The dark side of conscientiousness: conscientious people experience greater drops in life satisfaction following unemployment

Journal of Research in Personality 44: 535–539, DOI 10.1016/j.jrp.2010.05.001. University of Warwick and University of Manchester. Read as the accepted manuscript ("Accepted for publication in: Journal of Research in Personality"), posted on the first author's own website (journeyforhappiness.co.uk, publication list). The file there is named "...2011_jrp...". The published version was not read. Provenance: papers/carry_on/boyce2010_conscientiousness_unemployment.provenance.json.

What was read

All 814 lines of pdftotext -layout output (20 pages): abstract, introduction, method, results, discussion, references, author notes, footnotes, Table 1 and the Figure 1 caption. Figure 1 was rendered and read.

Question

Does conscientiousness measured before job loss change how much life satisfaction falls during unemployment? The hypothesis: more for conscientious people, for whom job loss is failure at an important goal.

Method

• Data. SOEP 2005–2008: 9,570 people, all employed in 2005 (age 17–83, mean 40).
  • Each later year is coded employed, or unemployed for 1, 2 or 3 years so far.
  • The number of people who became unemployed is not reported.
• Measures.
  • Life satisfaction: single 0–10 item.
  • Conscientiousness: the 3 SOEP BFI-S items from 2005 (α = .63), standardized.
• Model. Multilevel model of life satisfaction on 2005 life satisfaction, conscientiousness, the three unemployment-duration dummies and conscientiousness × dummies, with random slopes.
  • Regression 2 adds pre-unemployment age, gender and income, with their interactions.

Results

• Average drops against the employed (0–10 points): −0.60 in year 1, −0.83 in year 2, −0.72 in year 3 (d −0.37, −0.50, −0.44).

• Conscientiousness deepens the drop over time.

  • Interactions: −0.13 (p < .001), −0.15 (p < .05) and −0.32 (p < .01) for years 1–3.
  • With controls: −0.13, −0.14 (n.s.) and −0.40 (p < .001).

  Group	Year 1	Year 2	Year 3
  +1 SD conscientious	−0.69	−0.93	−1.00
  −1 SD conscientious	−0.52	−0.72	−0.45

  • In year 3 the high-conscientiousness drop is "120% higher".
  • Figure 1: the low-conscientiousness group starts recovering in year 3. The high group keeps falling.

• Selection.

  • People who later became unemployed had lower life satisfaction in 2005 (6.41 against 7.23), and slightly lower conscientiousness (17.53 against 17.74).
  • Conscientiousness predicted unemployment without controls (β = 0.07) but not with age, gender and income (β = −0.02).

• Conscientiousness correlates with life satisfaction at r = .03 in this sample. Footnote 2: among the unemployed the correlation is negative but not significant.

Limits

• Three years, a single-item outcome and a 3-item trait scale (α .63).
• The number of unemployed people and person-years per duration is never given. The year-3 estimate (SE 0.13) rests on few people, and the year-2 interaction is not robust to controls.
• What happened to people who left the labour force is not stated.
• The authors call the dummy coefficients "interpretable as causal effects" because of the lagged control. That is an overstatement: controlling for prior life satisfaction does not remove all selection, and the paper shows selection on both life satisfaction and conscientiousness.
• Inconsistencies found:
  • Direction of selection. The descriptives say future unemployed people were less conscientious (17.53 against 17.74, "statistically"). Yet the logistic regression reports conscientiousness predicting unemployment with a positive β = 0.07. Either the logit predicts staying employed, or one of the two is misreported. The paper does not say which.
  • Positive correlation. The text calls the conscientiousness–life-satisfaction correlation "significant positive" and builds on "employed conscientious individuals tend to have a higher life satisfaction". It is r = .03, an order of magnitude below the .2–.3 the introduction cites.
• Checked:
  • Simple slopes, including the conscientiousness main effect of 0.04: −0.69 / −0.51, −0.94 / −0.72, −1.00 / −0.44. These match the text and the plot.
  • 1.00 / 0.44 = 2.27, i.e. "120% higher" (the text uses 0.45).
  • d = b / 1.65: −0.36, −0.50, −0.44, matching Table 1.

What it means for Kurisutina

• Q1 and the pilot's lost-job stratum: trait direction can run against the stereotype.
  • A generic "conscientious people cope better" story predicts the wrong direction here. Conscientious people lost more life satisfaction after job loss, and the gap grew with time (verified).
  • A base model given a conscientious persona might well tell the stereotyped story (inferred). A persona probe along these lines could check it, and so could the pilot's empty-versus-own read-out if a conscientiousness proxy were available (GSS has none).
• Weigh it against the rest of the reading.
  • Anusic 2014 found little personality moderation of life satisfaction (1 of 40 terms). Specht 2011 found that unemployment changed no trait.
  • Boyce's two moderation results (conscientiousness here, agreeableness after disability) come from one group, on SOEP, with short scales, few event cases, and effects appearing late.
  • Inferred: plausible, not established. A replica should not encode them as rules. They are hypotheses for LISS T1 to test with declared analyses.
• Timing again. As with disability, the differences appear in years 2–3, not at the event. A design with one post-event wave would miss them.

Cross-references

• summaries/carry_on/boyce2011_disability_agreeableness.md: the same design for disability and agreeableness.
• summaries/carry_on/anusic2014_hilda_personality.md: little moderation in HILDA.
• summaries/carry_on/specht2011_personality_life_events.md: SOEP events and trait change.
• summaries/carry_on/haehner2024_event_characteristics_swb.md: for job loss, looking for work predicted recovery.
• summaries/carry_on/luhmann2012_adaptation.md: slow recovery from unemployment on average.