Immune Neglect: A Source of Durability Bias in Affective Forecasting
Journal of Personality and Social Psychology, 1998 · paper · Psychology
Six studies examine why people overestimate how long they will feel bad after negative events, such as a breakup or electoral defeat. The authors conclude people neglect the psychological immune system that softens negative emotions.
Checked: PDF from dtg.sites.fas.harvard.edu, fetched 5 Oct 2026. Page numbers are those of that PDF.
8 findings
Substance Study 3 text says "10 were losers and 15 were winners", but Table 4 gives n = 15 for Losers and n = 10 for Winners in every column.
- p. 9 Results, Response Rates
10 were losers and 15 were winners
- p. 9 Table 4, n rows
Losers
- p. 9 Table 4
Losers ... n 15 15 15; Winners ... n 10 10 10
- p. 9 Table 4, Losers n / Winners n
n 15 15 15
Only the authors can say what is right here.
Judge's note: Checked on the p. 9 image. The text and table disagree on the group sizes, and the text says all analyses used these 25 participants. Losers (Richards supporters) were the majority at recruitment (39 vs 18), which favours the table's split. Only the authors can confirm which is correct.
Second check: another model agreed
Report a problem with this finding
Substance Study 5 text gives experiencers' baseline "(M = 2.16, SD = 1.42)", but the Table 6 experiencer baseline cells (2.50, SD 1.83, n = 16; 1.88, SD 0.96, n = 13) pool to M = 2.22, SD = 1.51.
- p. 14 Baseline Affect
experiencers (M = 2.16, SD = 1.42)
- p. 15 Table 6, Experiencers baseline
1.88
- p. 15 Table 6, Experiencers baseline
Blameworthy story (n = 16) M 2.50 ... Blameless story (n = 13) M 1.88
- p. 15 Table 6 Experiencers baseline
Blameworthy story (n = 16) ... 2.50 ... Blameless story (n = 13) ... 1.88
Only the authors can say what is right here.
Judge's note: Checked the Table 6 values on the p. 15 image. The forecaster pooled mean reproduces as 2.74, as printed. The experiencer weighted mean is 2.22 and cannot reach 2.16 within rounding of the cell means. The pooled SD of about 1.51 also differs from 1.42. The df (139) matches the total n of 143 minus 4, so no missing participants are indicated.
Second check: another model agreed · checked by hand
Report a problem with this finding
Reference or label The reference list prints Rosenthal & Rosnow (1985) as "Content analyses: Forced comparisons in the analysis of variance".
- p. 21 References
Content analyses: Forced comparisons in the analysis of variance
- p. 5 Footnote 2
focused contrasts that use error terms from an ANOVA (Rosenthal & Rosnow, 1985)
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that the book's actual title is "Contrast Analysis: Focused Comparisons in the Analysis of Variance". The page image shows the printed wording, so this is not an OCR artefact. Footnote 2 cites the book for focused contrasts, which fits the reviewer's proposed title.
Second check: another model agreed
Report a problem with this finding
Reference or label Study 3 evaluation results say "As the lower portion of Table 3 suggests", but the Evaluation of Bush data are in the lower portion of Table 4; Table 3 is the Study 2 tenure table.
- p. 10 Evaluation Measures, Experiences
As the lower portion of Table 3 suggests
- p. 10 Evaluation Measures, Experiences
As the lower portion of Table 3 suggests, winners evaluated
Judge's note: The p. 10 image shows 'Table 3'. Table 3 (p. 8) contains only Study 2 happiness data. Table 4 (p. 9) has an 'Evaluation of Bush' lower portion.
Second check: another model agreed
Report a problem with this finding
Reference or label Study 4 Method says the happiness forecasts were made on scales "ranging from very happy (1) to very unhappy (7)" (experiencers used "the same scale"), while the Table 5 note says "Greater values indicate greater actual or predicted happiness".
- p. 12 Procedure, Forecasters
ranging from very happy ( 1 ) to very unhappy (7)
- p. 13 Table 5 note
Greater values indicate greater actual or predicted happiness.
Only the authors can say what is right here.
Judge's note: Checked the p. 12 and p. 13 images. The text treats lower Table 5 values as less happy (infallible-source experiencers 3.80 'less happy' than fallible-source 5.31). No reverse-scoring is mentioned. The scale anchors in the Method conflict with the table note.
Second check: another model agreed
Report a problem with this finding
Reference or label Study 6 results say "As shown in Table 6, dual-role participants in the fair and unfair decision conditions...", but Table 6 is the Study 5 table; the Study 6 data are in Table 7.
- p. 17 Accuracy of Affective Forecasts
As shown in Table 6, dual-role participants
- p. 17 Accuracy of Affective Forecasts, Study 6
As shown in Table 6, dual-role participants in the fair and unfair decision
- p. 17 Study 6, Accuracy of Affective Forecasts
As shown in Table 6, dual-role participants in the fair and unfair
- p. 16 Table 7 caption
Affective Forecasts and Experiences of Participants in Study 6
Judge's note: The p. 17 image shows 'Table 6'. The next sentence reads 'As also shown in Table 7'. Table 6 (p. 15) is captioned Study 5 and has no fair or unfair decision conditions.
Second check: another model agreed
Report a problem with this finding
Reference or label Reference "McCrae, R. R., & Costa, P. T., Jr. (1994). The stability of personality: Observations and evaluations. Current Directions in Psychological Science, 6, 173-175." gives volume 6 for a 1994 article.
- p. 21 References
Current Directions in Psychological Science, 6, 173-175.
- p. 21 References
Current Directions in Psychological Science, 6, 173-175
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: Current Directions in Psychological Science began in 1992 (vol. 1), so 1994 is vol. 3, and this article is usually cited as 3(6), 173-175. The crop of the p. 21 reference list confirms that the entry prints '1994' and 'Science, 6, 173-175'. The year/volume mismatch, probably the issue number printed as the volume, cannot be checked from the files alone.
Second check: another model agreed
Report a problem with this finding
Wording "Although five of these mechanisms--misconstrual, inaccurate theories, motivated distortion, and focalism--may lead people to overestimate..." names only four mechanisms; undercorrection is omitted.
- p. 3 Immune Neglect, last paragraph
Although five of these mechanisms--misconstrual, inaccurate theories, moti- vated distortion, and focalism--
- p. 3 Immune Neglect section, last paragraph
Although five of these mechanisms--misconstrual, inaccurate theories, motivated distortion, and focalism--may lead people
- p. 3 Immune Neglect section
five of these mechanisms--misconstrual, inaccurate theories, moti- vated distortion, and focalism
- p. 3 same sentence
mechanisms--misconstrual, inaccurate theories, moti- vated distortion, and focalism--
Judge's note: Checked on the p. 3 image. Five is stated but four are listed. The paper describes five non-immune mechanisms, including Undercorrection (p. 3 heading). Page 4 says 'only five were capable of causing a positive durability bias', which supports the count of five, so the list is incomplete. The General Discussion (p. 20) groups undercorrection with immune neglect as 'overestimation only'. That is about under- versus overestimation and does not resolve the count.
Second check: another model agreed
Report a problem with this finding
11 minor (typos and the like; most checked by the judge only)
Reference or label The Kammann & Flett (1983) reference gives pages "35, 257-265". The external record gives first page 259.
- p. 21
Kammann, R., & Flea, R. (1983). Affectometer 2: A scale to measure current level of genera
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the Crossref record giving first page 259. The page image shows 257-265 as printed. The image shows "Flett", so the "Flea" in the text layer is an extraction artefact.
Second check: another model agreed, code check
Report a problem with this finding
Wording Study 1 contrast "old lovers, F(1, 529) = 29.16, p < .001" and the Table 1 summary statistics (old lovers M = 5.71, SD = 1.02, n = 141; loners M = 5.17, SD = 1.31, n = 334) do not correspond under the declared test (focused contrast using the one-way ANOVA error term).
- p. 5 Lovers and loners
old lovers, F(1,529) = 29.16, p < .001
Only the authors can say what is right here.
Judge's note: Recomputed from Table 1 (p. 5 image). The pooled MSE from the three groups is about 1.488. The old-lovers vs loners contrast gives F of about 19.4, and at most about 20.2 when the means are allowed to vary within rounding. The printed value is 29.16. The young-lovers contrast recomputes to about 17.9 against the printed 18.92, so the method does not reproduce exactly either, but the old-lovers gap is much larger. Both values remain p < .001. Query to authors.
Second check: another model agreed
Report a problem with this finding
Wording Footnote 4 says "The error term associated with the main effect of experience was used in these contrasts", but the forecasters' within-subject ANOVA has only predicted outcome and time as factors.
- p. 7 Footnote 4
The error term associated with the main effect of experience was used in these contrasts.
Only the authors can say what is right here.
Judge's note: The ANOVA factors are defined on page 7, and neither is 'experience'. It is unclear which of the three error terms was used, and only the authors can say.
Second check: another model agreed
Report a problem with this finding
Wording Discussion (p. 19) states "In half of our studies, we examined forecasts and experiences of positive affect, and across those studies we found no significant evidence of a positive durability bias", but Study 2 reports recent positive experiencers less happy than forecast, F(1, 56) = 4.14, p = .047, described as "evidence of the durability bias".
- p. 19 What Are the Limits of the Durability Bias?
we found no significant evidence of a positive durability bias
- p. 8 Accuracy of positive experiencers
F( 1, 56) = 4.14, p = .047
- p. 19 What Are the Limits of the Durability Bias?
across those studies we found no significant evidence of a positive durability bias
Only the authors can say what is right here.
Judge's note: Viewed both crops. Page 8 (Study 2, tenure) reports the significant short-term positive contrast (p = .047) and the authors call it evidence of the durability bias. Page 19 Discussion says no significant evidence was found across the positive-affect studies. The results section (text near line 1170) reports a meta-analysis giving an 'anemic' average positive effect (r = .02 to .12), so 'across those studies' may refer to that aggregate. The sentence does not say this, and as worded it conflicts with the Study 2 contrast. This is a wording/ambiguity issue, so the authors should say which meaning is intended.
Second check: another model agreed
Report a problem with this finding
Typo Page 617 prints "is ordinarily predicted on the belief that"; the intended word is "predicated", which page 618 uses in the same construction.
- p. 1 Introduction, right column
is ordinarily predicted on the belief
Judge's note: The crop and page 1 show 'predicted' as printed. This is a word-choice typo.
Second check: none (judge only)
Report a problem with this finding
Typo Page 617 prints "rather than a coronet player"; the musical instrument is a "cornet".
- p. 1 Introduction, right column
rather than a coronet player
Judge's note: The crop shows 'coronet player' as printed. A player of an instrument is meant, so 'cornet' is the intended word.
Second check: none (judge only)
Report a problem with this finding
Typo In-text citation on p. 2 reads "Kahana, Kahona, Harel, & Rosner, 1988", but the reference list entry is "Kahana, E., Kahana, B., Harel, Z., & Rosner, T. (1988)".
- p. 2 The Durability Bias
Kahana, Kahona, Harel, & Rosner, 1988
- p. 21 References
Kahana, E., Kahana, B., Harel, Z., & Rosner, T. (1988)
Judge's note: The page image on p. 2 shows 'Kahona' as printed. The reference list on p. 21 has 'Kahana, B.'. The year and the other co-authors match the single entry, so this is a misspelling of the second author's name in the citation.
Second check: none (judge only)
Report a problem with this finding
Typo Study 2 response rates are printed as "47 (51.10%)" and "20 (64.50%)"; at two decimals the counts give 51.09% and 64.52%.
- p. 7 Response Rates
47 (51.10%) of the positive experiencers, and 20 (64.50%)
Judge's note: The percentages appear to be rounded to one decimal and then padded with a zero, while 34.02% for the forecasters is exact. The pool sizes (92 and 31) are given on page 6.
Second check: none (judge only)
Report a problem with this finding
Typo Study 2 reports "F(1, 44) = 0.20, p < .66", using an inequality where the paper elsewhere reports exact p values (e.g. "p = .61").
- p. 8 Accuracy of negative experiencers
= 0.20, p < .66
- p. 8 Accuracy of negative experiencers
F( 1, 44) = 0.20, p < .66
Judge's note: The p. 8 image shows 'p < .66' as printed. The surrounding contrasts use 'p ='. This is a notation inconsistency and does not change the meaning.
Second check: none (judge only)
Report a problem with this finding
Typo Page 625 says "revealed only the predicted Group x Time interaction", but the ANOVA was defined as group x measure (preelection vs forecast happiness).
- p. 9 Happiness Measures, Forecasts
only the predicted Group x Time interaction
Judge's note: The crop confirms 'Group × Time'. All the parallel Study 3 analyses use 'Group × Measure'.
Second check: none (judge only)
Report a problem with this finding
Typo Study 4 reports the SLCS comparison for positive self-views as "did not differ, F = 1", while the parallel clause reads "F < 1".
- p. 12 Classification of Participants
did not differ, F = 1, and that the SLCS
Judge's note: The p. 12 image shows 'F = 1' as printed. An F of exactly 1 is possible, so this could be either a lost '<' or a legitimate rounded value. Minor consistency point.
Second check: none (judge only)