Review of Silverman 2020: "Multiple-systems analysis for the quantification of modern slavery: classical and Bayesian approaches"
Professor Bernard Silverman, Home Office Chief Scientific Advisor of the United Kingdom, is a statistician who researchs human trafficking and modern slavery. In 2020, he published a paper titled "Multiple‐systems analysis for the quantification of modern slavery: classical and Bayesian approaches" in which he reviews existing methodologies for quantifying the scale of modern slavery and proposes a new method.
I have written a review of Silverman's paper, titled "Review of Silverman (2020): 'Multiple-systems analysis for the quantification of modern slavery: classical and Bayesian approaches'". The review has seven parts, which are as follows:
1. Introduction: overview of report, motivation for the research
2. Background: overview of data sets, review of basic log-linear model proposed by Cormack (1989)
3. Frequentist approaches: summary of two frequentist approaches (stepwise parameter addition and information criterion)
4. Existing Bayesian approaches: summary of two common Bayesian approaches that are useful for human rights research (graphical models and dirichlet process mixtures), discussion of the performance of these approaches on Silverman's data sets
5. Bayesian-threshold approach: summary of Silverman's proposed method (Bayesian-threshold)
6. Results: main takeaways from Silverman's results
7. Conclusion: concluding thoughts
- silverman-review.pdf: the review in pdf format
- silverman-review.Rmd: the review in Rmd format
- bibliography.bib: the bibliography for the review (referenced in the Rmd)