RQ1a: moral-foundations cluster structure (across 16 configs, neutral)
  mfq30: mean r within-individualizing +0.38, within-binding +0.84, between +0.03
  mfq2: mean r within-individualizing +0.15, within-binding +0.82, between +0.54
  mfq2: PROP vs individualizing +0.35, PROP vs binding +0.64

RQ1b: convergent validity across instruments (same construct, across configs)
  mfq30:CARE <-> mfq2:CARE  r=+0.13
  mfq30:FAIR <-> mfq2:EQUA  r=-0.18
  mfq30:FAIR <-> mfq2:PROP  r=+0.38
  mfq30:LOYA <-> mfq2:LOYA  r=+0.84
  mfq30:AUTH <-> mfq2:AUTH  r=+0.70
  mfq30:SANC <-> mfq2:PURI  r=+0.91
  mean convergent r (mfq30<->mfq2): +0.46
  ess:SD <-> ssvs:SD  r=+0.36
  ess:ST <-> ssvs:ST  r=+0.66
  ess:HE <-> ssvs:HE  r=+0.54
  ess:AC <-> ssvs:AC  r=+0.63
  ess:PO <-> ssvs:PO  r=+0.42
  ess:SE <-> ssvs:SE  r=+0.50
  ess:CO <-> ssvs:CO  r=+0.51
  ess:TR <-> ssvs:TR  r=+0.61
  ess:BE <-> ssvs:BE  r=+0.02
  ess:UN <-> ssvs:UN  r=+0.08
  mean convergent r (ess<->ssvs): +0.43
  heterotrait baseline (mfq30<->mfq2 different foundations): mean r +0.41

RQ1c: Schwartz circumplex test (ESS+SSVS pooled, centered scores, across configs)
  circular distance 1: mean r +0.15 (n=10 pairs)
  circular distance 2: mean r +0.13 (n=10 pairs)
  circular distance 3: mean r -0.23 (n=10 pairs)
  circular distance 4: mean r -0.35 (n=10 pairs)
  circular distance 5: mean r -0.32 (n=5 pairs)
  correlation(distance, r) = -0.43  (circumplex predicts strongly negative)

RQ2: within-config Cronbach alpha (reps as respondents; neutral; 6-item foundations)
  computable cells: 138; median alpha +0.000; cells >= 0.70: 0/138
  (degenerate/zero-variance cells excluded: 38 of 176 config-x-foundation cells)