Defensive Pessimism

What it is

The optimism literature is built on instruments that score a person on a single line, with optimism at one end and pessimism at the other. Defensive pessimism is the clearest demonstration that the line conflates two different things at its lower end. One kind of low expectation belongs to someone who is not doing well and does not expect to. The other belongs to someone who is doing well, knows it, and lowers the expectation on purpose because doing so makes the next few hours bearable.

Norem and Cantor built the distinction into the design rather than arguing for it. They screened 1,027 introductory psychology students on an eight-item academic optimism and pessimism questionnaire plus a choice among four strategy profiles, one of which, a "regular" pessimistic profile the authors describe as resembling depression, was deliberately excluded. Only students who strongly endorsed having generally done well academically in the past were kept, and only those with a grade point average of 3.0 or above, with no significant difference between the groups, 3.34 against 3.23. The selected pessimists sat at an average rank of 1,009 out of 1,027 on the questionnaire and the optimists at 22. So the two groups were as far apart as the instrument could place them, and equally able.

The proposed mechanism is rehearsal rather than despondency. By playing through potential bad outcomes before they occur, in the authors' phrase, the defensive pessimist gains some degree of control over his anxiety. They go further, suggesting the strategy seems to be acting as a form of natural cognitive therapy for those using it. The low expectation is not a prediction the person is committed to; it is a way of pre-paying the worry so that it is not all due at once.

Two features of the construct constrain how far it can be taken, and both are the authors' own. The first is that it is domain-specific. The academic and social screenings in the same sample correlated at only 0.38, and the paper says these are called strategies precisely because they can be used selectively in different situations. Somebody who talks down their chances before a presentation may do nothing of the kind before a race. The second is scale. The whole study runs on 69 participants, 35 optimists and 34 pessimists, doing anagrams with false feedback.

What the construct contributes to the wider literature on optimism and health is a reason to distrust the summed score. When the Life Orientation Test is split into its two halves, almost all the predictive relationship with physical health sits at the pessimistic pole. A natural reading is that pessimism is bad for you. Defensive pessimism shows that at least one thing scoring low on such a scale is a functioning strategy used by people performing at the same level as everybody else, which means the pessimistic pole is not one object either. The direction the evidence actually points, across the cohort subscales, the consortium reanalysis and the genetic composites, is towards something closer to hopelessness and depressive symptoms than towards low expectations as such.

In effect

Sixty-nine undergraduates and an anagram task

The research case is the original study. Pessimists predicted significantly worse performance than optimists, F(1, 68) = 6.15, p = .02, on equivalent past success and equivalent ability. Actual performance did not differ at either trial, with all F values below 1, which the authors use to argue that ability does not explain the difference in expectations. Optimists showed attributional egotism, reporting much more control after success than after failure, F(1, 34) = 14.50, p = .0006 on the first trial, while the pessimists' reports of control did not differ across the two. And the finding that does most of the work: pessimists who were told they had failed were no less satisfied than optimists who were told they had failed, with all F values below 1, despite doing none of the post hoc cushioning the optimists did. The authors' conclusion is that the pessimists are not simply prone to lower evaluations of their performance, or to a generally gloomy outlook.

Talking down your chances before the viva

The everyday version is familiar enough to be a social irritant. Somebody with a strong record spends the week before an examination, an interview or a recital listing what could go wrong and insisting they expect to do badly, and then performs normally. The well-meaning response, to tell them to think positively, is the one the research gives least support to: the low expectation appears to be doing a job, and in this study the people holding it were no worse off emotionally when things went wrong. It is worth saying what the 1986 paper does not establish, which is that intervening would hurt. The study most often cited for that claim, Spencer and Norem 1996, has not been read by this publication and is not relied on here.

The premortem on a product launch

The commercial version is an institutionalised defensive pessimism. In the planning exercise Gary Klein proposed and named the premortem, a team is told to imagine the launch has already failed and to write down why, before committing to the plan. Risk registers, pre-launch failure reviews and engineering blameless post-incident practice all run the same way round: rehearse the bad outcome in detail while there is still time to act on it. The parallel is a parallel and not a finding. Norem and Cantor measured expectations, satisfaction and attributions in individuals on a word puzzle, not the performance of teams, and no result on this page tests whether a premortem improves a launch.

What it does not say

It does not say that pessimism is good for you. The study measures expectations, satisfaction and attributions after a laboratory task. It says nothing about health, and nothing whatsoever about illness or recovery.

It does not say that defensive pessimists outperform optimists. Performance was statistically indistinguishable at both trials. The claim is that the low expectation costs them nothing in performance and appears to buy them something in anxiety management.

It does not say this is a personality type. The authors call it a strategy, deployed selectively by domain, and their own two screenings in the same sample correlate at only 0.38. Treating it as a trait goes beyond the paper.

It does not say that encouraging a defensive pessimist to think positively damages their performance. That claim is widely repeated and traces to a study this publication has not obtained. Until it is read, nothing here supports it.

And it does not generalise comfortably. The sample is 69 students at one American university, pre-selected for past academic success and screened to exclude a profile resembling depression. The 1986 paper read in full is good evidence for a narrow claim, which is why this page is marked mixed.

Who developed it

Julie Norem and Nancy Cantor, in two 1986 papers written while Norem was at the University of Michigan. The study described on this page is the one in Cognitive Therapy and Research.

---