Implications for personality development

Published in European Journal of Personality (2021), DOI 10.1177/08902070211045825; the published version was not read. Read as the PsyArXiv preprint of 1 October 2020 ("manuscript under review"; DOI 10.31234/osf.io/426g5, CC BY 4.0), distributed as a Word file. Michigan State and Ruhr University Bochum. Provenance: papers/carry_on/rakhshani2021_traits_event_perception.provenance.json.

What was read

All 891 lines of the pandoc plain-text conversion of the .docx. It covers the abstract, all sections, footnotes 1–5, references, Tables 1–4 and Supplementary Tables 1–8 (in the same file). The docx has no tracked changes or comments. Figures 1–5 (trait × ECQ correlation plots for event groups and single events) are images and were not inspected. Not read: Supplementary File 1 on OSF (osf.io/35dt9) and the data.

Questions

Six descriptive questions. The ones that matter here:

• Do Big Five traits predict how people perceive events?
• Do they predict which events people believe change personality?
• Do perceptions of imagined events relate to traits differently from perceptions of experienced events?

Method

• Sample. 1,599 US undergraduates, Fall 2019 and Spring 2020. 90 were removed on the honesty check, leaving 1,509. Mean age 19.9; 77% female, 72% white.
• Traits. BFI-2, 60 items (α .77–.90).
• Event checklist. 21 events, ever experienced.
• Beliefs. Whether each event "would change someone's personality", on a scale from −2 to +2.
• Perceptions. The ECQ, 18-item form (nine dimensions × 2 items), for one randomly chosen experienced event and two hypothetical ones (events the person had not had). Totals: 1,479 experienced and about 3,011 hypothetical ratings.
• Analysis. Descriptive correlations and odds ratios, with no correction for the many tests; the authors ask for replication of any single finding.

Results

• Traits and exposure (selection). Small but frequent associations.
  • Conscientiousness goes with having entered the workforce (OR 1.20).
  • Neuroticism goes with falling out with a close friend (OR 1.30).
  • Openness goes with moving, time abroad, serious illness (OR 1.41) and a relative jailed (OR 1.35).
  • Most events are rare in students. The text says 13 of 21 were experienced by 10% or fewer; Table 1 gives 12.
• Beliefs. Every event is believed to change personality: mean d = 1.02. The strongest are the death of a partner (1.52), becoming a parent (1.46), a friend's death and a parent's death. Traits barely relate to these beliefs (|r| ≤ .12, average .06; 101 of 105 correlations positive).
• Perceptions predict beliefs better than traits do. Average |r| is .13 against .06. Within events, challenge, emotional significance, impact and world-view change correlate .18–.20 on average with the belief; valence is inconsistent. Across the 20 event means: impact r = .87, emotional significance .82, challenge .72, world views .69, valence −.40.
• Traits barely predict perceptions (Table 4, all events pooled).
  • The largest correlations: neuroticism × challenge .22 (experienced) and .13 (hypothetical); agreeableness × social status change −.18 (experienced).
  • The average |r| per dimension is .02–.10.
  • Content matching fails. Conscientiousness is no more related to perceptions of work events than other traits are, and extraversion's links to social events are weak and inconsistent. Openness does relate to "new experience" events on several dimensions.
• Hypothetical against experienced. Trait–perception associations did not differ systematically. One exception is "had a falling out with a close friend", where agreeableness and openness show several divergences.
• Mean ratings (Supplementary Table 3; mixed event sets, so not a clean comparison). Hypothetical events were rated more challenging (4.08 against 3.68), more status-changing (2.66 against 2.03), more impactful and more extraordinary.

Limits

• Cross-sectional; students; many rare events have few experienced ratings (n = 4–14 for marriage, parenthood, divorce, graduation). Two-item ECQ scales, some with low α (world views .56–.61, impact .56–.59).
• "Beliefs about personality change" are lay predictions, not measured change.
• Experienced and hypothetical ratings cover different event mixes, so their means cannot be compared directly.
• Inconsistencies found:
  • The text says 13 events were experienced by ≤ 10%; by Table 1's frequencies it is 12 (close friend died is 10.14%).
  • The text reads OR 1.20 as a "20% greater chance" (it is odds).
  • The text calls the perception–belief associations "moderate" at an average |r| of .13.
  • Arithmetic checked: 1,599 − 63 − 27 = 1,509, and 101 of 105 trait–belief correlations are positive (Supplementary Table 7).

What it means for Kurisutina

• Question 1: Big Five scores will not tell the replica how its person appraises events. Traits explain at most about 5% of the variance in any appraisal dimension (r ≤ .22), usually under 1%.
  • Together with Kritzler et al. (appraisals vary a lot within an event) and Yap and Anusic (traits barely moderate reactions), this is the third independent null for "traits carry the person's way of reacting".
  • The slot has to carry the person's own appraisals, their stories of how past events felt, not trait summaries.
• A test that needs no waiting. If the elicitation interviews collect the person's appraisal of each past event they tell (the ECQ's nine dimensions), some events can be held out.
  • Ask the replica, given the rest of the slot, to rate the held-out events.
  • Compare against the empty-slot and other-slot replicas and against the typical profile (Kritzler).
  • This measures "reacts like her" on the appraisal step, within one interview. Proposal only.
• Hypothetical ratings as a cheaper probe, with a caveat. A person's ratings of events they have not had are easy to collect. But imagined events are rated as harsher and more impactful than lived ones, and "how would I feel" matched "how do people feel" on average (Kritzler). Their person-specific part may be small and must be checked for test–retest reliability before it serves as a target.
• Neuroticism does colour lived events as challenging (.22). This is one of the few trait–appraisal links worth building in.

Cross-references

• summaries/carry_on/kritzler2022_event_perception_profiles.md: within-event variability; outsiders' ratings.
• summaries/carry_on/haehner2022_event_perception.md: perceived valence and trait change.
• summaries/carry_on/yap2012_personality_moderation.md, summaries/carry_on/anusic2014_hilda_personality.md: traits as moderators of reactions.
• summaries/carry_on/sliwinski2009_stress_bursts.md: neuroticism and daily reactivity.
• summaries/carry_on/hout2016_gss_reliability.md: reliability ceilings for any person-level target.