Review Papers Against Their Own Results

Visual-first critique plan (no paper found)

What I can do now — methodological checklist & likely evaluation areas

1. Data sources — Rigorous SNA on historical/textual figures must state primary sources (which edition/translations/manuscripts of the biblical text were used), parsing rules for extracting interpersonal links, and inter-annotator agreement if manual coding was used; otherwise automated parsing algorithms and their parameters must be specified. Claims about network structure without this are weak (see methodological guidance below).
   
   Newman (2003) Network Review
2. Operationalization of ties — Key design choices: are ties direct speech acts, co-mention within the same verse/chapter, kinship, conflict/cooperation, or inferred influence? Different choices produce qualitatively different network topologies; robust papers report sensitivity analyses across tie-definitions.
   
   Text-as-data methods guidance
3. Temporal dimension — Biblical narratives often span episodes; static networks can mislead. Papers should either model time-slices or show temporal robustness. Absence of temporal analysis reduces interpretability.
   
   Temporal network cautions
4. Statistical baselines and null models — Good SNA compares observed metrics to null models (e.g., degree-preserving randomizations) to test if features like high centralization or modularity are unexpected given simple constraints. Without nulls, observed structure may be trivial.
   
   Network null-model methods
5. Interpretive limits — For historical/textual subjects, network centrality does not equal historical causal influence; networks describe textual connectivity as defined by the authors' operationalization. Papers must avoid social-scientific causal language unless triangulated with independent historical/archaeological evidence.
   
   Quantitative-historical inference limits
6. Reproducibility — Code, parsing scripts, and the exact text edition used must be published (or included as supplements) so others can reproduce node/edge construction and calculated metrics. Without these, results are effectively irreproducible.
   
   Reproducible computational research guidance

If I had the paper: exact actions I would perform (visual-first)

1. Extract the node list and edge list from the appendix or reconstruct them from the text; if absent, re-run extraction under documented rules and produce an edge CSV.
2. Plot degree distribution, compute centralities (degree, betweenness, eigenvector), and compare each metric to degree-preserving nulls (1000 randomizations) — produce Plotly figures for each metric and a table of z-scores.
3. Run community detection (Louvain/Infomap), then visualize communities overlaid on a force-directed layout; compute modularity and compare to null distributions.
4. If temporal information exists, produce animated or time-sliced graphs and show how Moses’ centrality changes over narrative episodes.
5. Produce an itemized reproducibility report: required inputs, preprocessing steps, code links, and parameter settings.

Immediate recommendations (what to send me)

• Paper PDF or DOI / official URL (required).
• If available: the node/edge CSVs used by the authors, the code repository, and the exact Bible edition/version used for extraction.

Confidence and likely evaluation if the paper follows best practice

If the authors provide (A) explicit extraction rules, (B) code & datasets, (C) null-model comparisons, and (D) temporal sensitivity analyses, the study is likely methodologically sound and interpretable; without these elements it will be difficult to accept strong historical or social-scientific claims from the network results (Network method caveats).

Next steps & options I can execute once you provide the paper or data

1. Full visual paper analysis with Plotly graphs (degree histogram, centrality barplots, community network map, temporal centrality plots) and a reproducibility checklist.
2. Critical appraisal with inline citations to methodological literature, explicit blindspots and alternative interpretations, and a short list of experiments/analyses that would falsify the paper's strongest claims.
3. If you want an automated reanalysis, I can run the analyses and provide CSVs and interactive Plotly plots.