., Isip, J., Keum, J., Raw, C., Wenban-Smith, E., Dicks, L. V., Hui, C., Jones, J. I., Woodcock, B., Isaac, N. J., & Purvis, A. (2025). A Multithreat Meta‐Analytic Database for Understanding Insect Biodiversity Change. Diversity and Distributions. DOI: https://doi.org/10.1111/ddi.70025
- Skinner, G., Cooke, R., Junghyuk, K., Purvis, A., Raw, C., Woodcock, B.A., Millard, J. (2023). Dynameta: a dynamic platform for ecological meta-analyses in R Shiny. SoftwareX. DOI: https://doi.org/10.1016/j.softx.2023.101439
How do human activities affect insects?
Click the insects below to find out how they are each affected by human activities.
Click here to find out more about insect Orders
The colour of the icons represents how they are affected by the threat: red indicates a decrease and blue indicates an increase in sites with the threat compared to sites without. Grey insect orders are not significantly affected by the threat.
The information displayed above does not distinguish between insects that are harmful vs. those that are beneficial to humans. For example, the order Lepidoptera (butterflies and moths) includes species and moths whose caterpillars are major crop pests, as well as species which are important pollinators. Hence, viewing the effect of a threat on a whole insect Order may hide important differences within that Order in terms of the insects' effect on people and ecosystems. Please also note that this does not represent the entirety of scientific knowledge, but only what the GLiTRS project has found and synthesised.
Scientific notes:
Please note that this only uses abundance data with logRR errors at the moment. Best and worst case scenarios correspond to upper and lower 95% confidence intervals, respectively, calculated automatically by the function metafor::rma.mv
Overview
Disclaimer
Due to the limited amount of data in the GLiTRS Dynameta database at this time, and the risk of multiple testing issues, this tool is best used for hypothesis generation rather than meta-analysis for publication at this time. If the results are to be published as they stand, it is very important to formulate, and ideally pre-register, the research question prior to interacting with the dynamic meta-analysis tool, to avoid (founded or unfounded) accusations of p-hacking (i.e. manipulating the search strings until results meet expectations and/or statistical significance). On the contrary, if the results are used only for hypothesis generation, this may in some circumstances be valid as long as it is declared as such (i.e. 'We searched for trends within the GLiTRS Dyameta database, to identify fields in which further study might be fruitful').
The GLiTRS Dynameta data
View original Dynameta source code.
Run meta-analyses
Use this tab to investigate how different anthropogenic threats impact insect biodiversity. These models are multilevel meta-analytic models, run using the metafor package. They account for the non-independence of the data by specifying Paper_ID and Observation_ID as nested random effects. The effect size used to compare biodiversity is, for some of the data, the log transformed Ratio Of Means (ROM) (also known as the log response ratio), and for the rest, Hedges' D (aka Hedges' g). The log response ratio quantifies proportionate change between treatments, while Hedges' D quantifies the difference between the means in terms of their standard deviations, with a correction for small sample size.
Based on your research question, below you can filter the data by threat, taxonomic order, and biodiversity metric. Once you have made your selections, click 'Run custom model'. The model will then run in real-time (this may take a while), with the results presented as a forest plot.
Note that users should be conscious of the dangers of multiple testing, as described here by Dr Matthew J Grainger (who is unaffiliated with this project), when using GLiTRS Dynameta to run dynamic meta-analyses. Please see our publication in SoftwareX, entitled 'Dynameta: a dynamic platform for ecological meta-analyses in R Shiny', for more details.
Use this section to download the results.
- Click 'Download R custom model summary' to download a .txt file containing the output of the summary() function applied to the custom model object. This provides a results summary of the model fitting.
- Click 'Download R custom model object' to download a .rds file containing the model object. This has additional attributes attached, which specify the date and time the model was run, the filters that were applied, and the R session information. Once downloaded, use the readRDS() and attributes() functions to load the model object and view its attributes. By downloading, it allows the same analysis to be repeated at a later date (perhaps after more data has become available).
- Click 'Download forest plot' to download a .png file of your forest plot.
Read more
All meta-analyses performed by the GLiTRS team were carried out in line with our protocol, preregistered on the Open Science Framework (OSF) here . The OSF project contains: 1) a detailed guidance document; 2) a skeleton spreadheet to complete; 3) a skeleton protocol to complete; 4) a spot check spreadsheet to complete; and 5) a set of prior contributor protocols.
Please refer to and follow the guidance developed by communities of practice when conducting your own evidence syntheses using this resource:
- Collaboration for Environmental Evidence (CEE) - For conducting environmental evidence syntheses. See here for full guidance document.
- PRISMA - For reporting systematic reviews and meta-analysis (developed for medical field but see ecology specific version below).
- PRISMA Extension for Ecology and Evolution
- ROSES - For reporting systematic reviews and meta-analysis (developed for environmental research).
- Cochrane Handbook for Systematic Reviews of Interventions - For conducting systematic reviews (developed for medical field).
- Campbell Collaboration - For conducting evidence syntheses (developed for social sciences field).
- Dynamic Meta-analysis: When Transparency Meets Multiplicity by Dr Matthew J Grainger - Blog on the dangers of multiple testing in dynamic meta-analysis
The following are useful guides for conducting meta-analyses:
The data used here come from seven meta-analyses, four from pre-existing literature and three which were carried out by GLiTRS contributors.
Prior meta-analyses:
Please note that we have only used a small subset of data from this paper. This is because we found a bug in the function they used to impute the standard deviation where it was not found in the literature, so we have only used data on which they did not use this function.
GLiTRS meta-analyses: