Sorry Victor, pessimists don’t really live longer

‘Pessimists are more likely to live longer’, the Mail Online tells us, while The Daily Telegraph claims, “Victory for Victor Meldrew, as pessimistic people 'live longer'”.

These headlines are based on a wide-ranging study into the associations between people’s expectations of their life and how accurate their predictions turn out to be, as well as various health outcomes.

The researchers found that the more participants overestimated their future satisfaction, the higher their risk of disability or death over the following decade. They speculate that people with a ‘happy-go-lucky’ attitude may cut corners when it comes to personal health and safety, which may increase their risk of disability or death.

Yet despite the headlines, there was no significant association between underestimating future satisfaction (‘being pessimistic’) and risk of disability or death when compared to people who accurately predicted future satisfaction.

The research has a number of limitations. It is unclear how accurately it measured a person’s optimism or pessimism. The reliability of the measures of disability or mortality is also unclear.

Sadly for the Victor Meldrews and Eeyores of this world, this study does not prove that a dark and dreary outlook will lead to a long and healthy life.

Where did the story come from?

The study was carried out by researchers from the University of Erlangen-Nuremberg, the University of Zurich, Humboldt-University of Berlin, the German Institute for Economic Research and the Max Planck Institute for Human Development. The research was funded by the Volkswagen Foundation.

The study was published in the peer-reviewed medical journal Psychology and Ageing.

The headlines declaring, “Being negative is good for you” do not really reflect the research results. The study found the more that people overestimated their future happiness (a group deemed optimists), the higher their risk of disability and death. However, no significant differences were seen among individuals who underestimated their future satisfaction (dubbed pessimists). So, headline writers would have been better off claiming ‘hubris confirmed’ or ‘pride comes before a fall’.

However, the journalists and editors can be forgiven to a certain extent as they may have been misled by the title of the research paper: ‘Forecasting life satisfaction across adulthood: benefits of seeing a dark future?’.

What kind of research was this?

This was a prospective cohort study assessing people’s abilities to predict their future satisfaction with life, and to determine whether their predictions were associated with future health.

The researchers suggest that our ability to anticipate our future state of mind “may have a strong impact on health and longevity”, but most people who try to predict how they’ll feel in the future get it wrong, both in terms of general and emotional wellbeing.

There are different schools of thought about how the way we predict our likely future outcomes will influence our health. Some suggest that an optimistic outlook may be protective in the face of unalterable circumstances, such as developing a long-term disease or experiencing the breakdown of a relationship. This may then help reduce feelings of uncertainty, anxiety and stress.

Others suggest having pessimistic or realistic views may aid in coping with anxiety or uncertainty.

The authors also suggest that one’s age may influence one’s outlook, with younger people tending to be more optimistic about their future selves, and older people tending to be more realistic.

The researchers investigated the relationships between the accuracy of predicting life satisfaction, and how this was linked to health. They also assessed how these predictions varied in different age groups, and whether other factors may influence the accuracy of these predictions.

An inherent limitation of this type of research is that it can tell us about whether there are associations between outlook and future health, but cannot tell us whether one causes the other.

What did the research involve?

The researchers enrolled over 10,000 individuals aged 18 to 96 years and investigated differences in anticipated life-satisfaction across several age groups.

At the beginning of the study, they collected data on education levels, self-rated health, and income. Each year for 11 years, they collected information on current life satisfaction (on a scale of 0 [totally unsatisfied] to 10 [totally satisfied]) and anticipated satisfaction in five years time (using the same scale). At the end of the study, the researchers collected information on participant health, including data on any disabilities and deaths that occurred.

Analysing the difference between people’s current life satisfaction and predicted life satisfaction

Researchers first analysed the data to determine whether there were differences in how people rated their satisfaction with life or with their predicted satisfaction in different age groups. They expected that there would be no differences in current measures, but that older adults would anticipate a decrease in their future satisfaction, while younger adults would anticipate an increase.

Determining the accuracy of people’s life satisfaction predictions

The second analysis assessed the accuracy of the participants’ predictions, and whether or not this accuracy changed over time. To determine accuracy, the researchers calculated the difference between the future life satisfaction rating and the ‘current’ life satisfaction rating measured five years later. A positive value represented an overestimation of future satisfaction (overly optimistic), while a negative value indicated the person underestimated their future satisfaction (overly pessimistic). A value at or near zero indicated an accurate prediction (realistic outlook).

The researchers expected younger adults to overestimate their future satisfaction, and older adults to underestimate it.

Analysing external influences on prediction accuracy

In the third analysis, they used the data collected at the beginning of the study on education, income, and subjective health to determine if any of these factors contributed to the accuracy of individual predictions.

