Declinism: Why the World Feels Worse Than the Numbers Say

Declinism is built out of three ordinary habits that each work well on their own. The first is what reaches you. Deaths, collapses and crimes are events and get reported; a child who does not die of measles and a village that gets a power line are non-events and do not. A person who follows the news closely therefore accumulates a long list of things that went wrong and almost no list of things that quietly went right, and the ratio between those two lists feels like information about the world.

The second is memory. The sharpness of bad feelings fades faster than the sharpness of good ones, so a decade you lived through comes back with its vacations intact and its anxieties flattened. Compare that edited past with an unedited present, full of problems that have not been resolved yet, and the present loses. This is the engine of rosy retrospection, and declinism is what happens when the comparison is extended forward into a prediction.

The third is that your own life changes. Richard Eibach, Lisa Libby and Thomas Gilovich showed in 2003 that people who had recently become parents were more likely than people who had not to say that crime had risen over the previous eight years, during a period when it had fallen. Becoming responsible for a small child makes the world feel more dangerous, and the extra danger is attributed to the world rather than to the change in the person looking at it.

None of this means that nothing gets worse. Plenty of measures do get worse, and which ones depends on the decade and the place. What declinism adds is the direction of the default: when a person has no figure in front of them, the answer they reach for is that the line is going the wrong way. The demo above uses six measures that moved the other way, which is why most visitors land on the gloomy side; a demo built from the measures that did deteriorate would not show the same pattern, and an honest reading of the research is about the default rather than about any one number.

The closest thing to a single founding experiment is a survey. Gapminder, the foundation set up by Hans Rosling, Ola Rosling and Anna Rosling Ronnlund, has been putting multiple-choice questions about world trends to national samples since 2013. In the Gapminder Test of 2017, run with Ipsos MORI and Novus, about 12,000 people across 14 countries answered 12 fact questions, and the average score was 2.2 correct. On the question about extreme poverty over the previous 20 years, 9 percent chose the right answer, that it had almost halved, while 60 percent said it had almost doubled. Scores that low are not what random guessing produces: on three-option questions, chance alone would have put about a third of the answers right, so the respondents were not merely uninformed but leaning in one direction.

Rosling's own reading of the pattern, set out in Factfulness in 2018, was that the wrong answers follow a shape. People do not scatter their guesses; they cluster on the side where the world is poorer, sicker and more dangerous than it is.

For the mechanism, the useful study is Eibach, Libby and Gilovich's analysis of American opinion survey data from 2000, published in the Journal of Personality and Social Psychology in 2003. Among 1,620 respondents, 38 percent of those who had become parents in the previous eight years said crime had increased over that period, against 29 percent of those who had not; 32 percent of the new parents said crime had declined, against 39 percent of the others. Crime had in fact fallen, so the most common answer among the new parents was the wrong one and the most common answer among the rest was the right one.

The largest test of the perception itself is Adam Mastroianni and Daniel Gilbert's study in Nature in 2023. Pooling archival and new survey data covering more than 12 million responses, they found that people in at least 60 nations believe morality is declining, that they have said so for at least 70 years, and that their ratings of the morality of the people around them at the time have not declined at all. The belief is steady; the thing it reports is not moving.

Declinism shows up wherever a judgment about a trend is made without looking the trend up. The rows below are illustrations, not findings from the studies cited on this page.

The counter to declinism is procedural: a feeling about a trend is not a trend, so the fix is to make the number easy to reach before the judgment gets made.

• Ask "worse than when, exactly, and measured how." A claim about decline that cannot name a year and a measure is not yet a claim that can be checked.
• Look it up, even roughly. Child deaths, crime, literacy and poverty are all published series. Thirty seconds with the source beats an hour of argument from memory.
• Read the series, not the headline. A single year is noise. What tells you about a direction is a line with a decade or more on it.
• Count the non-events. Ask what would have been reported if it had gone wrong, and whether the absence of that report means the thing went right.
• Check yourself for a recent change. A new job, a new baby, an illness or a move changes what you notice. Eibach and colleagues found exactly that pattern in new parents' judgments about crime.
• Hold the mixed picture. Some measures improve while others deteriorate, and the honest summary usually has both halves in it. Treating every number as part of one story is the error in both directions.

Declinism sits among several biases about time and about bad news, and it is worth keeping them apart.

• It is not rosy retrospection, which is the warm edit memory applies to a period you lived through. Rosy retrospection supplies one half of the comparison; declinism is the conclusion drawn from it, including the forecast.
• It is not pessimism bias, which is overestimating the chance that bad things happen to you personally. Declinism is a judgment about the state of the world, and people who are optimistic about their own lives routinely hold it.
• It is not negativity bias, the greater weight bad information carries compared with good. Negativity bias is one of the inputs to declinism rather than the same thing: it explains why the gloomy items are available, not why the trend gets called downward.