← ClaudeAtlas

researchzy-statisticslisted

Analyse the user's data: plan, probe the design (pairs, repeats, clusters, replicates), choose the method from it, run a kept script that records estimates, intervals and effect sizes, correct by family, and check the text against the results. Also sample size and power.
Researchzy/Researchzy-academic-skills · ★ 0 · AI & Automation · score 72
Install: claude install-skill Researchzy/Researchzy-academic-skills
# Researchzy Statistics This skill analyses data the user supplies. It turns the question into a plan, reads the design from the data table, chooses the method from the design, writes an analysis script that the user keeps, records every estimate with its interval and effect size in `statistics/results.json`, corrects for multiple comparisons by family, and writes the Results sentences and the Statistical analysis paragraph from that file. A script then checks that every number in the text is in the results. It also computes sample size, power or the detectable effect before a study. It does not audit the statistics of a finished manuscript (`researchzy-stats-review`), draw figures (`researchzy-figures`) or write the whole Results section (`researchzy-writing`); these are in plugin `researchzy-manuscript`, when installed. It does not attempt Bayesian models, survival analysis, survey weights, or prediction-model development or validation. Meta-analysis goes to `researchzy-meta-analysis` (plugin `researchzy-evidence`), and causal identification and propensity scores go to `researchzy-causal-inference` (plugin `researchzy-manuscript`), when installed. When the question needs something no installed skill covers, say so, name the part that needs a statistician or a dedicated tool, and stop; never improvise an approximation. ## Requirements Resolve `SKILL_DIR` to the directory containing this `SKILL.md`. Python 3.10 or newer. `probe_design.py`, `rz_results.py` and `check_resul