The researchers expected better baseline health, more education and higher income to be associated with a less pessimistic view of the future.

Determining whether accuracy of predictions influence death or disability

In the fourth analysis, the study authors assessed whether the accuracy of predictions was associated with the risk of disability or death over 10 years. This was calculated as the risk of disability over 11 years, and the risk of death over 12 years. The reported hazard ratios (HR) represent the increase in risk of disability or death for each standard deviation increase above the group average in an individual’s future life satisfaction estimate.

They expected that in old age, a realistic or pessimistic outlook would be associated with better health and lower risk of dying.

What were the basic results?

Current and future satisfaction across age groups

When assessing the differences in current and future satisfaction across age groups, the researchers found there were no significant differences in reported levels of current satisfaction with life. However, younger adults both rated their predicted future life satisfaction to be higher than other age groups, and their predictions declined at a lower rate. Older adults had the lowest levels of anticipated future satisfaction, and this declined at the highest rate over time.

Accuracy of predictions

When assessing the accuracy of predictions of future life satisfaction, the researchers found that:

• younger adults (aged 18 to 39 years) tended to be over estimate their future satisfaction – or be overly-optimistic
• middle aged individuals were more realistic in predicting future feelings
• older adults were found to underestimate future satisfaction – or be overly-pessimistic

Factors influencing predictions

The researchers then assessed the correlation with accuracy and personal characteristics, and found that older age, less education, higher levels of self-reported health, less decline in self-reported health, higher income and increases in income were each associated with an underestimation of future satisfaction. The strength of these associations was less pronounced in older people.

Effect of predictions on future health outcomes

Finally, when assessing the association between predictive accuracy and future health, the researchers found that overestimating one’s future life satisfaction was associated with higher disability over 11 years (Hazard Ratio [HR] 1.095, 95% confidence interval (CI) 1.018 to 1.178). This represents a 9.5% relative increase in risk of disability over 11 years the more the participants overestimated their future satisfaction.

The researchers found a similar increase in mortality risk (HR 1.103, 95% CI 1.038 to 1.172), with optimists having a 10.3% higher risk of dying over 12 years the more they overestimated future satisfaction. On the other hand, no significant differences in disability or mortality were seen as individuals underestimated their future satisfaction. The outcomes among this group were also not significantly different from those of individuals who accurately predicted future satisfaction levels.

How did the researchers interpret the results?

The researchers concluded that “foreseeing a dark future is beneficial for survival”, and that taken together, their results “suggest that the accuracy of predicting future life satisfaction has functional implications and consequences”.

Conclusion

This research suggests our ability to accurately predict our future satisfaction may be linked to our future health. However, the limitations of this study should be considered when interpreting the results.

First, the researchers used different numbers of participants in their analyses for each of their four questions. This makes it difficult to compare results across the four analyses as the same individuals were not included in each assessment, and may introduce bias into the analysis.

For instance:

• the first analysis included 11,131 individuals with data on current and future satisfaction estimates
• the final analysis included 6,749 individuals with data across the entire study and disability data, as well as 7,920 individuals with satisfaction and mortality data

While including only individuals with the relevant data has clear practical advantages, making no attempt to model or account for missing information can bias the results, as the individuals who continuously participated in the study over 11 years may be quite different from those who drop out. If this difference is linked to either of the factors under investigation, this may undermine the results. For instance, if participants with depression were both more likely to report a pessimistic outlook and to drop out of the study and therefore not be included in the analyses, this may obscure the relationship between outlook and disability or mortality.

Another problem with interpreting this research is the question of whether accurately being able to predict future satisfaction truly represents a pessimistic or optimistic outlook. Indeed, the researchers also included an item in their interview that intended to more directly measure self-reported optimism (by asking individuals “when thinking about the future in general, how optimistic are you?”). This measure of optimism was only moderately associated with the measure of further life satisfaction, which was the measure used for all the analyses. Whether the more direct measure of optimism was associated with future disability or mortality was not reported.

It’s also worth considering the fact that disability was assessed with a single self-reported measure: asking whether the person had been “officially certified as having a reduced capacity to work or as being severely handicapped”. There are other ways of measuring disability which are likely to be more reliable. Deaths were also only determined by interviews with family or neighbours, or from city registries and this approach may not reliably identify all deaths.

Overall, this study suggests that one’s ability to predict future satisfaction is related to one’s age, and may be correlated with future health.

Given the limitations of the study, there is probably not sufficient evidence to support claims that “the Victor Meldrews of the world finally have something to rejoice about”, not that they’d be inclined to rejoice in any case.