Compute per-theme statistics for a Mode 1 run

Mode 1's analog of aggregate_theme_statistics for Mode 2/3. Mode 1 has no sentiment / emotions / intensity (the AI didn't run coding or sentiment over the corpus – the researcher did), so this helper returns only what is meaningful in Mode 1: per-theme entry count, T0.2 participant spread, and provocation rollups (count by category + total + drop count).

Usage

compute_mode1_theme_stats(
  data,
  theme_set,
  reflection_log,
  quotes_per_theme = 3L,
  config = NULL
)

Arguments

data

Tibble with std_id, std_text, plus theme_membership_* columns or an emerged_themes column. Must carry std_author when T0.2 spread is desired.

theme_set

Researcher-authored ThemeSet.

reflection_log

Populated ResearcherReflectionLog (post provocateur loop).

quotes_per_theme

Integer; number of representative quotes to select per theme. Wired through from config$analysis$themes$quotes_per_theme; defaults to 3.

config

Optional ThematicConfig (or config list). When supplied config$data$column_mappings$metric_columns is used to detect dataset-specific metric columns for the per-theme Median(MAD) + Mean(SD) summary line in the Mode 1 report. When NULL or empty, metrics auto-detect from the data via .detect_metric_columns.

Value

Named list keyed by theme name. Each value carries n_entries, participant_spread, provocations (count by category + total), quotes (raw representative quotes – NOT sentiment-sorted because Mode 1 has no sentiment), plus the metric fields: metric_cols (character vector) and metric_stats (named list of per-metric Median/MAD/Mean/SD/ n_observed records).