Contribution
It joins two literatures. Social identity research had studied proximal causes of group behaviour: identification, entitativity, intergroup context. Terror management theory had found that reminders of death raise in-group favouritism, and explained it as defence of a worldview. The paper asks whether death reminders act on the link between person and group itself: whether they raise identification with one’s group and the perception that the group is a real entity. It answers yes in one experiment, and proposes a new function for belonging: a group gives the self a part that does not die.
Key insight
“Through identification with social groups, the individuals can project themselves in space and time, beyond their personal death. They participate in an entity that is not subject to the mortal fate that characterizes them as human beings” (p. 137). For that to work, the group “must be perceived as a ‘real’ entity” (p. 137). So a death reminder should raise both identification and perceived entitativity.
Assumptions
- Terror management theory. Awareness of death threatens paralysing terror, buffered by faith in a cultural worldview and by self-esteem (pp. 136–137). Reminders of death raise the need for the buffer.
- Social identity theory’s definition: “that part of an individual’s self-concept which derives from his knowledge of his membership in a social group” together with its value and emotional significance (Tajfel 1981, quoted p. 136).
- Entitativity, after Campbell (1958): “the extent to which a group is perceived as having real existence, as being a real entity” (p. 136).
- Mediators cause bias, not the reverse, on the strength of earlier work and of measuring them first (note 3).
Key results
-
Design (pp. 137–138). 48 undergraduates at the University of Padua, approached in libraries (95% agreed); 26 women, 22 men, aged 19–33 (M = 23); all Italian citizens; 24 per condition.
- Mortality salient: “Please write a short paragraph describing the emotions that the thought of your own death arouses in you.”
- Control: the same about “reading a book”.
- Then, in one booklet: entitativity of Italians (a modified scale; “Italians have many characteristics in common”, “Italians have a sense of common fate”, “Italy has real existence as a group”); six identification items (“I identify with Italians”, “Being Italian has nothing to do with my identity”); ratings of Italians and Germans on 10 traits. No delay or distracting task between the manipulation and the measures is reported, and no mood measure.
-
Scales (p. 138). Entitativity 19 of 20 items, α = .89; identification α = .90; trait ratings 9 of 10 items, α = .82 for Italians and .68 for Germans. Response ranges are not stated.
-
Identification and entitativity (p. 138). Identification 6.19 against 5.09, F(1, 46) = 4.86, reported p < .03. Entitativity 5.40 against 4.77, F(1, 46) = 4.15, p < .05.
-
In-group bias (p. 138, Table 1). Ratings:
Italians Germans bias mortality salient 6.70 (.75) 5.69 (.81) 1.01 (.93) control 5.88 (.98) 5.43 (.91) 0.45 (.98) Target F(1, 46) = 27.39, p < .001; condition F(1, 46) = 6.63, reported p < .01; interaction F(1, 46) = 4.11, p < .05. Bias was present in the control group, t(23) = 2.20, p < .05, and larger after the death prompt, t(23) = 5.28, p < .0001. The death prompt raised ratings of Italians, t = 3.24, p < .01, and not of Germans, t = 1.03, p > .29.
-
Mediation (pp. 138–139, Figure 1, unstandardized).
- Entitativity: condition → entitativity .62; entitativity → bias .32 (.27 with condition); condition → bias .56, falling to .39, not significant. Sobel Z = 1.71, p < .08. Path model χ²(1) = 2.05, p = .15, R² = .12.
- Identification: condition → identification 1.09; identification → bias .21 (.17); condition → bias .56, falling to .37. Sobel Z = 1.83, p < .06. χ²(1) = 1.80, p = .18, R² = .14.
-
Distinct constructs (p. 139). A two-factor model fitted better than a one-factor model (χ²(64) = 108.51 against χ²(65) = 138.86, a difference of 30.35 on 1 df; CFI .84 against .73), with the factors correlated .59. Both fits are poor by the authors’ own description.
-
Shared variance does the mediating (pp. 139–140). Neither identification’s residual (after entitativity) nor entitativity’s residual (after identification) mediated. Their predicted values did.
-
Two readings, not separated (pp. 140–141).
- Worldview defence: the group matters because it is “the milieu where worldviews are created and maintained”; identification is then an epiphenomenon of that need.
- The group as a buffer in its own right: social identities are “noncorporeal and highly symbolic” and “can be perceived as transcending the individual mortal fate”; raising entitativity is “an attempt to reify the ingroup” (p. 141).
The authors favour the second for the entitativity result and call it “tentative” (p. 141).
-
Self-sacrifice (p. 141). If the fear of death is the fear of losing one’s identity, “self-immolation for the ingroup is quite a reasonable choice: Through the social extension of the self implemented by group belonging, individuals may have the feeling that they are offered transcendence.”
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Reminders of one’s own death raise identification with one’s national group | weak to moderate: one study, N = 48, p ≈ .03 | Results |
| C2 | Reminders of one’s own death raise the perceived entitativity of one’s national group | weak to moderate: one study, N = 48, p < .05 | Results |
| C3 | The rise in in-group bias after a death reminder is in-group enhancement, not out-group derogation | moderate within the study: a clean ingroup/outgroup split, consistent with the authors’ account of earlier studies | Table 1 |
| C4 | Identification and entitativity mediate the effect of mortality salience on bias, through their shared variance | weak: marginal Sobel tests, mediators measured not manipulated | Figure 1, pp. 139–140 |
| C5 | Identification and entitativity are distinct constructs | moderate: a nested model comparison, though neither model fits well | p. 139 |
| C6 | A group seen as a lasting entity lets the self project beyond death, so that dying for it can serve the self | conjecture: proposed as tentative, not tested against worldview defence | Discussion |
Concepts
- entitativity: how far a collection is perceived as a real entity, with real existence; a degree (Campbell 1958).
