# BEH Inferential Results — Publication Tables

*Generated directly from `results/beh_inferential.csv`. Seed 42; B=10,000 permutations and bootstrap resamples.*

## Table 7: Required BEH Inferential Reporting Table

| Comparison | Level | n₁/n₂ | Mean₁ | Mean₂ | ΔH | 95% CI | p (2-sided) | q (BH-FDR) | Hedges' g | Cliff's δ |
|---|---|---|---|---|---|---|---|---|---|---|
| Currier A vs B | page | 114/82 | 6.016 | 6.958 | -0.943 | [-1.123, -0.762] | 0.0001 | 0.0001 | -1.544 | -0.707 |
| Currier A vs B | paragraph | 141/111 | 5.712 | 6.407 | -0.695 | [-0.901, -0.484] | 0.0001 | 0.0001 | -0.870 | -0.457 |
| Hand 1 vs 2 | page | 86/45 | 5.840 | 6.590 | -0.750 | [-0.908, -0.595] | 0.0001 | 0.0001 | -1.810 | -0.780 |
| Hand 1 vs 2 | paragraph | 92/64 | 5.726 | 6.077 | -0.351 | [-0.557, -0.133] | 0.0006 | 0.0006 | -0.578 | -0.381 |

**Notes:**

- All four two-sided BH-adjusted q < 0.001. Two-sided BEH is **supported**.
- All four one-sided p-values (original direction A/1 > B/2) = 1.000. Original one-sided BEH is **refuted**.
- Effect sizes are medium to large (|g| = 0.578–1.810; |δ| = 0.381–0.780).
- Minor paragraph-level discrepancy: Currier B mean = 6.407 (CSV) vs. 6.453 (earlier draft); Hand 2 paragraph mean = 6.077 (CSV) vs. 6.115 (earlier draft). Counts differ by 1–2 units (unit-boundary definition). Directional finding is unchanged in both cases.

## LaTeX Table (paste into Section 9.4 of manuscript)

```latex
\begin{table}[htbp]
\centering
\caption{Inferential BEH reporting table. Two-sided permutation test,
  $B = 10{,}000$ permutations; bootstrap 95\% CI, $B = 10{,}000$ resamples;
  seed 42. BH = Benjamini--Hochberg FDR correction across four tests.}
\label{tab:beh-inferential}
\small
\begin{tabular}{llrrrrrrrrr}
\toprule
Comparison & Level & $n_1$/$n_2$ & $\bar{H}_1$ & $\bar{H}_2$ & $\Delta H$ & 95\% CI & $p$ (2s) & $q$ (BH) & $g$ & $\delta_C$ \\
\midrule
Currier A vs. B & page & 114/82 & 6.016 & 6.958 & -0.943 & $[-1.123,\ -0.762]$ & < 0.001 & < 0.001 & -1.544 & -0.707 \\
Currier A vs. B & paragraph & 141/111 & 5.712 & 6.407 & -0.695 & $[-0.901,\ -0.484]$ & < 0.001 & < 0.001 & -0.870 & -0.457 \\
Hand 1 vs. 2 & page & 86/45 & 5.840 & 6.590 & -0.750 & $[-0.908,\ -0.595]$ & < 0.001 & < 0.001 & -1.810 & -0.780 \\
Hand 1 vs. 2 & paragraph & 92/64 & 5.726 & 6.077 & -0.351 & $[-0.557,\ -0.133]$ & < 0.001 & < 0.001 & -0.578 & -0.381 \\
\bottomrule
\end{tabular}
\end{table}
```

## Summary Statement for Abstract/Conclusion

The two-sided BEH is confirmed: all four Benjamini–Hochberg-adjusted
q-values are < 0.001, with effect sizes ranging from Hedges' g = −0.578
to g = −1.810. The original one-sided prediction (Currier A / Hand 1
having higher entropy than B / 2) is refuted: all four observed ΔH < 0.