Linear models would completely miss self-investment’s real effect on guilt

Evidence

Notably, the linear effect of self-investment on guilt was non-significant (Step 1). Conventional statistical tests assuming a linear relationship between the two variables would therefore have failed to detect a significant effect of self-investment on guilt. — Torsten Masson and Markus Barth, The identification-guilt paradox…, 61

Categorization

• QUESTION: This implies prior null findings on attachment-to-guilt links may be artifacts of testing linear-only — a methodological caution for this research’s design.
• SHARE: Worth flagging to colleagues running linear regressions on group identification where populations span both sides of the curve.
• XREF: connects to: Group-based guilt peaks at medium self-investment and drops at high levels — Torsten Masson and Markus Barth, The identification-guilt paradox…, 7; Group self-investment relates to collective guilt in an inverted-U curve. — Torsten Masson and Markus Barth, The identification-guilt paradox…, 52; Labor market experiments contradict the self-interested prediction of standard economic theory — Nicolas Baumard, A mutualistic approach to morality: The…, 150 (+5 more)

Related

• Group self-investment predicts guilt up to a tipping point where further investment suppresses it (similarity 0.84)
• Group self-investment relates to collective guilt in an inverted-U curve. (similarity 0.83)
• Group-based guilt peaks at medium self-investment and drops at high levels (similarity 0.83)
• The identification-guilt relation is curvilinear and inverted U-shaped, not linear. (similarity 0.83)
• Identification relates to collective guilt along a nonlinear inverted-U curve. (similarity 0.83)
• In-group identification relates to collective guilt in a non-linear fashion. (similarity 0.81)
• Affective group attachment, not mere membership perception, drives non-linear responses to collective guilt (similarity 0.81)
• Torsten Masson and Markus Barth — The identification-guilt paradox revisited – Evidence for a curvilinear relationship between group-level self-investment and group-based guilt (similarity 0.78)
• Both weak and strong group identifiers resist compensation but for opposite reasons (similarity 0.78)

Broader Connections

• Group-based guilt peaks at medium self-investment and drops at high levels #exemplifies (0.95) — Torsten Masson and Markus Barth, The identification-guilt paradox…, 7
• Group self-investment relates to collective guilt in an inverted-U curve. #contradicts (0.95) — Torsten Masson and Markus Barth, The identification-guilt paradox…, 52
• Group self-investment predicts guilt up to a tipping point where further investment suppresses it #exemplifies (0.85) — Torsten Masson and Markus Barth, The identification-guilt paradox…, 60
• A single-factor assumption reduces correlation estimation from factorial to linear scale. #corroborates (0.80) — Donald MacKenzie, An Engine, Not a Camera, 232
• Mass-psychology predictions missed mass man’s radical loss of self-interest #corroborates (0.75)
• Labor market experiments contradict the self-interested prediction of standard economic theory #exemplifies (0.72) — Nicolas Baumard, A mutualistic approach to morality, 150
• Self-interested rational-actor models cannot explain devotion-driven human behavior. #corroborates (0.72) — Norenzayan, Ara, Big gods - how religion transformed…, 545
• Economic models resist testing because agents self-select into treatments #corroborates (0.70) — Philip Mirowski & Dieter Plehwe (eds.), The Road from Mont Pèlerin, 1935