- ingroup identification: here, the importance of the group to one’s identity, measured by six verbal items.
- mortality salience: an experimental reminder of one’s own death.
- worldview defence: protecting the cultural belief system that buffers the terror of death, by favouring those who share it.
- proximal and distal causes: the immediate cognitive and motivational causes of group behaviour, against fundamental needs such as the management of death anxiety.
Connections
- Terror management, measured (Burke et al., LIT-1687, unread). The meta-analysis reports moderate mortality-salience effects across paradigms. The Many Labs 4 replication of the classic mortality-salience effect on worldview defence, which failed, may bear on how much weight this study can carry; another reader is drafting that note, and the record has no code for it yet.
- Lifton’s symbolic immortality (LIT-1714), unread. The paper does not cite Lifton, but its “transcendence” through the group, and its quotation of Anthony Smith on nationalism’s “promise of collective and terrestrial immortality” (p. 140), are the same idea.
- A group entity versus a category of members (TERM-096). Italians, a nationality, is the term’s own example of a category. The paper measures entitativity as a degree, which the term sets aside as different from its difference in kind. Its scale mixes likeness (“many characteristics in common”, a category’s mark) with common fate and “real existence as a group”, closer to an entity’s.
- Skorich’s group-entity account (THEORY-1112). That account says perceiving a group entity needs cues of shared mental states. Here a death reminder, not a cue about the group, moved perceived entitativity: the perceiver’s need moved it.
- Identity fusion (NOTE-1256). Swann et al. explain self-sacrifice for the group by self-verification. Castano et al. explain it, more speculatively, by transcendence through an extended self. Both make sacrifice serve the self as a self-concept, not as a welfare.
- Self-categorization theory (THEORY-1114). The authors write within social identity theory and do not use depersonalization; their closing hypothesis is a shift “from personal to social identities” (p. 141).
Bearing on the record
-
The legacy question (QUESTION-301). This is the study the question cites for where the temporal and social extensions of the self meet. Read, it supports that meeting as the authors’ proposal, not as a finding: see the thesis below.
-
The owner’s thesis (CLAIM-491). These are inferences.
- The paper states the thesis’s join of social and temporal extension in so many words: “the social extension of the self implemented by group belonging” offers transcendence of one’s death (p. 141). On it, dying for the group is not selfless but “a reasonable choice” for a self that fears annihilation. That is the thesis’s strongest form, and the record now holds it as a published conjecture.
- The data support a weaker claim. A death reminder raised identification and valuing of the in-group. That fits “the self turns to the part that does not die”, but equally fits “the self defends the worldview the group carries”. The authors say they cannot tell these apart (p. 141). Only the first is the thesis.
- Nothing measures the self’s extent. Identification is six verbal items; no overlap measure, no fusion item.
-
The triviality objection (CLAIM-492). Entitativity is a property of the group as perceived, and it separates from identification (r = .59). But the two mediate only through what they share, and there is no measure of caring about Italians apart from rating them. So the paper does not show a criterion of the self’s extent coming apart from valuing the group.
-
Batson’s objection (CLAIM-494). Not addressed.
-
The term (TERM-096). A dated note is proposed:
Castano et al. (NOTE-1297) found that a reminder of one’s own death raised the perceived entitativity of Italians, a category in this term’s sense, in 48 Italian students (5.40 against 4.77). They read it as an attempt “to reify the ingroup” into something that outlasts the self. If so, the difference in kind between a category and a group entity is perceived as a degree, and motivation can move a category toward an entity. Their scale mixes likeness with common fate and “real existence”, so it does not say which side of the term moved.
-
No THEORY filed.
-
No ML instruction.
Limitations
- One study, 48 participants, 24 per cell, Italian students recruited in libraries. The bias effect rests on one trait list and one out-group.
- A mortality-salience effect, not a replication-tested one. No delay or distractor before the measures is reported, and mood was not measured (note 1 answers the mood objection from the literature).
- Mediators measured, not manipulated; the Sobel tests were marginal (p < .08, p < .06), and the shared-variance analysis leaves open what the common factor is.
- The proposal is not tested. Nothing measures transcendence, symbolic immortality or how lasting the group is believed to be.
- Reporting slips.
- The definitions are swapped on p. 137: entitativity is glossed as “the importance of the group for the identity of the individual” and identification as “the perception of the ingroup as a real entity”.
- Figure 1’s note reads “*p < .5”.
- The model difference is printed “χ²(30.35)”; it is χ²(1) = 30.35.
- The between-condition t-tests on ratings are printed t(24); with 48 participants in two groups the df would be 46.
- F(1, 46) = 4.86 is printed p < .03 and F(1, 46) = 6.63 p < .01; they are p ≈ .033 and p ≈ .013.
- Typos.
- “Univesité” in Yzerbyt’s affiliation; “the level identification” (p. 136); “particular relevant” and “the their cultural worldview” (p. 137); “where entered” and “goodness of fit-ness” (p. 139); “must be perceive” and “epiphenomenona” (p. 140).
- In the references: “Pyszcsynzski”, “Hogart”, “Wheterel”, and “n the national system.:” in Kelman’s title.
Open questions
- Does the effect replicate, given the doubts Many Labs 4 raises about mortality salience?
- Which reading holds? Two groups with the same worldview, one seen as lasting and one as about to dissolve, would separate transcendence through the group from defence of the worldview.
- Would a death reminder raise self–group overlap or fusion, not only verbal identification? That would test the thesis’s version directly.