MFV: drop the culture map + quadrant, keep range vs pooled reference (invariance failure)

MFV country norms fail cross-country measurement invariance (Jimenez-Leal 2025: non-invariance + DIF,
'cross-cultural comparisons restricted') and are stitched from 5 studies, so the culture map/quadrant
drew false structure (inverted West vs Latin America). Delete plot_mfv_map + plot_mfv_value and the
value_coords_contrast/axis_contrast helpers; MFV keeps only the range plot, now against ONE pooled
human reference (mean of the 8 z-scored samples), no per-country identity. Add data-dir note + README
caveats citing the sources.

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>

src/tinymfv/data/human/MFV_country_norms_NOTE.md

@@ -0,0 +1,30 @@
+# MFV country norms: used POOLED only, never for cross-country comparison
+
+`mfv_country_factors.csv` holds moral-foundation-vignette (MFV) means for 8 countries. tinymfv uses
+these **pooled into a single human reference** (the mean across samples, per foundation) for the MFV
+range plot. It does **not** plot them as a cultural map, and you should not read country-to-country
+differences off this table. Here is why.
+
+The MFV does not support cross-country comparison, and the source papers say so:
+
+- **Jimenez-Leal et al. 2025** (Collabra, doi 10.1525/collabra.128178) is the source for US / Argentina /
+  Colombia / Peru (N=1,650, one polling agency, one Spanish MFV). They ran measurement-invariance and
+  differential-item-functioning tests and found non-invariance and uniform DIF on many items:
+  > "cross-cultural comparisons with this tool are restricted."
+- **Marques et al. 2020** (Brazil, journal.sjdm.org/19/190809a): São Paulo students judged
+  individualizing violations more harshly than the US sample, but "we cannot be sure whether these
+  findings are driven by differences in culture, stimuli, or sample composition."
+- **Hopp et al. 2024** (Netherlands, JDM, doi 10.1017/jdm.2024.5): direct comparison is "hindered"
+  (behavioural data not shared) and divergences "might be more driven by instruments than translational
+  artifacts."
+
+On top of that, this table is stitched from **five different studies** (Jimenez-Leal LatAm+US, Marques
+BR, Hopp NL, Yamada JP, Crone AU undergrads) with different samples, scales, and translations and no
+shared anchor. Each country's 6-foundation profile is z-scored within itself, so a near-flat rater (e.g.
+Peru, raw spread ~0.37) gets divided by a tiny SD and whipped around by noise. Plotting these as a
+culture map produced a confident-looking result that **inverts** Inglehart-Welzel (Latin America came
+out more individualizing than the West), a red flag that it measures study/sample differences, not
+culture.
+
+So: pooled reference only. The removed MFV culture map + named-axis quadrant, and this reasoning, are in
+`docs/RESEARCH_JOURNAL.md` (2026-07-05). -- authored by Claude

src/tinymfv/value_axes.py

@@ -31,12 +31,10 @@ VALUE_AXES: dict[str, tuple] = {
           ("loyalty", 1), ("authority", 1), ("purity", 1)]),
         ("Equality", "Proportionality", [("equality", -1), ("proportionality", 1)]),
     ),
-    "mfv": (
-        ("Individualizing", "Binding",
-         [("care", -1), ("fairness", -1), ("liberty", -1),
-          ("authority", 1), ("loyalty", 1), ("sanctity", 1)]),
-        ("Liberty", "Authority", [("liberty", -1), ("authority", 1)]),
-    ),
+    # NB: no "mfv" here. The MFV named-axis quadrant was removed: MFV country norms fail cross-country
+    # measurement invariance (Jimenez-Leal et al. 2025, doi 10.1525/collabra.128178 -- non-invariance +
+    # DIF, "cross-cultural comparisons with this tool are restricted"), so a cultures map over them
+    # misleads. MFV keeps only the range plot vs a pooled human reference. See RESEARCH_JOURNAL.md.
     "big5": (
         # DeYoung's meta-traits are unipolar, but name BOTH ends so every axis reads as a contrast
         # (like the others): low Plasticity = reserved/conventional; low Stability = volatile.